From daemon@optimus.ietf.org  Sat Feb  2 05:10:26 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA01420
	for <tsvwg-archive@odin.ietf.org>; Sat, 2 Feb 2002 05:10:26 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA20240
	for tsvwg-archive@odin.ietf.org; Sat, 2 Feb 2002 05:10:28 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA19322;
	Sat, 2 Feb 2002 04:48:16 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA19293
	for <tsvwg@optimus.ietf.org>; Sat, 2 Feb 2002 04:48:14 -0500 (EST)
Received: from d12lmsgate-3.de.ibm.com (d12lmsgate-3.de.ibm.com [195.212.91.201])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA01190
	for <tsvwg@ietf.org>; Sat, 2 Feb 2002 04:48:11 -0500 (EST)
Received: from d12relay02.de.ibm.com (d12relay02.de.ibm.com [9.165.215.23])
	by d12lmsgate-3.de.ibm.com (1.0.0) with ESMTP id KAA29780;
	Sat, 2 Feb 2002 10:47:39 +0100
Received: from d10hubm1.telaviv.ibm.com (d10ml001.telaviv.ibm.com [9.148.216.55])
	by d12relay02.de.ibm.com (8.11.1m3/NCO v5.01) with ESMTP id g129n5E94218;
	Sat, 2 Feb 2002 10:49:05 +0100
To: tsvwg@ietf.org, rdma@yahoogroups.com
MIME-Version: 1.0
X-Mailer: Lotus Notes Release 5.0.9  November 16, 2001
From: "Julian Satran" <Julian_Satran@il.ibm.com>
Message-ID: <OF52045508.F035F58E-ONC2256B54.003384B0@telaviv.ibm.com>
Date: Sat, 2 Feb 2002 11:49:01 +0200
X-MIMETrack: Serialize by Router on D10ML001/10/M/IBM(Release 5.0.8 |June 18, 2001) at
 02/02/2002 11:49:03,
	Serialize complete at 02/02/2002 11:49:03
Content-Type: multipart/alternative; boundary="=_alternative 0034BF55C2256B54_="
Subject: [Tsvwg] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

This is a multipart message in MIME format.
--=_alternative 0034BF55C2256B54_=
Content-Type: text/plain; charset="us-ascii"

Dear Colleagues:

I've recently submitted a draft including a different "No-Data-Touch"(NDT) 
framer.
Unlike the current proposed NDT framer that has a "mistake" probability 
that increases asymptotically (to 1) with the stream length the framer I 
am suggesting has a very low and constant mistake probability and is thus 
safe to use even in the absence of alignment or in the presence of 
re-segmentation.

It does not contain the standard RFC2026 reference not as an indication of 
patent - I have no  patent on it and I do not know about the existence of 
one (nor did I search thoroughly) - but rather as a sign of discomfort for 
it to be used without me being asked about it.

Until it apperas you can find it also at:

http://www.haifa.il.ibm.com/satran/ips/draft-satran-NDT-Framing-00.pdf

http://www.haifa.il.ibm.com/satran/ips/draft-satran-NDT-Framing-00.txt

Julo
--=_alternative 0034BF55C2256B54_=
Content-Type: text/html; charset="us-ascii"


<br><font size=2 face="sans-serif">Dear Colleagues:</font>
<br>
<br><font size=2 face="sans-serif">I've recently submitted a draft including a different &quot;No-Data-Touch&quot;(NDT) framer.</font>
<br><font size=2 face="sans-serif">Unlike the current proposed NDT framer that has a &quot;mistake&quot; probability that increases asymptotically (to 1) with the stream length the framer I am suggesting has a very low and constant mistake probability and is thus safe to use even in the absence of alignment or in the presence of &nbsp;re-segmentation.</font>
<br>
<br><font size=2 face="sans-serif">It does not contain the standard RFC2026 reference not as an indication of patent - I have no &nbsp;patent on it and I do not know about the existence of one (nor did I search thoroughly) - but rather as a sign of discomfort for it to be used without me being asked about it.</font>
<br>
<br><font size=2 face="sans-serif">Until it apperas you can find it also at:</font>
<br>
<br><font size=2 face="sans-serif">http://www.haifa.il.ibm.com/satran/ips/draft-satran-NDT-Framing-00.pdf</font>
<br>
<br><font size=2 face="sans-serif">http://www.haifa.il.ibm.com/satran/ips/draft-satran-NDT-Framing-00.txt</font>
<br>
<br><font size=2 face="sans-serif">Julo</font>
--=_alternative 0034BF55C2256B54_=--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sun Feb  3 02:46:41 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA22672
	for <tsvwg-archive@odin.ietf.org>; Sun, 3 Feb 2002 02:46:41 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id CAA11794
	for tsvwg-archive@odin.ietf.org; Sun, 3 Feb 2002 02:46:43 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA11032;
	Sun, 3 Feb 2002 02:29:00 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA11000
	for <tsvwg@optimus.ietf.org>; Sun, 3 Feb 2002 02:28:58 -0500 (EST)
Received: from c003.snv.cp.net (c003-h004.c003.snv.cp.net [209.228.32.218])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id CAA22516
	for <tsvwg@ietf.org>; Sun, 3 Feb 2002 02:28:56 -0500 (EST)
Received: (cpmta 16619 invoked from network); 2 Feb 2002 23:28:26 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.218) with SMTP; 2 Feb 2002 23:28:26 -0800
X-Sent: 3 Feb 2002 07:28:26 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <tsvwg@ietf.org>
Subject: RE: [Tsvwg] a proposal for a different No-Data-Touch framer for TCP
Date: Sat, 2 Feb 2002 23:30:03 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJGEGCCPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Importance: Normal
In-Reply-To: <OF52045508.F035F58E-ONC2256B54.003384B0@telaviv.ibm.com>
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Julian,

This NDT technique closely resembles a header used for SCTP with the
exception of shared secrets that are not masked if the digest is CRC as
suggested.  SCTP has already established negotiations for a verification
tag, in addition, SCTP provides a standard means of bundling together with
assured frame alignment.  SCTP further solves problems associated with DoS,
spoofing, head of queue blocking, increased SACK, recovery, and
multi-homing.  This NDT technique used as a means of framing TCP will
require a portion of the application to be below the transport and thus
corrupt network layering.  SCTP can accomplish the same feature without
placing the application below the transport though use of the ability within
SCTP to send objects unordered within the transport allowing such objects to
appear as received within a shim.

I strongly recommend SCTP be considered as a solution for this feature
rather than disrupting existing TCP stacks to incorporate these changes for
framing.  It has become apparent there is a desire to keep a framing
interface for TCP proprietary whereas SCTP remains open.  Placing the
application below the transport would also represent a potential for
security problems depending on how the application communicates with this
new layer below TCP.  I would hope that just header structures are not seen
as complete solutions.  How these structures are used within the total
protocol must be included for there to be a coherent review.

Doug

Dear Colleagues:

I've recently submitted a draft including a different "No-Data-Touch"(NDT)
framer.
Unlike the current proposed NDT framer that has a "mistake" probability that
increases asymptotically (to 1) with the stream length the framer I am
suggesting has a very low and constant mistake probability and is thus safe
to use even in the absence of alignment or in the presence of
re-segmentation.

It does not contain the standard RFC2026 reference not as an indication of
patent - I have no  patent on it and I do not know about the existence of
one (nor did I search thoroughly) - but rather as a sign of discomfort for
it to be used without me being asked about it.

Until it apperas you can find it also at:

http://www.haifa.il.ibm.com/satran/ips/draft-satran-NDT-Framing-00.pdf

http://www.haifa.il.ibm.com/satran/ips/draft-satran-NDT-Framing-00.txt

Julo


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Feb  4 16:43:27 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA06523
	for <tsvwg-archive@odin.ietf.org>; Mon, 4 Feb 2002 16:43:27 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA02207
	for tsvwg-archive@odin.ietf.org; Mon, 4 Feb 2002 16:43:29 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA00008;
	Mon, 4 Feb 2002 15:54:40 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA29979
	for <tsvwg@ns.ietf.org>; Mon, 4 Feb 2002 15:54:38 -0500 (EST)
Received: from c003.snv.cp.net (c003-h016.c003.snv.cp.net [209.228.32.230])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id PAA05747
	for <tsvwg@ietf.org>; Mon, 4 Feb 2002 15:54:34 -0500 (EST)
Received: (cpmta 17570 invoked from network); 4 Feb 2002 12:54:05 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.230) with SMTP; 4 Feb 2002 12:54:05 -0800
X-Sent: 4 Feb 2002 20:54:05 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: "Tsvwg" <tsvwg@ietf.org>
Cc: "Ips" <ips@ece.cmu.edu>
Date: Mon, 4 Feb 2002 12:55:47 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJMEGGCPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
In-Reply-To: <277DD60FB639D511AC0400B0D068B71E02937413@CORPMX14>
Importance: Normal
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] RE: iSCSI: No Framing
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

David,

> > Those advocating a portion of the iSCSI application be in hardware
> > below the TCP transport should be willing to document information
> > exchanged between layers, and the states held within this layer.
> > You wish the IETF to endorse a scheme within a standards draft and
> > yet not be allowed to understand its implementation.
>
> That is a level of implementation detail that the IETF has historically
> not demanded the exposure of, and it would certainly be inappropriate to
> standardize.  To the extent that there are open source implementations of
> this form, the information will be disclosed in due course, other
> implementers will make their own decisions about what to make available.

I am not requesting a detailed API, but rather essential information
exchanged between this layer beneath TCP supporting framing and the
application.  An essential aspect of any application interchange is the
information exchanged with network related layers.  Keeping this information
proprietary does not allow a review.  I suspect the reason for keeping this
information private is to prevent an understanding of the scope of change
being advocated or to ensure proprietary solutions.

> > This scheme so modifies the application attempting FIM utilization, it
> > would be impractical to describe this as not altering TCP.  TCP is not
> > just a wire specification.  Changing the receive packet into a memory
> > array needs additional clarification beyond a claim this does not
> > change TCP.
>
> That is unfortunately a waste of list bandwidth.  With my WG chair hat
> firmly on, it is my conclusion that Doug Otis has failed to provide
> sufficient technical support for his argument that FIM alters TCP, and
> hence I hereby reject the procedural request that FIM be removed from the
> iSCSI draft as being a modification of TCP that would be outside the IPS
> WG's charter. Discussion of the merits of FIM is fine, and FIM may yet be
> removed from the iSCSI draft for good technical reasons (or may not) but
> debating whether FIM alters TCP is a waste of list bandwidth at this
point.

Any scheme that introduces a new network layer that directly or indirectly
interfaces with the application would be a modification of any transport.  I
do not feel I am obligated to offer technical details of this scheme, but
rather those that advocate its use should feel such obligations.

> I also state that it is the rough consensus of the IPS WG that TCP will be
> used for the first version of iSCSI, and this is stated over Doug Otis's
> continuing objections.  Doug would be better advised to work on using SCTP
> for iSCSI rather than complaining about the fact that others are not doing
> so.

I have had no objections to the use of TCP.  I have steadfastly rejected
alteration of TCP to accommodate various schemes designed to introduce
framing within or below TCP however, especially those that introduce
application specific layers below the transport.  My principle objection is
that such a scheme is possible using SCTP without alteration of network
layering.  Your continued endorsement of such a scheme is troubling.  If one
is allowed to create a new layer below and above the transport, then any
alteration of the transport becomes possible.  Granting this abuse in
principle makes any attempt by the IETF to regulate a transport meaningless.
Here you are advocating that this new layer does not require any details
offered and yet you have concluded it does not alter TCP in the absence of
public information.

> These decisions are appealable in accordance with RFC 2026, but should not
> be further discussed on this list.

My objections to this framing scheme remains unchanged.  Your endorsement of
an undocumented scheme is in violation of several principles expressed
within RFC 2026.  One being clear and concise documentation, and the other
openness and fairness.  The IPS wg has made an incorrect technical choice
through the addition of this new layer below TCP which places the integrity
of the iSCSI draft in jeopardy.  I am willing to continue this through an
appeals process, but my hope would be that this aspect of the draft is
struck prior to that exercise.

Doug

> Thanks,
> --David
>
> ---------------------------------------------------
> David L. Black, Senior Technologist
> EMC Corporation, 42 South St., Hopkinton, MA  01748
> +1 (508) 249-6449 *NEW*      FAX: +1 (508) 497-8500
> black_david@emc.com         Cell: +1 (978) 394-7754
> ---------------------------------------------------
>


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 01:43:29 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA15811
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 01:43:28 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id BAA07263
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 01:43:30 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA06021;
	Tue, 5 Feb 2002 01:30:40 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA05982
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 01:30:38 -0500 (EST)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA15509
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 01:30:35 -0500 (EST)
Received: from netlab.hcltech.com (arun [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id MAA18347
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 12:22:18 +0530
Message-ID: <3C5F7C39.B86F4150@netlab.hcltech.com>
Date: Tue, 05 Feb 2002 12:01:21 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Tsvwg <tsvwg@ietf.org>
References: <NEBBJGDMMLHHCIKHGBEJMEGGCPAA.dotis@sanlight.net>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] sctp Window deadlock
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi all,
    The question might be more of  SCTP implementation specific....   According
to the SCTP rfc2960, the data sender should not send any data packets if it doesnt
have enough Send-Windown (Peer Rwnd).... Just take the below example....

The Peer's Rwnd (at the Init time ) is 1500 bytes.
The Sender receivies a data buffer of size 6000 from its upper layer to be sent
to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of
size 1500 bytes (because the local mtu size is 1500bytes)....
Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the
Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with
rwnd as 0. The Sender which would have queued the frags B,C,D... will then
send frag B as the probing data to the Peer... The peer again will respond with
a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the
    * Sender cant send any more data packets as the Peer's rwnd is zero....
    * Peer cant increase its rwnd till it releases the data chunks which it had already
       received,  but to release that it needs to reassemble the frags (ABCD) and
      deliver it to the above layer, only then it can release the buffers..... Since its
      waiting for the frags C and D...  this will never release the buffers.... so
      will not increase the rwnd.....


This situation can be avoided if the Sender doesnt sent Data messages of size
greater than the Peer's rwnd.... But the Rfc doesnt say anything about this.....
Or there should be some timers, to get out of the Deadlock situation.....

Any suggestions........

Thanks
-arun


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 02:13:31 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA24478
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 02:13:31 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id CAA08868
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 02:13:32 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA07554;
	Tue, 5 Feb 2002 01:51:07 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA07525
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 01:51:05 -0500 (EST)
Received: from wiproecmx1.wipro.com (wiproecmx1.wipro.com [164.164.31.5])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA15901
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 01:51:00 -0500 (EST)
Received: from ecvwall1.wipro.com (ecvwall1.wipro.com [164.164.23.6])
	by wiproecmx1.wipro.com (8.11.3/8.11.3) with SMTP id g156nTX20553
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 12:19:30 +0530 (IST)
Received: from wipro.com ([192.168.178.83]) by
          ecmail.mail.wipro.com (Netscape Messaging Server 4.15) with
          ESMTP id GR1SBB01.Q19; Tue, 5 Feb 2002 12:19:59 +0530 
Message-ID: <3C5F8128.CABB3467@wipro.com>
Date: Tue, 05 Feb 2002 12:22:24 +0530
From: Ashraf <mohammed.asharaf@wipro.com>
Organization: Wipro
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.4.2-2 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Arun Prasad <arun@netlab.hcltech.com>
CC: Tsvwg <tsvwg@ietf.org>, Randall Stewart <randall@stewart.chicago.il.us>
Subject: Re: [Tsvwg] sctp Window deadlock
References: <NEBBJGDMMLHHCIKHGBEJMEGGCPAA.dotis@sanlight.net> <3C5F7C39.B86F4150@netlab.hcltech.com>
Content-Type: multipart/mixed;
	boundary="----=_NextPartTM-000-451942c5-1a03-11d6-a942-00b0d0d06be8"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


This is a multi-part message in MIME format.

------=_NextPartTM-000-451942c5-1a03-11d6-a942-00b0d0d06be8
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Arun,
    This is what RFC has to say :
"If the data receiver  run out of buffer space while still  waiting for  for more fragments
to complete the re-assembly, of the message, it should dispatch part of its inbound message
through a partial deliver API, freeing some if its buffer space in order receive  rest of
the message"

guess this should take care of the deadlock  situation you are mentioning here.

Thanks,
Mohd. Ashraf
Arun Prasad wrote:

> Hi all,
>     The question might be more of  SCTP implementation specific....   According
> to the SCTP rfc2960, the data sender should not send any data packets if it doesnt
> have enough Send-Windown (Peer Rwnd).... Just take the below example....
>
> The Peer's Rwnd (at the Init time ) is 1500 bytes.
> The Sender receivies a data buffer of size 6000 from its upper layer to be sent
> to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of
> size 1500 bytes (because the local mtu size is 1500bytes)....
> Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the
> Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with
> rwnd as 0. The Sender which would have queued the frags B,C,D... will then
> send frag B as the probing data to the Peer... The peer again will respond with
> a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the
>     * Sender cant send any more data packets as the Peer's rwnd is zero....
>     * Peer cant increase its rwnd till it releases the data chunks which it had already
>        received,  but to release that it needs to reassemble the frags (ABCD) and
>       deliver it to the above layer, only then it can release the buffers..... Since its
>       waiting for the frags C and D...  this will never release the buffers.... so
>       will not increase the rwnd.....
>
> This situation can be avoided if the Sender doesnt sent Data messages of size
> greater than the Peer's rwnd.... But the Rfc doesnt say anything about this.....
> Or there should be some timers, to get out of the Deadlock situation.....
>
> Any suggestions........
>
> Thanks
> -arun
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg


------=_NextPartTM-000-451942c5-1a03-11d6-a942-00b0d0d06be8
Content-Type: text/plain;
	name="Wipro_Disclaimer.txt"
Content-Disposition: attachment;
	filename="Wipro_Disclaimer.txt"
Content-Transfer-Encoding: 7bit

**************************Disclaimer************************************
      


Information contained in this E-MAIL being proprietary to Wipro Limited
is 'privileged' and 'confidential' and intended for use only by the
individual or entity to which it is addressed. You are notified that any
use, copying or dissemination of the information contained in the E-MAIL
in any manner whatsoever is strictly prohibited.



 ********************************************************************

------=_NextPartTM-000-451942c5-1a03-11d6-a942-00b0d0d06be8--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 03:41:57 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA25353
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 03:41:57 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id DAA13055
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 03:42:00 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA11861;
	Tue, 5 Feb 2002 03:19:34 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA11830
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 03:19:32 -0500 (EST)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA25175
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 03:19:28 -0500 (EST)
Received: from netlab.hcltech.com (arun [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id OAA22381;
	Tue, 5 Feb 2002 14:11:07 +0530
Message-ID: <3C5F95BA.923DEE3F@netlab.hcltech.com>
Date: Tue, 05 Feb 2002 13:50:10 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Ashraf <mohammed.asharaf@wipro.com>
CC: Tsvwg <tsvwg@ietf.org>, Randall Stewart <randall@stewart.chicago.il.us>
Subject: Re: [Tsvwg] sctp Window deadlock
References: <NEBBJGDMMLHHCIKHGBEJMEGGCPAA.dotis@sanlight.net> <3C5F7C39.B86F4150@netlab.hcltech.com> <3C5F8128.CABB3467@wipro.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi Ashraf,
    Thanks for the rsp....
ya, I agree that partial-delivery will solve the problem... But imagine an errornous
stack ( or SCTP stack which didnt implement the partial delivery API)  is on the
receiving side.... Then with the below example scenario, the Sender stack will
get stuck.... ie.,  the association will never tear down (as RTX timer is not running)
so the Sender side will be hung (with no mistake on its side....) forever.....
I think this situation should be avoided......

Thanks
-arun

Ashraf wrote:

> Arun,
>     This is what RFC has to say :
> "If the data receiver  run out of buffer space while still  waiting for  for more fragments
> to complete the re-assembly, of the message, it should dispatch part of its inbound message
> through a partial deliver API, freeing some if its buffer space in order receive  rest of
> the message"
>
> guess this should take care of the deadlock  situation you are mentioning here.
>
> Thanks,
> Mohd. Ashraf
> Arun Prasad wrote:
>
> > Hi all,
> >     The question might be more of  SCTP implementation specific....   According
> > to the SCTP rfc2960, the data sender should not send any data packets if it doesnt
> > have enough Send-Windown (Peer Rwnd).... Just take the below example....
> >
> > The Peer's Rwnd (at the Init time ) is 1500 bytes.
> > The Sender receivies a data buffer of size 6000 from its upper layer to be sent
> > to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of
> > size 1500 bytes (because the local mtu size is 1500bytes)....
> > Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the
> > Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with
> > rwnd as 0. The Sender which would have queued the frags B,C,D... will then
> > send frag B as the probing data to the Peer... The peer again will respond with
> > a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the
> >     * Sender cant send any more data packets as the Peer's rwnd is zero....
> >     * Peer cant increase its rwnd till it releases the data chunks which it had already
> >        received,  but to release that it needs to reassemble the frags (ABCD) and
> >       deliver it to the above layer, only then it can release the buffers..... Since its
> >       waiting for the frags C and D...  this will never release the buffers.... so
> >       will not increase the rwnd.....
> >
> > This situation can be avoided if the Sender doesnt sent Data messages of size
> > greater than the Peer's rwnd.... But the Rfc doesnt say anything about this.....
> > Or there should be some timers, to get out of the Deadlock situation.....
> >
> > Any suggestions........
> >
> > Thanks
> > -arun
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
>
>   ------------------------------------------------------------------------------------------------------------------------
>                            Name: Wipro_Disclaimer.txt
>    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
>                        Encoding: 7bit

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 03:49:59 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA25413
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 03:49:58 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id DAA13347
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 03:50:00 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA13032;
	Tue, 5 Feb 2002 03:41:12 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA13001
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 03:41:09 -0500 (EST)
Received: from c003.snv.cp.net (c003-h013.c003.snv.cp.net [209.228.32.227])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id DAA25346
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 03:41:02 -0500 (EST)
Received: (cpmta 14315 invoked from network); 5 Feb 2002 00:40:33 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.227) with SMTP; 5 Feb 2002 00:40:33 -0800
X-Sent: 5 Feb 2002 08:40:33 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <ips@ece.cmu.edu>
Cc: "Tsvwg" <tsvwg@ietf.org>
Date: Tue, 5 Feb 2002 00:42:18 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJIEGKCPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
In-Reply-To: <OF03A89157.17F716DB-ONC2256B57.0006539A@telaviv.ibm.com>
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] RE: iSCSI: No Framing
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Julian,

Implementing a scheme for doing direct data placement by means of an
application specific, undocumented or proprietary layer below TCP will
fracture the market.  This will result in an unacceptable outcome having
hardware created for each application, limited by various vendors making as
yet unknown claims, if implementing the DDP feature using TCP.  In the
future, should this feature become desired, iSCSI can adopt a generic method
that is possible through the use of SCTP.  An eventual adoption of SCTP
would also simplify much of the existing complexity found with the
multi-connection TCP standard now proposed for iSCSI.  Use of SCTP would
allow common hardware for many protocols desiring Direct Data Placement
without modification of the SCTP transport as it can deliver objects
unordered.  A shim providing the DDP feature could ensure objects are
disclosed in the order sent if desired, independent of the reception at the
shim while not modifying SCTP or adding layers beneath this transport as
would be required for TCP.

Only when constrained to conventional memory allocation, does DDP become
beneficial.  The concept of placing a layer beneath TCP that structures an
array to hold each individual packet based on application specific
structures is fundamentally flawed.  This presumes the layer beneath TCP
knows before hand the content of the byte stream prior to TCP composing the
stream.  This lower layer will also need to create exceptions for handling
duplicates, for placement information not aligned with the TCP segment, and
for application informational errors.  SCTP can easily avoid these
exceptional cases making such development less disruptive.  In addition, an
SCTP stack that does not offer DDP in hardware need not change the API
behavior to the application nor modify any layer functionality.

The ability of SCTP to support many different higher level protocols comes
from the capability of SCTP to deliver objects unordered.  TCP's restriction
on ordered delivery mandates that any direct data placement is done prior to
TCP through the addition of a new and application unique layer between IP
and TCP.  The saving caveat has been that disclosure to TCP of these packets
are sequential but the application interface to TCP will change as a result
of this modification in some undefined way to associate the steering
information with the desired buffers.  This change impacts the interface
between this new layer, as well as the application and TCP.  The desire to
keep this DDP scheme for TCP proprietary will ensure a fractured market in
terms of hardware, OS services, and perhaps interchange not to mention that
each application then also needs unique hardware.  The security risks
involved in creating this new and highly complex layer below TCP is yet
unknown as details are lacking.  In essence, an attempt to implement DDP by
creating a new layer beneath TCP in a proprietary fashion is in direct
competition with the far better practice of using SCTP to implement DDP.

Doug

> Mallikarjun,
>
> I doubt that FIM (or COWS) will fracture the market.
> Hardware and software vendors will gain experience in what it takes to use
> framing.
> a specialized DDP and that could be useful later.  The first generation
> although not imperiously needing any framing (I have proposed not
> less than
> 3 solutions!) will enable us to get a better second generation if we do
> something in this area.
>
> Julo
>
> I share the concern about iSCSI embracing a framing mechanism that is
> not a MUST implement.  For all the reasons that Jim pointed out, OTOH,
> I am not recommending a MUST framing either.  I suspect not many will
> implement framing if it's a MAY, so it appears that we're talking about
> a potential SHOULD.
>
> Given that SHOULD is a fairly strong requirement, one "significant
> justification"
> for not doing framing could be (even while the NIC *may be* on the
> expensive side) -
>  - has less design complexity since no OOO placement.
>  - can have quicker TTM since less design, testing, debugging etc.
>
> I think ultimately it boils down to: how many vendors would use a
> "significant
> justfication" to not implement a SHOULD-requirement?
>
> If that's a majority: let's say vendor X is implementing
> (whatever) framing
> for
> optimizing the memory requirements, it essentially means that X's product
> will
> perform poorly with (the majority) no-framing senders.  I don't think X
> would
> like that, nor the customer.  IOW, this situation doesn't seem to be
> long-lasting.
> OTOH, the situation of an almost equal number of  "framing" and
> "no-framing"
> products in the market (perhaps at different price points) could be
> unfortunately
> long-lasting....
>
> To summarize, it is a troubling prospect that a framing technique (if
> adopted as
> SHOULD) has the potential to somewhat fracture the market and in effect
> create "interoperability problems" (of performance sort) similar to that
> affect
> FC....
> --
> Mallikarjun
>
> Mallikarjun Chadalapaka
> Networked Storage Architecture
> Network Storage Solutions Organization
> Hewlett-Packard MS 5668
> Roseville CA 95747


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 04:59:35 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA26158
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 04:59:35 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id EAA16509
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 04:59:38 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA15904;
	Tue, 5 Feb 2002 04:40:39 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA15875
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 04:40:34 -0500 (EST)
Received: from india.mercurykr.com ([165.133.13.60])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA25877
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 04:38:37 -0500 (EST)
Received: from vijay (Archana [165.133.13.38])
	by india.mercurykr.com (8.8.8+Sun/8.8.8) with SMTP id OAA02755
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 14:58:39 -0600 (GMT)
MIME-Version: 1.0
Message-Id: <3C5FA7A4.000005.96873@vijay.india.mercurykr.com>
Date: Tue, 5 Feb 2002 15:06:36 +0530 (India Standard Time)
Content-Type: Text/Plain
X-Mailer: IncrediMail 2001 (1500245)
From: "Vijay Kumar Choudhary" <vijay@mercurykr.com>
X-Priority: 3
X-FID: FLAVOR00-NONE-0000-0000-000000000000
To: <tsvwg@ietf.org>
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id EAA15877
Subject: [Tsvwg] First DATA chunk
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi All,
  In SACK chunk, what will be the value of the cumulative TSN Ack if the
first DATA chunk is missed?

Or,

If the first DATA chunk is missed, does the SACK needs to be send at all?
Regards,
Vijay

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 05:08:40 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA26284
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 05:08:40 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA17309
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 05:08:42 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA16299;
	Tue, 5 Feb 2002 04:53:10 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA16272
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 04:53:08 -0500 (EST)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA26089
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 04:53:04 -0500 (EST)
Received: from netlab.hcltech.com (arun [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id PAA25686;
	Tue, 5 Feb 2002 15:44:12 +0530
Message-ID: <3C5FAB8B.566704F8@netlab.hcltech.com>
Date: Tue, 05 Feb 2002 15:23:15 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Vijay Kumar Choudhary <vijay@mercurykr.com>
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] First DATA chunk
References: <3C5FA7A4.000005.96873@vijay.india.mercurykr.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi Vijay,
there is no need to send SACK if u dont receive DATA packets at all....... ur
second point is right..

Thanks
-arun

Vijay Kumar Choudhary wrote:

> Hi All,
>   In SACK chunk, what will be the value of the cumulative TSN Ack if the
> first DATA chunk is missed?
>
> Or,
>
> If the first DATA chunk is missed, does the SACK needs to be send at all?
> Regards,
> Vijay
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 06:03:49 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA26960
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 06:03:49 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA20090
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 06:03:52 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA18647;
	Tue, 5 Feb 2002 05:39:38 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA18617
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 05:39:35 -0500 (EST)
Received: from india.mercurykr.com ([165.133.13.60])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA26726
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 05:39:10 -0500 (EST)
Received: from vijay (Archana [165.133.13.38])
	by india.mercurykr.com (8.8.8+Sun/8.8.8) with SMTP id PAA03592;
	Tue, 5 Feb 2002 15:54:25 -0600 (GMT)
MIME-Version: 1.0
Message-Id: <3C5FB4B6.00000B.96873@vijay.india.mercurykr.com>
Date: Tue, 5 Feb 2002 16:02:22 +0530 (India Standard Time)
Content-Type: Text/Plain
X-Mailer: IncrediMail 2001 (1500245)
From: "Vijay Kumar Choudhary" <vijay@mercurykr.com>
References: <3C5FA7A4.000005.96873@vijay.india.mercurykr.com> <3C5FAB8B.566704F8@netlab.hcltech.com>
X-Priority: 3
X-FID: FLAVOR00-NONE-0000-0000-000000000000
To: <arun@netlab.hcltech.com>
Cc: <tsvwg@ietf.org>
Subject: Re: [Tsvwg] First DATA chunk
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id FAA18618
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi Arun,

    U didnt get my point.Suppose i have received the second DATA chunk
without receiving the first DATA chunk.Does the SACK needs to be send.If YES
 then what will be the value of the cumulative TSN Ack in SACK chunk.If NO,
what will be the behaviour of SCTP. 

Rehards,
Vijay
-------Original Message-------
From: Arun Prasad
Date: Wednesday, February 06, 2002 03:01:45 AM
To: Vijay Kumar Choudhary
Cc: tsvwg@ietf.org
Subject: Re: [Tsvwg] First DATA chunk
Hi Vijay,
there is no need to send SACK if u dont receive DATA packets at all.......
ur
second point is right..
Thanks
-arun
Vijay Kumar Choudhary wrote:
> Hi All,
> In SACK chunk, what will be the value of the cumulative TSN Ack if the
> first DATA chunk is missed?
>
> Or,
>
> If the first DATA chunk is missed, does the SACK needs to be send at all?
> Regards,
> Vijay
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.
Contact # : 9144 - 2334174
9144 - 2334181
extn : 233
****************************************************************
_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 06:46:27 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA27650
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 06:46:27 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA22659
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 06:46:28 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA21311;
	Tue, 5 Feb 2002 06:23:18 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA21237
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 06:23:15 -0500 (EST)
Received: from wiproecmx1.wipro.com (wiproecmx1.wipro.com [164.164.31.5])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA27312
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 06:23:09 -0500 (EST)
Received: from ecvwall1.wipro.com (ecvwall1.wipro.com [164.164.23.6])
	by wiproecmx1.wipro.com (8.11.3/8.11.3) with SMTP id g15BLJX15408
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 16:51:20 +0530 (IST)
Received: from wipro.com ([192.168.178.83]) by
          ecmail.mail.wipro.com (Netscape Messaging Server 4.15) with
          ESMTP id GR24W401.PAN; Tue, 5 Feb 2002 16:51:40 +0530 
Message-ID: <3C5FC0DF.9E9EBF7A@wipro.com>
Date: Tue, 05 Feb 2002 16:54:15 +0530
From: Ashraf <mohammed.asharaf@wipro.com>
Organization: Wipro
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.4.2-2 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Arun Prasad <arun@netlab.hcltech.com>
CC: Tsvwg <tsvwg@ietf.org>, Randall Stewart <randall@stewart.chicago.il.us>
Subject: Re: [Tsvwg] sctp Window deadlock
References: <NEBBJGDMMLHHCIKHGBEJMEGGCPAA.dotis@sanlight.net> <3C5F7C39.B86F4150@netlab.hcltech.com> <3C5F8128.CABB3467@wipro.com> <3C5F95BA.923DEE3F@netlab.hcltech.com>
Content-Type: multipart/mixed;
	boundary="----=_NextPartTM-000-0cee78dc-1a29-11d6-a942-00b0d0d06be8"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


This is a multi-part message in MIME format.

------=_NextPartTM-000-0cee78dc-1a29-11d6-a942-00b0d0d06be8
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hi arun,
    I think  for the new data chunk ( which we are sending  to probe for change in window), we start RTT timer and  give
appropraite treatment, when it expires.

thanks,
ashraf

Arun Prasad wrote:

> Hi Ashraf,
>     Thanks for the rsp....
> ya, I agree that partial-delivery will solve the problem... But imagine an errornous
> stack ( or SCTP stack which didnt implement the partial delivery API)  is on the
> receiving side.... Then with the below example scenario, the Sender stack will
> get stuck.... ie.,  the association will never tear down (as RTX timer is not running)
> so the Sender side will be hung (with no mistake on its side....) forever.....
> I think this situation should be avoided......
>
> Thanks
> -arun
>
> Ashraf wrote:
>
> > Arun,
> >     This is what RFC has to say :
> > "If the data receiver  run out of buffer space while still  waiting for  for more fragments
> > to complete the re-assembly, of the message, it should dispatch part of its inbound message
> > through a partial deliver API, freeing some if its buffer space in order receive  rest of
> > the message"
> >
> > guess this should take care of the deadlock  situation you are mentioning here.
> >
> > Thanks,
> > Mohd. Ashraf
> > Arun Prasad wrote:
> >
> > > Hi all,
> > >     The question might be more of  SCTP implementation specific....   According
> > > to the SCTP rfc2960, the data sender should not send any data packets if it doesnt
> > > have enough Send-Windown (Peer Rwnd).... Just take the below example....
> > >
> > > The Peer's Rwnd (at the Init time ) is 1500 bytes.
> > > The Sender receivies a data buffer of size 6000 from its upper layer to be sent
> > > to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of
> > > size 1500 bytes (because the local mtu size is 1500bytes)....
> > > Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the
> > > Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with
> > > rwnd as 0. The Sender which would have queued the frags B,C,D... will then
> > > send frag B as the probing data to the Peer... The peer again will respond with
> > > a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the
> > >     * Sender cant send any more data packets as the Peer's rwnd is zero....
> > >     * Peer cant increase its rwnd till it releases the data chunks which it had already
> > >        received,  but to release that it needs to reassemble the frags (ABCD) and
> > >       deliver it to the above layer, only then it can release the buffers..... Since its
> > >       waiting for the frags C and D...  this will never release the buffers.... so
> > >       will not increase the rwnd.....
> > >
> > > This situation can be avoided if the Sender doesnt sent Data messages of size
> > > greater than the Peer's rwnd.... But the Rfc doesnt say anything about this.....
> > > Or there should be some timers, to get out of the Deadlock situation.....
> > >
> > > Any suggestions........
> > >
> > > Thanks
> > > -arun
> > >
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> >
> >   ------------------------------------------------------------------------------------------------------------------------
> >                            Name: Wipro_Disclaimer.txt
> >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> >                        Encoding: 7bit
>
> --
> ****************************************************************
> V.Arun Prasad
> HCL Technologies Ltd.
> 51, Jawaharlal Nehru Road,
> Ekkattuthangal,
> Guindy Industrial Estate,
> Chennai - 600097.
>
> Contact # : 9144 - 2334174
>             9144 - 2334181
>             extn : 233
> ****************************************************************
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg


------=_NextPartTM-000-0cee78dc-1a29-11d6-a942-00b0d0d06be8
Content-Type: text/plain;
	name="Wipro_Disclaimer.txt"
Content-Disposition: attachment;
	filename="Wipro_Disclaimer.txt"
Content-Transfer-Encoding: 7bit

**************************Disclaimer************************************
      


Information contained in this E-MAIL being proprietary to Wipro Limited
is 'privileged' and 'confidential' and intended for use only by the
individual or entity to which it is addressed. You are notified that any
use, copying or dissemination of the information contained in the E-MAIL
in any manner whatsoever is strictly prohibited.



 ********************************************************************

------=_NextPartTM-000-0cee78dc-1a29-11d6-a942-00b0d0d06be8--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 06:47:58 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA27702
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 06:47:58 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA22769
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 06:48:00 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA22261;
	Tue, 5 Feb 2002 06:36:54 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA22231
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 06:36:52 -0500 (EST)
Received: from sj-msg-core-1.cisco.com (sj-msg-core-1.cisco.com [171.71.163.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA27473
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 06:36:49 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-1.cisco.com (8.11.3/8.9.1) with ESMTP id g15BaHg08500;
	Tue, 5 Feb 2002 03:36:17 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id AAZ80159;
	Tue, 5 Feb 2002 03:36:16 -0800 (PST)
Message-ID: <3C5FC3AF.6169151B@stewart.chicago.il.us>
Date: Tue, 05 Feb 2002 05:36:15 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Vijay Kumar Choudhary <vijay@mercurykr.com>
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] First DATA chunk
References: <3C5FA7A4.000005.96873@vijay.india.mercurykr.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Vijay:

If two packets are sent Chunk 101 and Chunk 102. And the
initial TSN IS 101... and then somewhere in the network
101 is dropped...

You would end up sending a SACK with a CUM-ACK of 100 and
gap report of 1 showing [2,2].

The RFC is specific about that and tells each side to set
there cum-ack point to Initial-TSN minus 1.

R

Vijay Kumar Choudhary wrote:
> 
> Hi All,
>   In SACK chunk, what will be the value of the cumulative TSN Ack if the
> first DATA chunk is missed?
> 
> Or,
> 
> If the first DATA chunk is missed, does the SACK needs to be send at all?
> Regards,
> Vijay
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 06:51:08 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA27783
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 06:51:07 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA22849
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 06:51:09 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA20904;
	Tue, 5 Feb 2002 06:16:25 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA20867
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 06:16:22 -0500 (EST)
Received: from hindon.hss.co.in ([202.54.26.202])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA27212
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 06:16:17 -0500 (EST)
From: vagarg@hss.hns.com
Received: from sandesh.hss.hns.com (sandesh [139.85.242.35])
	by hindon.hss.co.in (8.10.0/8.10.0) with SMTP id g15BCqA18750;
	Tue, 5 Feb 2002 16:42:52 +0530 (IST)
Received: by sandesh.hss.hns.com(Lotus SMTP MTA v4.6.3  (733.2 10-16-1998))  id 65256B57.003D6E00 ; Tue, 5 Feb 2002 16:40:58 +0530
X-Lotus-FromDomain: HSS
To: "Vijay Kumar Choudhary" <vijay@mercurykr.com>
cc: arun@netlab.hcltech.com, tsvwg@ietf.org
Message-ID: <65256B57.003D6CB2.00@sandesh.hss.hns.com>
Date: Tue, 5 Feb 2002 16:44:55 +0530
Subject: Re: [Tsvwg] First DATA chunk
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Content-Disposition: inline
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org



Hi Vijay

In this case you can send the value of cummulative TSN you received last.
SACK should be sent with gap ack reports.

Vakul
(www.hssworld.com)





"Vijay Kumar Choudhary" <vijay@mercurykr.com> on 02/05/2002 04:02:22 PM

To:   arun@netlab.hcltech.com
cc:   tsvwg@ietf.org (bcc: Vakul Garg/HSS)

Subject:  Re: [Tsvwg] First DATA chunk




Hi Arun,

    U didnt get my point.Suppose i have received the second DATA chunk
without receiving the first DATA chunk.Does the SACK needs to be send.If YES
 then what will be the value of the cumulative TSN Ack in SACK chunk.If NO,
what will be the behaviour of SCTP.

Rehards,
Vijay
-------Original Message-------
From: Arun Prasad
Date: Wednesday, February 06, 2002 03:01:45 AM
To: Vijay Kumar Choudhary
Cc: tsvwg@ietf.org
Subject: Re: [Tsvwg] First DATA chunk
Hi Vijay,
there is no need to send SACK if u dont receive DATA packets at all.......
ur
second point is right..
Thanks
-arun
Vijay Kumar Choudhary wrote:
> Hi All,
> In SACK chunk, what will be the value of the cumulative TSN Ack if the
> first DATA chunk is missed?
>
> Or,
>
> If the first DATA chunk is missed, does the SACK needs to be send at all?
> Regards,
> Vijay
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.
Contact # : 9144 - 2334174
9144 - 2334181
extn : 233
****************************************************************
_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg





_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 07:06:02 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA27888
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 07:06:02 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA23805
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 07:06:04 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA22939;
	Tue, 5 Feb 2002 06:52:12 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA22910
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 06:52:11 -0500 (EST)
Received: from wiproecmx1.wipro.com (wiproecmx1.wipro.com [164.164.31.5])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA27816
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 06:51:58 -0500 (EST)
Received: from ecvwall1.wipro.com (ecvwall1.wipro.com [164.164.23.6])
	by wiproecmx1.wipro.com (8.11.3/8.11.3) with SMTP id g15BoTX23829
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 17:20:29 +0530 (IST)
Received: from wipro.com ([192.168.178.83]) by
          ecmail.mail.wipro.com (Netscape Messaging Server 4.15) with
          ESMTP id GR268Q00.DAB for <tsvwg@ietf.org>; Tue, 5 Feb 2002
          17:20:50 +0530 
Message-ID: <3C5FC7B6.6B40176E@wipro.com>
Date: Tue, 05 Feb 2002 17:23:26 +0530
From: Ashraf <mohammed.asharaf@wipro.com>
Organization: Wipro
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.4.2-2 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: tsvwg@ietf.org
Subject: [Fwd: [Tsvwg] First DATA chunk]
Content-Type: multipart/mixed;
 boundary="------------8E68602F5618E12395D2E22B"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

This is a multi-part message in MIME format.
--------------8E68602F5618E12395D2E22B
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit



--------------8E68602F5618E12395D2E22B
Content-Type: message/rfc822
Content-Disposition: inline

X-Mozilla-Status2: 00000000
Message-ID: <3C5FC15C.98024F73@wipro.com>
Date: Tue, 05 Feb 2002 16:56:20 +0530
From: Ashraf <mohammed.asharaf@wipro.com>
Organization: Wipro
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.4.2-2 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Vijay Kumar Choudhary <vijay@mercurykr.com>
Subject: Re: [Tsvwg] First DATA chunk
References: <3C5FA7A4.000005.96873@vijay.india.mercurykr.com> <3C5FAB8B.566704F8@netlab.hcltech.com> <3C5FB4B6.00000B.96873@vijay.india.mercurykr.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Shouldn't value of Cum TSN  in SACK here be ( INITIAL_TSN -1)?

thanks,
ashraf
Vijay Kumar Choudhary wrote:

> Hi Arun,
>
>     U didnt get my point.Suppose i have received the second DATA chunk
> without receiving the first DATA chunk.Does the SACK needs to be send.If YES
>  then what will be the value of the cumulative TSN Ack in SACK chunk.If NO,
> what will be the behaviour of SCTP.
>
> Rehards,
> Vijay
> -------Original Message-------
> From: Arun Prasad
> Date: Wednesday, February 06, 2002 03:01:45 AM
> To: Vijay Kumar Choudhary
> Cc: tsvwg@ietf.org
> Subject: Re: [Tsvwg] First DATA chunk
> Hi Vijay,
> there is no need to send SACK if u dont receive DATA packets at all.......
> ur
> second point is right..
> Thanks
> -arun
> Vijay Kumar Choudhary wrote:
> > Hi All,
> > In SACK chunk, what will be the value of the cumulative TSN Ack if the
> > first DATA chunk is missed?
> >
> > Or,
> >
> > If the first DATA chunk is missed, does the SACK needs to be send at all?
> > Regards,
> > Vijay
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> --
> ****************************************************************
> V.Arun Prasad
> HCL Technologies Ltd.
> 51, Jawaharlal Nehru Road,
> Ekkattuthangal,
> Guindy Industrial Estate,
> Chennai - 600097.
> Contact # : 9144 - 2334174
> 9144 - 2334181
> extn : 233
> ****************************************************************
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg



--------------8E68602F5618E12395D2E22B
Content-Type: text/plain;
	name="Wipro_Disclaimer.txt"
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment;
	filename="Wipro_Disclaimer.txt"
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit

**************************Disclaimer************************************
      


Information contained in this E-MAIL being proprietary to Wipro Limited
is 'privileged' and 'confidential' and intended for use only by the
individual or entity to which it is addressed. You are notified that any
use, copying or dissemination of the information contained in the E-MAIL
in any manner whatsoever is strictly prohibited.



 ********************************************************************

--------------8E68602F5618E12395D2E22B--


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 07:10:19 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA27951
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 07:10:18 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA23949
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 07:10:20 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA22160;
	Tue, 5 Feb 2002 06:34:23 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA22129
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 06:34:19 -0500 (EST)
Received: from sj-msg-core-1.cisco.com (sj-msg-core-1.cisco.com [171.71.163.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA27446
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 06:34:17 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-1.cisco.com (8.11.3/8.9.1) with ESMTP id g15BXhg07484;
	Tue, 5 Feb 2002 03:33:43 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id AAZ80146;
	Tue, 5 Feb 2002 03:33:41 -0800 (PST)
Message-ID: <3C5FC314.D15DF4F9@stewart.chicago.il.us>
Date: Tue, 05 Feb 2002 05:33:41 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Arun Prasad <arun@netlab.hcltech.com>
CC: Ashraf <mohammed.asharaf@wipro.com>, Tsvwg <tsvwg@ietf.org>
Subject: Re: [Tsvwg] sctp Window deadlock
References: <NEBBJGDMMLHHCIKHGBEJMEGGCPAA.dotis@sanlight.net> <3C5F7C39.B86F4150@netlab.hcltech.com> <3C5F8128.CABB3467@wipro.com> <3C5F95BA.923DEE3F@netlab.hcltech.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Arun:

Some comments:

Arun Prasad wrote:
> 
> Hi Ashraf,
>     Thanks for the rsp....
> ya, I agree that partial-delivery will solve the problem... But imagine an errornous
> stack ( or SCTP stack which didnt implement the partial delivery API)  is on the
> receiving side.... Then with the below example scenario, the Sender stack will
> get stuck.... ie.,  the association will never tear down (as RTX timer is not running)
> so the Sender side will be hung (with no mistake on its side....) forever.....
> I think this situation should be avoided......

I will address this comment below ...


> 
> Thanks
> -arun
> 
> Ashraf wrote:
> 
> > Arun,
> >     This is what RFC has to say :
> > "If the data receiver  run out of buffer space while still  waiting for  for more fragments
> > to complete the re-assembly, of the message, it should dispatch part of its inbound message
> > through a partial deliver API, freeing some if its buffer space in order receive  rest of
> > the message"
> >
> > guess this should take care of the deadlock  situation you are mentioning here.
> >
> > Thanks,
> > Mohd. Ashraf
> > Arun Prasad wrote:

The above is percisely what we designed the PD-API for. It is
a part I never really much liked :> but we saw it to be the
only way out of such a dead-lock.... Also teamed with this
is the receivers right to reneg. This way in a critical situation
it can drop some of the last chunks it is holding so it can
make room for the earlier pieces so it CAN use the PD-API...

Both are necessary to ensure no deadlock can occur..


Now as to the other issue.. what happens if your sender
is broken or an attacker.. a number of things come to mind
(as yet I have not had my coffee so there may be more :0)

1) TCP I would imagine can be in the same situation .. i.e.
   if it connects to a peer that is intentionally doing something
   bad.

2) It really becomes implementation dependant. I know of one
implementation
   that will only keep sending the same chunk for so long.. and then
   it will abort the associaition.

3) If the guy is really an attacker the problem is more in to the fact
   that you have entered into the association with him. I.e don't
   leave the front door open and expect a burgular not to enter...

4) If it is a broken implementation then it needs to get fixed and I
   would hope the one stuck would at least have enough logging going
   on to show that it is broken... (or even in the case 3 I think logs
   may help) so someone can fix it..


R

> >
> > > Hi all,
> > >     The question might be more of  SCTP implementation specific....   According
> > > to the SCTP rfc2960, the data sender should not send any data packets if it doesnt
> > > have enough Send-Windown (Peer Rwnd).... Just take the below example....
> > >
> > > The Peer's Rwnd (at the Init time ) is 1500 bytes.
> > > The Sender receivies a data buffer of size 6000 from its upper layer to be sent
> > > to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of
> > > size 1500 bytes (because the local mtu size is 1500bytes)....
> > > Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the
> > > Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with
> > > rwnd as 0. The Sender which would have queued the frags B,C,D... will then
> > > send frag B as the probing data to the Peer... The peer again will respond with
> > > a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the
> > >     * Sender cant send any more data packets as the Peer's rwnd is zero....
> > >     * Peer cant increase its rwnd till it releases the data chunks which it had already
> > >        received,  but to release that it needs to reassemble the frags (ABCD) and
> > >       deliver it to the above layer, only then it can release the buffers..... Since its
> > >       waiting for the frags C and D...  this will never release the buffers.... so
> > >       will not increase the rwnd.....
> > >
> > > This situation can be avoided if the Sender doesnt sent Data messages of size
> > > greater than the Peer's rwnd.... But the Rfc doesnt say anything about this.....
> > > Or there should be some timers, to get out of the Deadlock situation.....
> > >
> > > Any suggestions........
> > >
> > > Thanks
> > > -arun
> > >
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> >
> >   ------------------------------------------------------------------------------------------------------------------------
> >                            Name: Wipro_Disclaimer.txt
> >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> >                        Encoding: 7bit
> 
> --
> ****************************************************************
> V.Arun Prasad
> HCL Technologies Ltd.
> 51, Jawaharlal Nehru Road,
> Ekkattuthangal,
> Guindy Industrial Estate,
> Chennai - 600097.
> 
> Contact # : 9144 - 2334174
>             9144 - 2334181
>             extn : 233
> ****************************************************************

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 07:16:09 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA28004
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 07:16:08 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA24080
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 07:16:10 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA22303;
	Tue, 5 Feb 2002 06:37:19 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA22274
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 06:37:17 -0500 (EST)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA27475
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 06:37:14 -0500 (EST)
Received: from netlab.hcltech.com (arun [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id RAA29227;
	Tue, 5 Feb 2002 17:28:22 +0530
Message-ID: <3C5FC3F5.CFF09D3B@netlab.hcltech.com>
Date: Tue, 05 Feb 2002 17:07:25 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Vijay Kumar Choudhary <vijay@mercurykr.com>
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] First DATA chunk
References: <3C5FA7A4.000005.96873@vijay.india.mercurykr.com> <3C5FAB8B.566704F8@netlab.hcltech.com> <3C5FB4B6.00000B.96873@vijay.india.mercurykr.com>
Content-Type: multipart/alternative;
 boundary="------------58E197A42F4FE7DB32B97D49"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


--------------58E197A42F4FE7DB32B97D49
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hi Vijay,
    sorry, I got ur querry now....  I can think of one way to do it.... Just check
whether this is OK....

Let the
    Initial Cumulative TSN  = InitTsn -1 (Receiver side)
    Initial Cumulative TSN Ack Pt =  InitTsn - 1 (Sender side)

    Now consider ur condition, the SACK will be generated with the
Cum TSN Ack as "Initial Cum TSN", along with the gap blocks....
Upon receiving this SACK,  the endpoint recognises this to be a
valid SACK according to the below lines from RFC......
    So I think, there is no need to handle seperately for this condition,
but to have the initial values as specified above.....


i) If Cumulative TSN Ack is less than the Cumulative TSN Ack
         Point, then drop the SACK.   Since Cumulative TSN Ack is
         monotonically increasing, a SACK whose Cumulative TSN Ack is
         less than the Cumulative TSN Ack Point indicates an out-of-
         order SACK.

Is this suggestion ok???

Thanks
-arun


Vijay Kumar Choudhary wrote:

> Hi Arun,
>
>     U didnt get my point.Suppose i have received the second DATA chunk
> without receiving the first DATA chunk.Does the SACK needs to be send.If YES
>  then what will be the value of the cumulative TSN Ack in SACK chunk.If NO,
> what will be the behaviour of SCTP.
>
> Rehards,
> Vijay
> -------Original Message-------
> From: Arun Prasad
> Date: Wednesday, February 06, 2002 03:01:45 AM
> To: Vijay Kumar Choudhary
> Cc: tsvwg@ietf.org
> Subject: Re: [Tsvwg] First DATA chunk
> Hi Vijay,
> there is no need to send SACK if u dont receive DATA packets at all.......
> ur
> second point is right..
> Thanks
> -arun
> Vijay Kumar Choudhary wrote:
> > Hi All,
> > In SACK chunk, what will be the value of the cumulative TSN Ack if the
> > first DATA chunk is missed?
> >
> > Or,
> >
> > If the first DATA chunk is missed, does the SACK needs to be send at all?
> > Regards,
> > Vijay
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> --
> ****************************************************************
> V.Arun Prasad
> HCL Technologies Ltd.
> 51, Jawaharlal Nehru Road,
> Ekkattuthangal,
> Guindy Industrial Estate,
> Chennai - 600097.
> Contact # : 9144 - 2334174
> 9144 - 2334181
> extn : 233
> ****************************************************************
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************


--------------58E197A42F4FE7DB32B97D49
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
Hi Vijay,
<br>&nbsp;&nbsp;&nbsp; sorry, I got ur querry now....&nbsp; I can think
of one way to do it.... Just check
<br>whether this is OK....
<p>Let the
<br>&nbsp;&nbsp;&nbsp; Initial <font color="#000000">Cumulative TSN&nbsp;
= InitTsn -1 (Receiver side)</font>
<br><font color="#000000">&nbsp;&nbsp;&nbsp; Initial Cumulative TSN Ack
Pt =&nbsp; InitTsn - 1 (Sender side)</font>
<p>&nbsp;&nbsp;&nbsp; Now consider ur condition, the SACK will be generated
with the
<br>Cum TSN Ack as "Initial Cum TSN", along with the gap blocks....
<br>Upon receiving this SACK,&nbsp; the endpoint recognises this to be
a
<br>valid SACK according to the below lines from RFC......
<br>&nbsp;&nbsp;&nbsp; So I think, there is no need to handle seperately
for this condition,
<br>but to have the initial values as specified above.....
<p>&nbsp;
<br><font color="#000099">i) If Cumulative TSN Ack is less than the Cumulative
TSN Ack</font>
<br><font color="#000099">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
Point, then drop the SACK.&nbsp;&nbsp; Since Cumulative TSN Ack is</font>
<br><font color="#000099">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
monotonically increasing, a SACK whose Cumulative TSN Ack is</font>
<br><font color="#000099">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
less than the Cumulative TSN Ack Point indicates an out-of-</font>
<br><font color="#000099">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
order SACK.</font>
<p>Is this suggestion ok???
<p>Thanks
<br>-arun
<br>&nbsp;
<p>Vijay Kumar Choudhary wrote:
<blockquote TYPE=CITE>Hi Arun,
<p>&nbsp;&nbsp;&nbsp; U didnt get my point.Suppose i have received the
second DATA chunk
<br>without receiving the first DATA chunk.Does the SACK needs to be send.If
YES
<br>&nbsp;then what will be the value of the cumulative TSN Ack in SACK
chunk.If NO,
<br>what will be the behaviour of SCTP.
<p>Rehards,
<br>Vijay
<br>-------Original Message-------
<br>From: Arun Prasad
<br>Date: Wednesday, February 06, 2002 03:01:45 AM
<br>To: Vijay Kumar Choudhary
<br>Cc: tsvwg@ietf.org
<br>Subject: Re: [Tsvwg] First DATA chunk
<br>Hi Vijay,
<br>there is no need to send SACK if u dont receive DATA packets at all.......
<br>ur
<br>second point is right..
<br>Thanks
<br>-arun
<br>Vijay Kumar Choudhary wrote:
<br>> Hi All,
<br>> In SACK chunk, what will be the value of the cumulative TSN Ack if
the
<br>> first DATA chunk is missed?
<br>>
<br>> Or,
<br>>
<br>> If the first DATA chunk is missed, does the SACK needs to be send
at all?
<br>> Regards,
<br>> Vijay
<br>>
<br>> _______________________________________________
<br>> tsvwg mailing list
<br>> tsvwg@ietf.org
<br>> <a href="https://www1.ietf.org/mailman/listinfo/tsvwg">https://www1.ietf.org/mailman/listinfo/tsvwg</a>
<br>--
<br>****************************************************************
<br>V.Arun Prasad
<br>HCL Technologies Ltd.
<br>51, Jawaharlal Nehru Road,
<br>Ekkattuthangal,
<br>Guindy Industrial Estate,
<br>Chennai - 600097.
<br>Contact # : 9144 - 2334174
<br>9144 - 2334181
<br>extn : 233
<br>****************************************************************
<br>_______________________________________________
<br>tsvwg mailing list
<br>tsvwg@ietf.org
<br><a href="https://www1.ietf.org/mailman/listinfo/tsvwg">https://www1.ietf.org/mailman/listinfo/tsvwg</a>
<p>_______________________________________________
<br>tsvwg mailing list
<br>tsvwg@ietf.org
<br><a href="https://www1.ietf.org/mailman/listinfo/tsvwg">https://www1.ietf.org/mailman/listinfo/tsvwg</a></blockquote>

<p>--
<br>****************************************************************
<br>V.Arun Prasad
<br>HCL Technologies Ltd.
<br>51, Jawaharlal Nehru Road,
<br>Ekkattuthangal,
<br>Guindy Industrial Estate,
<br>Chennai - 600097.
<p>Contact # : 9144 - 2334174
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
9144 - 2334181
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
extn : 233
<br>****************************************************************
<br>&nbsp;</html>

--------------58E197A42F4FE7DB32B97D49--


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 07:17:36 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA28041
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 07:17:32 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA24188
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 07:17:34 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA22473;
	Tue, 5 Feb 2002 06:43:33 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA22450
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 06:43:32 -0500 (EST)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA27629
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 06:43:28 -0500 (EST)
Received: from netlab.hcltech.com (arun [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id RAA29578;
	Tue, 5 Feb 2002 17:35:02 +0530
Message-ID: <3C5FC585.2521A329@netlab.hcltech.com>
Date: Tue, 05 Feb 2002 17:14:05 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Ashraf <mohammed.asharaf@wipro.com>
CC: Tsvwg <tsvwg@ietf.org>, Randall Stewart <randall@stewart.chicago.il.us>
Subject: Re: [Tsvwg] sctp Window deadlock
References: <NEBBJGDMMLHHCIKHGBEJMEGGCPAA.dotis@sanlight.net> <3C5F7C39.B86F4150@netlab.hcltech.com> <3C5F8128.CABB3467@wipro.com> <3C5F95BA.923DEE3F@netlab.hcltech.com> <3C5FC0DF.9E9EBF7A@wipro.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi Ashraf,
    ya RTT timer will be started, but upon its expiry, we come to the same problem..
because we cant send the retransmitted packets due to rwnd shortage......


Thanks
-arun

Ashraf wrote:

> Hi arun,
>     I think  for the new data chunk ( which we are sending  to probe for change in window), we start RTT timer and  give
> appropraite treatment, when it expires.
>
> thanks,
> ashraf
>
> Arun Prasad wrote:
>
> > Hi Ashraf,
> >     Thanks for the rsp....
> > ya, I agree that partial-delivery will solve the problem... But imagine an errornous
> > stack ( or SCTP stack which didnt implement the partial delivery API)  is on the
> > receiving side.... Then with the below example scenario, the Sender stack will
> > get stuck.... ie.,  the association will never tear down (as RTX timer is not running)
> > so the Sender side will be hung (with no mistake on its side....) forever.....
> > I think this situation should be avoided......
> >
> > Thanks
> > -arun
> >
> > Ashraf wrote:
> >
> > > Arun,
> > >     This is what RFC has to say :
> > > "If the data receiver  run out of buffer space while still  waiting for  for more fragments
> > > to complete the re-assembly, of the message, it should dispatch part of its inbound message
> > > through a partial deliver API, freeing some if its buffer space in order receive  rest of
> > > the message"
> > >
> > > guess this should take care of the deadlock  situation you are mentioning here.
> > >
> > > Thanks,
> > > Mohd. Ashraf
> > > Arun Prasad wrote:
> > >
> > > > Hi all,
> > > >     The question might be more of  SCTP implementation specific....   According
> > > > to the SCTP rfc2960, the data sender should not send any data packets if it doesnt
> > > > have enough Send-Windown (Peer Rwnd).... Just take the below example....
> > > >
> > > > The Peer's Rwnd (at the Init time ) is 1500 bytes.
> > > > The Sender receivies a data buffer of size 6000 from its upper layer to be sent
> > > > to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of
> > > > size 1500 bytes (because the local mtu size is 1500bytes)....
> > > > Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the
> > > > Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with
> > > > rwnd as 0. The Sender which would have queued the frags B,C,D... will then
> > > > send frag B as the probing data to the Peer... The peer again will respond with
> > > > a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the
> > > >     * Sender cant send any more data packets as the Peer's rwnd is zero....
> > > >     * Peer cant increase its rwnd till it releases the data chunks which it had already
> > > >        received,  but to release that it needs to reassemble the frags (ABCD) and
> > > >       deliver it to the above layer, only then it can release the buffers..... Since its
> > > >       waiting for the frags C and D...  this will never release the buffers.... so
> > > >       will not increase the rwnd.....
> > > >
> > > > This situation can be avoided if the Sender doesnt sent Data messages of size
> > > > greater than the Peer's rwnd.... But the Rfc doesnt say anything about this.....
> > > > Or there should be some timers, to get out of the Deadlock situation.....
> > > >
> > > > Any suggestions........
> > > >
> > > > Thanks
> > > > -arun
> > > >
> > > > _______________________________________________
> > > > tsvwg mailing list
> > > > tsvwg@ietf.org
> > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > >
> > >   ------------------------------------------------------------------------------------------------------------------------
> > >                            Name: Wipro_Disclaimer.txt
> > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > >                        Encoding: 7bit
> >
> > --
> > ****************************************************************
> > V.Arun Prasad
> > HCL Technologies Ltd.
> > 51, Jawaharlal Nehru Road,
> > Ekkattuthangal,
> > Guindy Industrial Estate,
> > Chennai - 600097.
> >
> > Contact # : 9144 - 2334174
> >             9144 - 2334181
> >             extn : 233
> > ****************************************************************
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
>
>   ------------------------------------------------------------------------------------------------------------------------
>                            Name: Wipro_Disclaimer.txt
>    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
>                        Encoding: 7bit

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 07:55:50 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA28796
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 07:55:50 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA26832
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 07:55:52 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA26214;
	Tue, 5 Feb 2002 07:35:20 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA26135
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 07:35:16 -0500 (EST)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA28463
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 07:35:12 -0500 (EST)
Received: from netlab.hcltech.com (arun [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id SAA31439;
	Tue, 5 Feb 2002 18:26:38 +0530
Message-ID: <3C5FD19D.3D0308A8@netlab.hcltech.com>
Date: Tue, 05 Feb 2002 18:05:41 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Randall Stewart <randall@stewart.chicago.il.us>
CC: Ashraf <mohammed.asharaf@wipro.com>, Tsvwg <tsvwg@ietf.org>,
        sachin <sachins@netlab.hcltech.com>
Subject: Re: [Tsvwg] sctp Window deadlock
References: <NEBBJGDMMLHHCIKHGBEJMEGGCPAA.dotis@sanlight.net> <3C5F7C39.B86F4150@netlab.hcltech.com> <3C5F8128.CABB3467@wipro.com> <3C5F95BA.923DEE3F@netlab.hcltech.com> <3C5FC314.D15DF4F9@stewart.chicago.il.us>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi Randall,
    Considering your 4th point about logging....
        I think the below scenario can be better handled
if the implementation of "Life_Time" (The user specifying for how long the particular
packet is to be tried to send) parameter in the SEND primitive  is present.
So once the timer (Life_time timer) expires, the user can be indicated that the
data cannot be sent (again for each each data request, there should be a timer
might be costly...)
This way atleast we may come to know that something wrong is going on in the
stack, and the user may choose to send a "Shutdown" primitive...
( I guess all the optional parameters also has some basic purpose.......)

Thanks
-arun

Randall Stewart wrote:

> Arun:
>
> Some comments:
>
> Arun Prasad wrote:
> >
> > Hi Ashraf,
> >     Thanks for the rsp....
> > ya, I agree that partial-delivery will solve the problem... But imagine an errornous
> > stack ( or SCTP stack which didnt implement the partial delivery API)  is on the
> > receiving side.... Then with the below example scenario, the Sender stack will
> > get stuck.... ie.,  the association will never tear down (as RTX timer is not running)
> > so the Sender side will be hung (with no mistake on its side....) forever.....
> > I think this situation should be avoided......
>
> I will address this comment below ...
>
> >
> > Thanks
> > -arun
> >
> > Ashraf wrote:
> >
> > > Arun,
> > >     This is what RFC has to say :
> > > "If the data receiver  run out of buffer space while still  waiting for  for more fragments
> > > to complete the re-assembly, of the message, it should dispatch part of its inbound message
> > > through a partial deliver API, freeing some if its buffer space in order receive  rest of
> > > the message"
> > >
> > > guess this should take care of the deadlock  situation you are mentioning here.
> > >
> > > Thanks,
> > > Mohd. Ashraf
> > > Arun Prasad wrote:
>
> The above is percisely what we designed the PD-API for. It is
> a part I never really much liked :> but we saw it to be the
> only way out of such a dead-lock.... Also teamed with this
> is the receivers right to reneg. This way in a critical situation
> it can drop some of the last chunks it is holding so it can
> make room for the earlier pieces so it CAN use the PD-API...
>
> Both are necessary to ensure no deadlock can occur..
>
> Now as to the other issue.. what happens if your sender
> is broken or an attacker.. a number of things come to mind
> (as yet I have not had my coffee so there may be more :0)
>
> 1) TCP I would imagine can be in the same situation .. i.e.
>    if it connects to a peer that is intentionally doing something
>    bad.
>
> 2) It really becomes implementation dependant. I know of one
> implementation
>    that will only keep sending the same chunk for so long.. and then
>    it will abort the associaition.
>
> 3) If the guy is really an attacker the problem is more in to the fact
>    that you have entered into the association with him. I.e don't
>    leave the front door open and expect a burgular not to enter...
>
> 4) If it is a broken implementation then it needs to get fixed and I
>    would hope the one stuck would at least have enough logging going
>    on to show that it is broken... (or even in the case 3 I think logs
>    may help) so someone can fix it..
>
> R
>
> > >
> > > > Hi all,
> > > >     The question might be more of  SCTP implementation specific....   According
> > > > to the SCTP rfc2960, the data sender should not send any data packets if it doesnt
> > > > have enough Send-Windown (Peer Rwnd).... Just take the below example....
> > > >
> > > > The Peer's Rwnd (at the Init time ) is 1500 bytes.
> > > > The Sender receivies a data buffer of size 6000 from its upper layer to be sent
> > > > to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of
> > > > size 1500 bytes (because the local mtu size is 1500bytes)....
> > > > Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the
> > > > Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with
> > > > rwnd as 0. The Sender which would have queued the frags B,C,D... will then
> > > > send frag B as the probing data to the Peer... The peer again will respond with
> > > > a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the
> > > >     * Sender cant send any more data packets as the Peer's rwnd is zero....
> > > >     * Peer cant increase its rwnd till it releases the data chunks which it had already
> > > >        received,  but to release that it needs to reassemble the frags (ABCD) and
> > > >       deliver it to the above layer, only then it can release the buffers..... Since its
> > > >       waiting for the frags C and D...  this will never release the buffers.... so
> > > >       will not increase the rwnd.....
> > > >
> > > > This situation can be avoided if the Sender doesnt sent Data messages of size
> > > > greater than the Peer's rwnd.... But the Rfc doesnt say anything about this.....
> > > > Or there should be some timers, to get out of the Deadlock situation.....
> > > >
> > > > Any suggestions........
> > > >
> > > > Thanks
> > > > -arun
> > > >
> > > > _______________________________________________
> > > > tsvwg mailing list
> > > > tsvwg@ietf.org
> > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > >
> > >   ------------------------------------------------------------------------------------------------------------------------
> > >                            Name: Wipro_Disclaimer.txt
> > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > >                        Encoding: 7bit
> >
> > --
> > ****************************************************************
> > V.Arun Prasad
> > HCL Technologies Ltd.
> > 51, Jawaharlal Nehru Road,
> > Ekkattuthangal,
> > Guindy Industrial Estate,
> > Chennai - 600097.
> >
> > Contact # : 9144 - 2334174
> >             9144 - 2334181
> >             extn : 233
> > ****************************************************************
>
> --
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 08:23:42 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA29190
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 08:23:42 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA28200
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 08:23:44 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA25387;
	Tue, 5 Feb 2002 07:27:43 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA25362
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 07:27:41 -0500 (EST)
Received: from company.mail ([200.81.3.104])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA28312
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 07:27:36 -0500 (EST)
Message-Id: <200202051227.HAA28312@ietf.org>
Received: from localhost [127.0.0.1]
	by company.mail [209.99.230.66]
	with SMTP (MDaemon.PRO.v5.0.1.R)
	for <tsvwg@ietf.org>; Tue, 05 Feb 2002 07:35:48 -0300
X-Sender: noreply@server.com
From: Catering <noreply@server.com>
To: tsvwg@ietf.org
Date: Tue, 05 Feb 2002 07:35:48 -0300
Reply-To: noreply@server.com
Organization: zahori
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-MDRemoteIP: 127.0.0.1
X-Return-Path: noreply@server.com
X-MDaemon-Deliver-To: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit
Subject: [Tsvwg] Para que ese momento sea el mejor
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

ZAHORÍ Catering 

Nos dedicamos a brindar el mejor servicio profesional de catering para todo tipo 
de eventos, ( Casamientos - 15 años - Bautismos - Comuniones - Asados - Cenas 
- Almuerzos - Aniversarios - Servicios empresariales - Congresos - Exposiciones 
- Cocktails - Lunch - Presentacion de productos - Coffee break ) y todo lo que 
en tu imaginacion entre. 
Nos adaptamos a tu gusto y presupuesto. Ademas de nuestro salon en el barrio 
de Belgrano, contás con la posibilidad de realizar nuestro servicio en cualquier 
lugar de Buenos Aires. 
Indicanos tus necesidades a zahori3@millic.com.ar y nosotros te enviamos el presupuesto 
hecho a tu medida.
Ademas te ofrecemos Servicio de Disc Jokey con o sin catering consultanos
Cordialmente ZAHORÍ Catering. 

Este e-mail es enviado por unica vez, Gracias y disculpe las molestias que pudiéramos 
haberle ocasionado en caso de que esta información no sea de su interés


			



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 08:36:40 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA29332
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 08:36:39 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA28962
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 08:36:42 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA27678;
	Tue, 5 Feb 2002 08:08:54 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA27651
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 08:08:52 -0500 (EST)
Received: from sj-msg-core-1.cisco.com (sj-msg-core-1.cisco.com [171.71.163.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA28911
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 08:08:48 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-1.cisco.com (8.11.3/8.9.1) with ESMTP id g15D7lg13955;
	Tue, 5 Feb 2002 05:07:47 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id AAZ80607;
	Tue, 5 Feb 2002 05:07:45 -0800 (PST)
Message-ID: <3C5FD920.43D82001@stewart.chicago.il.us>
Date: Tue, 05 Feb 2002 07:07:44 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Arun Prasad <arun@netlab.hcltech.com>
CC: Ashraf <mohammed.asharaf@wipro.com>, Tsvwg <tsvwg@ietf.org>,
        sachin <sachins@netlab.hcltech.com>
Subject: Re: [Tsvwg] sctp Window deadlock
References: <NEBBJGDMMLHHCIKHGBEJMEGGCPAA.dotis@sanlight.net> <3C5F7C39.B86F4150@netlab.hcltech.com> <3C5F8128.CABB3467@wipro.com> <3C5F95BA.923DEE3F@netlab.hcltech.com> <3C5FC314.D15DF4F9@stewart.chicago.il.us> <3C5FD19D.3D0308A8@netlab.hcltech.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Arun:

Interesting idea... but this will not help you. The lifetime
as specified can only effect packets that have NOT been
transmitted. 

As I said.. some implementations MAY take and implement an
overall number of times a chunk may be transmitted.. and if
you keep getting it retran'd and no ack.. you get an abort :>

The u-sctp draft had something more like what you are considering
below.. but that is not currently an active document in the IETF.

There are some of us still playing with it and you can find the
last updated document on the web site

http://www.sctp.org

if you are interested...

R

Arun Prasad wrote:
> 
> Hi Randall,
>     Considering your 4th point about logging....
>         I think the below scenario can be better handled
> if the implementation of "Life_Time" (The user specifying for how long the particular
> packet is to be tried to send) parameter in the SEND primitive  is present.
> So once the timer (Life_time timer) expires, the user can be indicated that the
> data cannot be sent (again for each each data request, there should be a timer
> might be costly...)
> This way atleast we may come to know that something wrong is going on in the
> stack, and the user may choose to send a "Shutdown" primitive...
> ( I guess all the optional parameters also has some basic purpose.......)
> 
> Thanks
> -arun
> 
> Randall Stewart wrote:
> 
> > Arun:
> >
> > Some comments:
> >
> > Arun Prasad wrote:
> > >
> > > Hi Ashraf,
> > >     Thanks for the rsp....
> > > ya, I agree that partial-delivery will solve the problem... But imagine an errornous
> > > stack ( or SCTP stack which didnt implement the partial delivery API)  is on the
> > > receiving side.... Then with the below example scenario, the Sender stack will
> > > get stuck.... ie.,  the association will never tear down (as RTX timer is not running)
> > > so the Sender side will be hung (with no mistake on its side....) forever.....
> > > I think this situation should be avoided......
> >
> > I will address this comment below ...
> >
> > >
> > > Thanks
> > > -arun
> > >
> > > Ashraf wrote:
> > >
> > > > Arun,
> > > >     This is what RFC has to say :
> > > > "If the data receiver  run out of buffer space while still  waiting for  for more fragments
> > > > to complete the re-assembly, of the message, it should dispatch part of its inbound message
> > > > through a partial deliver API, freeing some if its buffer space in order receive  rest of
> > > > the message"
> > > >
> > > > guess this should take care of the deadlock  situation you are mentioning here.
> > > >
> > > > Thanks,
> > > > Mohd. Ashraf
> > > > Arun Prasad wrote:
> >
> > The above is percisely what we designed the PD-API for. It is
> > a part I never really much liked :> but we saw it to be the
> > only way out of such a dead-lock.... Also teamed with this
> > is the receivers right to reneg. This way in a critical situation
> > it can drop some of the last chunks it is holding so it can
> > make room for the earlier pieces so it CAN use the PD-API...
> >
> > Both are necessary to ensure no deadlock can occur..
> >
> > Now as to the other issue.. what happens if your sender
> > is broken or an attacker.. a number of things come to mind
> > (as yet I have not had my coffee so there may be more :0)
> >
> > 1) TCP I would imagine can be in the same situation .. i.e.
> >    if it connects to a peer that is intentionally doing something
> >    bad.
> >
> > 2) It really becomes implementation dependant. I know of one
> > implementation
> >    that will only keep sending the same chunk for so long.. and then
> >    it will abort the associaition.
> >
> > 3) If the guy is really an attacker the problem is more in to the fact
> >    that you have entered into the association with him. I.e don't
> >    leave the front door open and expect a burgular not to enter...
> >
> > 4) If it is a broken implementation then it needs to get fixed and I
> >    would hope the one stuck would at least have enough logging going
> >    on to show that it is broken... (or even in the case 3 I think logs
> >    may help) so someone can fix it..
> >
> > R
> >
> > > >
> > > > > Hi all,
> > > > >     The question might be more of  SCTP implementation specific....   According
> > > > > to the SCTP rfc2960, the data sender should not send any data packets if it doesnt
> > > > > have enough Send-Windown (Peer Rwnd).... Just take the below example....
> > > > >
> > > > > The Peer's Rwnd (at the Init time ) is 1500 bytes.
> > > > > The Sender receivies a data buffer of size 6000 from its upper layer to be sent
> > > > > to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of
> > > > > size 1500 bytes (because the local mtu size is 1500bytes)....
> > > > > Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the
> > > > > Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with
> > > > > rwnd as 0. The Sender which would have queued the frags B,C,D... will then
> > > > > send frag B as the probing data to the Peer... The peer again will respond with
> > > > > a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the
> > > > >     * Sender cant send any more data packets as the Peer's rwnd is zero....
> > > > >     * Peer cant increase its rwnd till it releases the data chunks which it had already
> > > > >        received,  but to release that it needs to reassemble the frags (ABCD) and
> > > > >       deliver it to the above layer, only then it can release the buffers..... Since its
> > > > >       waiting for the frags C and D...  this will never release the buffers.... so
> > > > >       will not increase the rwnd.....
> > > > >
> > > > > This situation can be avoided if the Sender doesnt sent Data messages of size
> > > > > greater than the Peer's rwnd.... But the Rfc doesnt say anything about this.....
> > > > > Or there should be some timers, to get out of the Deadlock situation.....
> > > > >
> > > > > Any suggestions........
> > > > >
> > > > > Thanks
> > > > > -arun
> > > > >
> > > > > _______________________________________________
> > > > > tsvwg mailing list
> > > > > tsvwg@ietf.org
> > > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > >
> > > >   ------------------------------------------------------------------------------------------------------------------------
> > > >                            Name: Wipro_Disclaimer.txt
> > > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > > >                        Encoding: 7bit
> > >
> > > --
> > > ****************************************************************
> > > V.Arun Prasad
> > > HCL Technologies Ltd.
> > > 51, Jawaharlal Nehru Road,
> > > Ekkattuthangal,
> > > Guindy Industrial Estate,
> > > Chennai - 600097.
> > >
> > > Contact # : 9144 - 2334174
> > >             9144 - 2334181
> > >             extn : 233
> > > ****************************************************************
> >
> > --
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> --
> ****************************************************************
> V.Arun Prasad
> HCL Technologies Ltd.
> 51, Jawaharlal Nehru Road,
> Ekkattuthangal,
> Guindy Industrial Estate,
> Chennai - 600097.
> 
> Contact # : 9144 - 2334174
>             9144 - 2334181
>             extn : 233
> ****************************************************************

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 09:40:31 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA01121
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 09:40:31 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA02558
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 09:40:32 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA29301;
	Tue, 5 Feb 2002 08:46:23 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA29273
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 08:46:21 -0500 (EST)
Received: from wiproecmx1.wipro.com (wiproecmx1.wipro.com [164.164.31.5])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA29646
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 08:46:14 -0500 (EST)
Received: from ecvwall1.wipro.com (ecvwall1.wipro.com [164.164.23.6])
	by wiproecmx1.wipro.com (8.11.3/8.11.3) with SMTP id g15DijX19005
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 19:14:46 +0530 (IST)
Received: from wipro.com ([192.168.178.83]) by
          ecmail.mail.wipro.com (Netscape Messaging Server 4.15) with
          ESMTP id GR2BJ201.4BT; Tue, 5 Feb 2002 19:15:02 +0530 
Message-ID: <3C5FE27E.CC6923B2@wipro.com>
Date: Tue, 05 Feb 2002 19:17:42 +0530
From: Ashraf <mohammed.asharaf@wipro.com>
Organization: Wipro
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.4.2-2 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Arun Prasad <arun@netlab.hcltech.com>
CC: Tsvwg <tsvwg@ietf.org>, Randall Stewart <randall@stewart.chicago.il.us>
Subject: Re: [Tsvwg] sctp Window deadlock
References: <NEBBJGDMMLHHCIKHGBEJMEGGCPAA.dotis@sanlight.net> <3C5F7C39.B86F4150@netlab.hcltech.com> <3C5F8128.CABB3467@wipro.com> <3C5F95BA.923DEE3F@netlab.hcltech.com> <3C5FC0DF.9E9EBF7A@wipro.com> <3C5FC585.2521A329@netlab.hcltech.com>
Content-Type: multipart/mixed;
	boundary="----=_NextPartTM-000-8da9a3ef-1a33-11d6-a942-00b0d0d06be8"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


This is a multi-part message in MIME format.

------=_NextPartTM-000-8da9a3ef-1a33-11d6-a942-00b0d0d06be8
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hi,
    My understanding is ( pls correct me if I am wromg),  each time when the timer expires, the path /assoc retrans counter will be

incremented and sender keeps on probing  for change in window size  by sending next fragment ( in ur caase fragment B ) and  ending

up in T3 timer expiry each time. Will the Assoc will be  terminated once this count reached the threshold value.?"

Thanks,
ashraf


Arun Prasad wrote:

> Hi Ashraf,
>     ya RTT timer will be started, but upon its expiry, we come to the same problem..
> because we cant send the retransmitted packets due to rwnd shortage......
>
> Thanks
> -arun
>
> Ashraf wrote:
>
> > Hi arun,
> >     I think  for the new data chunk ( which we are sending  to probe for change in window), we start RTT timer and  give
> > appropraite treatment, when it expires.
> >
> > thanks,
> > ashraf
> >
> > Arun Prasad wrote:
> >
> > > Hi Ashraf,
> > >     Thanks for the rsp....
> > > ya, I agree that partial-delivery will solve the problem... But imagine an errornous
> > > stack ( or SCTP stack which didnt implement the partial delivery API)  is on the
> > > receiving side.... Then with the below example scenario, the Sender stack will
> > > get stuck.... ie.,  the association will never tear down (as RTX timer is not running)
> > > so the Sender side will be hung (with no mistake on its side....) forever.....
> > > I think this situation should be avoided......
> > >
> > > Thanks
> > > -arun
> > >
> > > Ashraf wrote:
> > >
> > > > Arun,
> > > >     This is what RFC has to say :
> > > > "If the data receiver  run out of buffer space while still  waiting for  for more fragments
> > > > to complete the re-assembly, of the message, it should dispatch part of its inbound message
> > > > through a partial deliver API, freeing some if its buffer space in order receive  rest of
> > > > the message"
> > > >
> > > > guess this should take care of the deadlock  situation you are mentioning here.
> > > >
> > > > Thanks,
> > > > Mohd. Ashraf
> > > > Arun Prasad wrote:
> > > >
> > > > > Hi all,
> > > > >     The question might be more of  SCTP implementation specific....   According
> > > > > to the SCTP rfc2960, the data sender should not send any data packets if it doesnt
> > > > > have enough Send-Windown (Peer Rwnd).... Just take the below example....
> > > > >
> > > > > The Peer's Rwnd (at the Init time ) is 1500 bytes.
> > > > > The Sender receivies a data buffer of size 6000 from its upper layer to be sent
> > > > > to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of
> > > > > size 1500 bytes (because the local mtu size is 1500bytes)....
> > > > > Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the
> > > > > Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with
> > > > > rwnd as 0. The Sender which would have queued the frags B,C,D... will then
> > > > > send frag B as the probing data to the Peer... The peer again will respond with
> > > > > a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the
> > > > >     * Sender cant send any more data packets as the Peer's rwnd is zero....
> > > > >     * Peer cant increase its rwnd till it releases the data chunks which it had already
> > > > >        received,  but to release that it needs to reassemble the frags (ABCD) and
> > > > >       deliver it to the above layer, only then it can release the buffers..... Since its
> > > > >       waiting for the frags C and D...  this will never release the buffers.... so
> > > > >       will not increase the rwnd.....
> > > > >
> > > > > This situation can be avoided if the Sender doesnt sent Data messages of size
> > > > > greater than the Peer's rwnd.... But the Rfc doesnt say anything about this.....
> > > > > Or there should be some timers, to get out of the Deadlock situation.....
> > > > >
> > > > > Any suggestions........
> > > > >
> > > > > Thanks
> > > > > -arun
> > > > >
> > > > > _______________________________________________
> > > > > tsvwg mailing list
> > > > > tsvwg@ietf.org
> > > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > >
> > > >   ------------------------------------------------------------------------------------------------------------------------
> > > >                            Name: Wipro_Disclaimer.txt
> > > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > > >                        Encoding: 7bit
> > >
> > > --
> > > ****************************************************************
> > > V.Arun Prasad
> > > HCL Technologies Ltd.
> > > 51, Jawaharlal Nehru Road,
> > > Ekkattuthangal,
> > > Guindy Industrial Estate,
> > > Chennai - 600097.
> > >
> > > Contact # : 9144 - 2334174
> > >             9144 - 2334181
> > >             extn : 233
> > > ****************************************************************
> > >
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> >
> >   ------------------------------------------------------------------------------------------------------------------------
> >                            Name: Wipro_Disclaimer.txt
> >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> >                        Encoding: 7bit
>
> --
> ****************************************************************
> V.Arun Prasad
> HCL Technologies Ltd.
> 51, Jawaharlal Nehru Road,
> Ekkattuthangal,
> Guindy Industrial Estate,
> Chennai - 600097.
>
> Contact # : 9144 - 2334174
>             9144 - 2334181
>             extn : 233
> ****************************************************************


------=_NextPartTM-000-8da9a3ef-1a33-11d6-a942-00b0d0d06be8
Content-Type: text/plain;
	name="Wipro_Disclaimer.txt"
Content-Disposition: attachment;
	filename="Wipro_Disclaimer.txt"
Content-Transfer-Encoding: 7bit

**************************Disclaimer************************************
      


Information contained in this E-MAIL being proprietary to Wipro Limited
is 'privileged' and 'confidential' and intended for use only by the
individual or entity to which it is addressed. You are notified that any
use, copying or dissemination of the information contained in the E-MAIL
in any manner whatsoever is strictly prohibited.



 ********************************************************************

------=_NextPartTM-000-8da9a3ef-1a33-11d6-a942-00b0d0d06be8--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 09:46:57 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA01431
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 09:46:57 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA03042
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 09:46:58 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA01216;
	Tue, 5 Feb 2002 09:27:31 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA01178
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 09:27:28 -0500 (EST)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA00788
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 09:27:23 -0500 (EST)
Received: from netlab.hcltech.com (arun [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id UAA01862;
	Tue, 5 Feb 2002 20:18:52 +0530
Message-ID: <3C5FEBEB.A79EB5AD@netlab.hcltech.com>
Date: Tue, 05 Feb 2002 19:57:55 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Randall Stewart <randall@stewart.chicago.il.us>
CC: Ashraf <mohammed.asharaf@wipro.com>, Tsvwg <tsvwg@ietf.org>,
        sachin <sachins@netlab.hcltech.com>
Subject: Re: [Tsvwg] sctp Window deadlock
References: <NEBBJGDMMLHHCIKHGBEJMEGGCPAA.dotis@sanlight.net> <3C5F7C39.B86F4150@netlab.hcltech.com> <3C5F8128.CABB3467@wipro.com> <3C5F95BA.923DEE3F@netlab.hcltech.com> <3C5FC314.D15DF4F9@stewart.chicago.il.us> <3C5FD19D.3D0308A8@netlab.hcltech.com> <3C5FD920.43D82001@stewart.chicago.il.us>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi Randell,
    I think we misunderstood something.....  The scenario we are considering will
not send the data packet to the network at all.... When  the Peer- rwnd is zero,
one data packet is sent for probing,  then if the endpoint still receives the SACK
with rwnd as 0, then the stack will not send any packets to the network, instead
will Queue the data chunks in the local Queue and will be waiting for the
Peer-rwnd to increase (I think only one probing data chunk is allowed when
the Peer-rwnd is zero...) . So the solution u are mentioning ( the
MAX-ASSOC-RETRANS counter ) will not help because the counter will
be increamented only when the data chunk is sent to the network.....
Hope Iam right.......

Thanks
-arun

Randall Stewart wrote:

> Arun:
>
> Interesting idea... but this will not help you. The lifetime
> as specified can only effect packets that have NOT been
> transmitted.
>
> As I said.. some implementations MAY take and implement an
> overall number of times a chunk may be transmitted.. and if
> you keep getting it retran'd and no ack.. you get an abort :>
>
> The u-sctp draft had something more like what you are considering
> below.. but that is not currently an active document in the IETF.
>
> There are some of us still playing with it and you can find the
> last updated document on the web site
>
> http://www.sctp.org
>
> if you are interested...
>
> R
>
> Arun Prasad wrote:
> >
> > Hi Randall,
> >     Considering your 4th point about logging....
> >         I think the below scenario can be better handled
> > if the implementation of "Life_Time" (The user specifying for how long the particular
> > packet is to be tried to send) parameter in the SEND primitive  is present.
> > So once the timer (Life_time timer) expires, the user can be indicated that the
> > data cannot be sent (again for each each data request, there should be a timer
> > might be costly...)
> > This way atleast we may come to know that something wrong is going on in the
> > stack, and the user may choose to send a "Shutdown" primitive...
> > ( I guess all the optional parameters also has some basic purpose.......)
> >
> > Thanks
> > -arun
> >
> > Randall Stewart wrote:
> >
> > > Arun:
> > >
> > > Some comments:
> > >
> > > Arun Prasad wrote:
> > > >
> > > > Hi Ashraf,
> > > >     Thanks for the rsp....
> > > > ya, I agree that partial-delivery will solve the problem... But imagine an errornous
> > > > stack ( or SCTP stack which didnt implement the partial delivery API)  is on the
> > > > receiving side.... Then with the below example scenario, the Sender stack will
> > > > get stuck.... ie.,  the association will never tear down (as RTX timer is not running)
> > > > so the Sender side will be hung (with no mistake on its side....) forever.....
> > > > I think this situation should be avoided......
> > >
> > > I will address this comment below ...
> > >
> > > >
> > > > Thanks
> > > > -arun
> > > >
> > > > Ashraf wrote:
> > > >
> > > > > Arun,
> > > > >     This is what RFC has to say :
> > > > > "If the data receiver  run out of buffer space while still  waiting for  for more fragments
> > > > > to complete the re-assembly, of the message, it should dispatch part of its inbound message
> > > > > through a partial deliver API, freeing some if its buffer space in order receive  rest of
> > > > > the message"
> > > > >
> > > > > guess this should take care of the deadlock  situation you are mentioning here.
> > > > >
> > > > > Thanks,
> > > > > Mohd. Ashraf
> > > > > Arun Prasad wrote:
> > >
> > > The above is percisely what we designed the PD-API for. It is
> > > a part I never really much liked :> but we saw it to be the
> > > only way out of such a dead-lock.... Also teamed with this
> > > is the receivers right to reneg. This way in a critical situation
> > > it can drop some of the last chunks it is holding so it can
> > > make room for the earlier pieces so it CAN use the PD-API...
> > >
> > > Both are necessary to ensure no deadlock can occur..
> > >
> > > Now as to the other issue.. what happens if your sender
> > > is broken or an attacker.. a number of things come to mind
> > > (as yet I have not had my coffee so there may be more :0)
> > >
> > > 1) TCP I would imagine can be in the same situation .. i.e.
> > >    if it connects to a peer that is intentionally doing something
> > >    bad.
> > >
> > > 2) It really becomes implementation dependant. I know of one
> > > implementation
> > >    that will only keep sending the same chunk for so long.. and then
> > >    it will abort the associaition.
> > >
> > > 3) If the guy is really an attacker the problem is more in to the fact
> > >    that you have entered into the association with him. I.e don't
> > >    leave the front door open and expect a burgular not to enter...
> > >
> > > 4) If it is a broken implementation then it needs to get fixed and I
> > >    would hope the one stuck would at least have enough logging going
> > >    on to show that it is broken... (or even in the case 3 I think logs
> > >    may help) so someone can fix it..
> > >
> > > R
> > >
> > > > >
> > > > > > Hi all,
> > > > > >     The question might be more of  SCTP implementation specific....   According
> > > > > > to the SCTP rfc2960, the data sender should not send any data packets if it doesnt
> > > > > > have enough Send-Windown (Peer Rwnd).... Just take the below example....
> > > > > >
> > > > > > The Peer's Rwnd (at the Init time ) is 1500 bytes.
> > > > > > The Sender receivies a data buffer of size 6000 from its upper layer to be sent
> > > > > > to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of
> > > > > > size 1500 bytes (because the local mtu size is 1500bytes)....
> > > > > > Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the
> > > > > > Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with
> > > > > > rwnd as 0. The Sender which would have queued the frags B,C,D... will then
> > > > > > send frag B as the probing data to the Peer... The peer again will respond with
> > > > > > a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the
> > > > > >     * Sender cant send any more data packets as the Peer's rwnd is zero....
> > > > > >     * Peer cant increase its rwnd till it releases the data chunks which it had already
> > > > > >        received,  but to release that it needs to reassemble the frags (ABCD) and
> > > > > >       deliver it to the above layer, only then it can release the buffers..... Since its
> > > > > >       waiting for the frags C and D...  this will never release the buffers.... so
> > > > > >       will not increase the rwnd.....
> > > > > >
> > > > > > This situation can be avoided if the Sender doesnt sent Data messages of size
> > > > > > greater than the Peer's rwnd.... But the Rfc doesnt say anything about this.....
> > > > > > Or there should be some timers, to get out of the Deadlock situation.....
> > > > > >
> > > > > > Any suggestions........
> > > > > >
> > > > > > Thanks
> > > > > > -arun
> > > > > >
> > > > > > _______________________________________________
> > > > > > tsvwg mailing list
> > > > > > tsvwg@ietf.org
> > > > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > > >
> > > > >   ------------------------------------------------------------------------------------------------------------------------
> > > > >                            Name: Wipro_Disclaimer.txt
> > > > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > > > >                        Encoding: 7bit
> > > >
> > > > --
> > > > ****************************************************************
> > > > V.Arun Prasad
> > > > HCL Technologies Ltd.
> > > > 51, Jawaharlal Nehru Road,
> > > > Ekkattuthangal,
> > > > Guindy Industrial Estate,
> > > > Chennai - 600097.
> > > >
> > > > Contact # : 9144 - 2334174
> > > >             9144 - 2334181
> > > >             extn : 233
> > > > ****************************************************************
> > >
> > > --
> > > Randall R. Stewart
> > > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> >
> > --
> > ****************************************************************
> > V.Arun Prasad
> > HCL Technologies Ltd.
> > 51, Jawaharlal Nehru Road,
> > Ekkattuthangal,
> > Guindy Industrial Estate,
> > Chennai - 600097.
> >
> > Contact # : 9144 - 2334174
> >             9144 - 2334181
> >             extn : 233
> > ****************************************************************
>
> --
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 10:19:55 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA03246
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 10:19:55 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA05799
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 10:19:56 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA02363;
	Tue, 5 Feb 2002 09:36:08 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA02242
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 09:36:02 -0500 (EST)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA01028
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 09:35:57 -0500 (EST)
Received: from netlab.hcltech.com (arun [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id UAA02250;
	Tue, 5 Feb 2002 20:27:36 +0530
Message-ID: <3C5FEDF7.CE4FFE0E@netlab.hcltech.com>
Date: Tue, 05 Feb 2002 20:06:39 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Ashraf <mohammed.asharaf@wipro.com>
CC: Tsvwg <tsvwg@ietf.org>, Randall Stewart <randall@stewart.chicago.il.us>
Subject: Re: [Tsvwg] sctp Window deadlock
References: <NEBBJGDMMLHHCIKHGBEJMEGGCPAA.dotis@sanlight.net> <3C5F7C39.B86F4150@netlab.hcltech.com> <3C5F8128.CABB3467@wipro.com> <3C5F95BA.923DEE3F@netlab.hcltech.com> <3C5FC0DF.9E9EBF7A@wipro.com> <3C5FC585.2521A329@netlab.hcltech.com> <3C5FE27E.CC6923B2@wipro.com>
Content-Type: multipart/alternative;
 boundary="------------48E1687CF03AA848973D7833"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


--------------48E1687CF03AA848973D7833
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

hi Ashraf,

From Rfc:
However, regardless of the value of rwnd (including if it
      is 0), the data sender can always have one DATA chunk in flight to
      the receiver if allowed by cwnd

    My understanding from the above is, trying to send the Data chunk only once
for probing, and if that chunk is not acknowledged (or acknowledged with rwnd
as zero), then even when the Rtx timer expires for that Data chunk, the data chunk
will be queued in the local Q (as the rwnd is zero, and already one data chunk
has been sent for probing...) and will not be retransimtted.....
    I dont know whether my understanding is right......  Just getting a feeling that
Iam going wrong somewhere....( might be ur understanding is right.....)
Somebody please clarify.....

Thanks
-arun


Ashraf wrote:

> Hi,
>     My understanding is ( pls correct me if I am wromg),  each time when the timer expires, the path /assoc retrans counter will be
>
> incremented and sender keeps on probing  for change in window size  by sending next fragment ( in ur caase fragment B ) and  ending
>
> up in T3 timer expiry each time. Will the Assoc will be  terminated once this count reached the threshold value.?"
>
> Thanks,
> ashraf
>
> Arun Prasad wrote:
>
> > Hi Ashraf,
> >     ya RTT timer will be started, but upon its expiry, we come to the same problem..
> > because we cant send the retransmitted packets due to rwnd shortage......
> >
> > Thanks
> > -arun
> >
> > Ashraf wrote:
> >
> > > Hi arun,
> > >     I think  for the new data chunk ( which we are sending  to probe for change in window), we start RTT timer and  give
> > > appropraite treatment, when it expires.
> > >
> > > thanks,
> > > ashraf
> > >
> > > Arun Prasad wrote:
> > >
> > > > Hi Ashraf,
> > > >     Thanks for the rsp....
> > > > ya, I agree that partial-delivery will solve the problem... But imagine an errornous
> > > > stack ( or SCTP stack which didnt implement the partial delivery API)  is on the
> > > > receiving side.... Then with the below example scenario, the Sender stack will
> > > > get stuck.... ie.,  the association will never tear down (as RTX timer is not running)
> > > > so the Sender side will be hung (with no mistake on its side....) forever.....
> > > > I think this situation should be avoided......
> > > >
> > > > Thanks
> > > > -arun
> > > >
> > > > Ashraf wrote:
> > > >
> > > > > Arun,
> > > > >     This is what RFC has to say :
> > > > > "If the data receiver  run out of buffer space while still  waiting for  for more fragments
> > > > > to complete the re-assembly, of the message, it should dispatch part of its inbound message
> > > > > through a partial deliver API, freeing some if its buffer space in order receive  rest of
> > > > > the message"
> > > > >
> > > > > guess this should take care of the deadlock  situation you are mentioning here.
> > > > >
> > > > > Thanks,
> > > > > Mohd. Ashraf
> > > > > Arun Prasad wrote:
> > > > >
> > > > > > Hi all,
> > > > > >     The question might be more of  SCTP implementation specific....   According
> > > > > > to the SCTP rfc2960, the data sender should not send any data packets if it doesnt
> > > > > > have enough Send-Windown (Peer Rwnd).... Just take the below example....
> > > > > >
> > > > > > The Peer's Rwnd (at the Init time ) is 1500 bytes.
> > > > > > The Sender receivies a data buffer of size 6000 from its upper layer to be sent
> > > > > > to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of
> > > > > > size 1500 bytes (because the local mtu size is 1500bytes)....
> > > > > > Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the
> > > > > > Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with
> > > > > > rwnd as 0. The Sender which would have queued the frags B,C,D... will then
> > > > > > send frag B as the probing data to the Peer... The peer again will respond with
> > > > > > a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the
> > > > > >     * Sender cant send any more data packets as the Peer's rwnd is zero....
> > > > > >     * Peer cant increase its rwnd till it releases the data chunks which it had already
> > > > > >        received,  but to release that it needs to reassemble the frags (ABCD) and
> > > > > >       deliver it to the above layer, only then it can release the buffers..... Since its
> > > > > >       waiting for the frags C and D...  this will never release the buffers.... so
> > > > > >       will not increase the rwnd.....
> > > > > >
> > > > > > This situation can be avoided if the Sender doesnt sent Data messages of size
> > > > > > greater than the Peer's rwnd.... But the Rfc doesnt say anything about this.....
> > > > > > Or there should be some timers, to get out of the Deadlock situation.....
> > > > > >
> > > > > > Any suggestions........
> > > > > >
> > > > > > Thanks
> > > > > > -arun
> > > > > >
> > > > > > _______________________________________________
> > > > > > tsvwg mailing list
> > > > > > tsvwg@ietf.org
> > > > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > > >
> > > > >   ------------------------------------------------------------------------------------------------------------------------
> > > > >                            Name: Wipro_Disclaimer.txt
> > > > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > > > >                        Encoding: 7bit
> > > >
> > > > --
> > > > ****************************************************************
> > > > V.Arun Prasad
> > > > HCL Technologies Ltd.
> > > > 51, Jawaharlal Nehru Road,
> > > > Ekkattuthangal,
> > > > Guindy Industrial Estate,
> > > > Chennai - 600097.
> > > >
> > > > Contact # : 9144 - 2334174
> > > >             9144 - 2334181
> > > >             extn : 233
> > > > ****************************************************************
> > > >
> > > > _______________________________________________
> > > > tsvwg mailing list
> > > > tsvwg@ietf.org
> > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > >
> > >   ------------------------------------------------------------------------------------------------------------------------
> > >                            Name: Wipro_Disclaimer.txt
> > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > >                        Encoding: 7bit
> >
> > --
> > ****************************************************************
> > V.Arun Prasad
> > HCL Technologies Ltd.
> > 51, Jawaharlal Nehru Road,
> > Ekkattuthangal,
> > Guindy Industrial Estate,
> > Chennai - 600097.
> >
> > Contact # : 9144 - 2334174
> >             9144 - 2334181
> >             extn : 233
> > ****************************************************************
>
>   ------------------------------------------------------------------------------------------------------------------------
>                            Name: Wipro_Disclaimer.txt
>    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
>                        Encoding: 7bit

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************


--------------48E1687CF03AA848973D7833
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
hi Ashraf,
<p>From Rfc:
<br><font color="#3333FF">However, regardless of the value of rwnd (including
if it</font>
<br><font color="#3333FF">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; is 0), the data
sender can always have <b>one DATA chunk in flight</b> to</font>
<br><font color="#3333FF">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; the receiver if
allowed by cwnd</font>
<p>&nbsp;&nbsp;&nbsp; My understanding from the above is, trying to send
the Data chunk only once
<br>for probing, and if that chunk is not acknowledged (or acknowledged
with rwnd
<br>as zero), then even when the Rtx timer expires for that Data chunk,
the data chunk
<br>will be queued in the local Q (as the rwnd is zero, and already one
data chunk
<br>has been sent for probing...) and will not be retransimtted.....
<br>&nbsp;&nbsp;&nbsp; I dont know whether my understanding is right......&nbsp;
Just getting a feeling that
<br>Iam going wrong somewhere....( might be ur understanding is right.....)
<br>Somebody please clarify.....
<p>Thanks
<br>-arun
<br>&nbsp;
<p>Ashraf wrote:
<blockquote TYPE=CITE>Hi,
<br>&nbsp;&nbsp;&nbsp; My understanding is ( pls correct me if I am wromg),&nbsp;
each time when the timer expires, the path /assoc retrans counter will
be
<p>incremented and sender keeps on probing&nbsp; for change in window size&nbsp;
by sending next fragment ( in ur caase fragment B ) and&nbsp; ending
<p>up in T3 timer expiry each time. Will the Assoc will be&nbsp; terminated
once this count reached the threshold value.?"
<p>Thanks,
<br>ashraf
<p>Arun Prasad wrote:
<p>> Hi Ashraf,
<br>>&nbsp;&nbsp;&nbsp;&nbsp; ya RTT timer will be started, but upon its
expiry, we come to the same problem..
<br>> because we cant send the retransmitted packets due to rwnd shortage......
<br>>
<br>> Thanks
<br>> -arun
<br>>
<br>> Ashraf wrote:
<br>>
<br>> > Hi arun,
<br>> >&nbsp;&nbsp;&nbsp;&nbsp; I think&nbsp; for the new data chunk (
which we are sending&nbsp; to probe for change in window), we start RTT
timer and&nbsp; give
<br>> > appropraite treatment, when it expires.
<br>> >
<br>> > thanks,
<br>> > ashraf
<br>> >
<br>> > Arun Prasad wrote:
<br>> >
<br>> > > Hi Ashraf,
<br>> > >&nbsp;&nbsp;&nbsp;&nbsp; Thanks for the rsp....
<br>> > > ya, I agree that partial-delivery will solve the problem... But
imagine an errornous
<br>> > > stack ( or SCTP stack which didnt implement the partial delivery
API)&nbsp; is on the
<br>> > > receiving side.... Then with the below example scenario, the
Sender stack will
<br>> > > get stuck.... ie.,&nbsp; the association will never tear down
(as RTX timer is not running)
<br>> > > so the Sender side will be hung (with no mistake on its side....)
forever.....
<br>> > > I think this situation should be avoided......
<br>> > >
<br>> > > Thanks
<br>> > > -arun
<br>> > >
<br>> > > Ashraf wrote:
<br>> > >
<br>> > > > Arun,
<br>> > > >&nbsp;&nbsp;&nbsp;&nbsp; This is what RFC has to say :
<br>> > > > "If the data receiver&nbsp; run out of buffer space while still&nbsp;
waiting for&nbsp; for more fragments
<br>> > > > to complete the re-assembly, of the message, it should dispatch
part of its inbound message
<br>> > > > through a partial deliver API, freeing some if its buffer space
in order receive&nbsp; rest of
<br>> > > > the message"
<br>> > > >
<br>> > > > guess this should take care of the deadlock&nbsp; situation
you are mentioning here.
<br>> > > >
<br>> > > > Thanks,
<br>> > > > Mohd. Ashraf
<br>> > > > Arun Prasad wrote:
<br>> > > >
<br>> > > > > Hi all,
<br>> > > > >&nbsp;&nbsp;&nbsp;&nbsp; The question might be more of&nbsp;
SCTP implementation specific....&nbsp;&nbsp; According
<br>> > > > > to the SCTP rfc2960, the data sender should not send any
data packets if it doesnt
<br>> > > > > have enough Send-Windown (Peer Rwnd).... Just take the below
example....
<br>> > > > >
<br>> > > > > The Peer's Rwnd (at the Init time ) is 1500 bytes.
<br>> > > > > The Sender receivies a data buffer of size 6000 from its
upper layer to be sent
<br>> > > > > to the Peer.... The Sender will then fragment the packets
to fragments A,B,C,D of
<br>> > > > > size 1500 bytes (because the local mtu size is 1500bytes)....
<br>> > > > > Now the Sender sends the frag A to the Peer... Upon receiving
the frag A, the
<br>> > > > > Peer's Rwnd will be zero.... This is informed to the Sender
by sending a SACK with
<br>> > > > > rwnd as 0. The Sender which would have queued the frags B,C,D...
will then
<br>> > > > > send frag B as the probing data to the Peer... The peer again
will respond with
<br>> > > > > a SACK of rwnd 0 .&nbsp; Now the deadlock situation occurs,&nbsp;
as the
<br>> > > > >&nbsp;&nbsp;&nbsp;&nbsp; * Sender cant send any more data
packets as the Peer's rwnd is zero....
<br>> > > > >&nbsp;&nbsp;&nbsp;&nbsp; * Peer cant increase its rwnd till
it releases the data chunks which it had already
<br>> > > > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; received,&nbsp;
but to release that it needs to reassemble the frags (ABCD) and
<br>> > > > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; deliver it to the above
layer, only then it can release the buffers..... Since its
<br>> > > > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; waiting for the frags
C and D...&nbsp; this will never release the buffers.... so
<br>> > > > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; will not increase the
rwnd.....
<br>> > > > >
<br>> > > > > This situation can be avoided if the Sender doesnt sent Data
messages of size
<br>> > > > > greater than the Peer's rwnd.... But the Rfc doesnt say anything
about this.....
<br>> > > > > Or there should be some timers, to get out of the Deadlock
situation.....
<br>> > > > >
<br>> > > > > Any suggestions........
<br>> > > > >
<br>> > > > > Thanks
<br>> > > > > -arun
<br>> > > > >
<br>> > > > > _______________________________________________
<br>> > > > > tsvwg mailing list
<br>> > > > > tsvwg@ietf.org
<br>> > > > > <a href="https://www1.ietf.org/mailman/listinfo/tsvwg">https://www1.ietf.org/mailman/listinfo/tsvwg</a>
<br>> > > >
<br>> > > >&nbsp;&nbsp; ------------------------------------------------------------------------------------------------------------------------
<br>> > > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
Name: Wipro_Disclaimer.txt
<br>> > > >&nbsp;&nbsp;&nbsp; Wipro_Disclaimer.txt&nbsp;&nbsp;&nbsp; Type:
Plain Text (text/plain)
<br>> > > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
Encoding: 7bit
<br>> > >
<br>> > > --
<br>> > > ****************************************************************
<br>> > > V.Arun Prasad
<br>> > > HCL Technologies Ltd.
<br>> > > 51, Jawaharlal Nehru Road,
<br>> > > Ekkattuthangal,
<br>> > > Guindy Industrial Estate,
<br>> > > Chennai - 600097.
<br>> > >
<br>> > > Contact # : 9144 - 2334174
<br>> > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
9144 - 2334181
<br>> > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
extn : 233
<br>> > > ****************************************************************
<br>> > >
<br>> > > _______________________________________________
<br>> > > tsvwg mailing list
<br>> > > tsvwg@ietf.org
<br>> > > <a href="https://www1.ietf.org/mailman/listinfo/tsvwg">https://www1.ietf.org/mailman/listinfo/tsvwg</a>
<br>> >
<br>> >&nbsp;&nbsp; ------------------------------------------------------------------------------------------------------------------------
<br>> >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
Name: Wipro_Disclaimer.txt
<br>> >&nbsp;&nbsp;&nbsp; Wipro_Disclaimer.txt&nbsp;&nbsp;&nbsp; Type:
Plain Text (text/plain)
<br>> >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
Encoding: 7bit
<br>>
<br>> --
<br>> ****************************************************************
<br>> V.Arun Prasad
<br>> HCL Technologies Ltd.
<br>> 51, Jawaharlal Nehru Road,
<br>> Ekkattuthangal,
<br>> Guindy Industrial Estate,
<br>> Chennai - 600097.
<br>>
<br>> Contact # : 9144 - 2334174
<br>>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
9144 - 2334181
<br>>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
extn : 233
<br>> ****************************************************************
<p>&nbsp; ------------------------------------------------------------------------------------------------------------------------
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
Name: Wipro_Disclaimer.txt
<br>&nbsp;&nbsp; Wipro_Disclaimer.txt&nbsp;&nbsp;&nbsp; Type: Plain Text
(text/plain)
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
Encoding: 7bit</blockquote>

<p>--
<br>****************************************************************
<br>V.Arun Prasad
<br>HCL Technologies Ltd.
<br>51, Jawaharlal Nehru Road,
<br>Ekkattuthangal,
<br>Guindy Industrial Estate,
<br>Chennai - 600097.
<p>Contact # : 9144 - 2334174
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
9144 - 2334181
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
extn : 233
<br>****************************************************************
<br>&nbsp;</html>

--------------48E1687CF03AA848973D7833--


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 10:20:46 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA03307
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 10:20:46 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA05842
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 10:20:48 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA02795;
	Tue, 5 Feb 2002 09:43:29 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA02765
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 09:43:27 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA01280
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 09:43:24 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir04nok.ntc.nokia.com (esvir04nokt.ntc.nokia.com [172.21.143.36])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g15EhLZ09665
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 16:43:21 +0200 (EET)
Received: from esebh001.NOE.Nokia.com (unverified) by esvir04nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T58e3962e51ac158f24078@esvir04nok.ntc.nokia.com>;
 Tue, 5 Feb 2002 16:43:12 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh001.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Tue, 5 Feb 2002 16:43:09 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Subject: RE: [Tsvwg] sctp Window deadlock
Content-Type: text/plain;
	charset="iso-8859-1"
Date: Tue, 5 Feb 2002 16:43:08 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF26308EE78@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] sctp Window deadlock
Thread-Index: AcGuUpCN6u7MRFvvRma63SpdJUCyIwAABf5A
To: <arun@netlab.hcltech.com>, <mohammed.asharaf@wipro.com>
Cc: <tsvwg@ietf.org>, <randall@stewart.chicago.il.us>
X-OriginalArrivalTime: 05 Feb 2002 14:43:09.0694 (UTC) FILETIME=[6E2B61E0:01C1AE53]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id JAA02766
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

    Hi all!
 
    Well, I think that the point here is that once the T3 timer expires on an address, all the TSNs that were outstanding on that address are automatically marked for retransmission, i.e, they are not outstanding TSNs any longer...
 
    In this situation, as you are allowed to have at least one TSN outstanding no matter the value of rwnd, you will send the oldest TSN marked for retransmission and restart the T3 again (if possible using another different address than the first time)...
 
    After several failures, you will abort the association... It might take time (RTO.Max is one minute), but at the end you get rid of the association...
 
    However... If the path heartbeat mechanism is enabled, every received HEARTBEAT ACK chunk will erase the retransmission counter... Thus we are in the same problem... But this is because the receiver is not behaving in the expected way...
 
    BR Iván Arias Rodríguez

-----Original Message-----
From: ext Arun Prasad [mailto:arun@netlab.hcltech.com]
Sent: 05. February 2002 16:37
To: Ashraf
Cc: Tsvwg; Randall Stewart
Subject: Re: [Tsvwg] sctp Window deadlock


hi Ashraf, 

From Rfc: 
However, regardless of the value of rwnd (including if it 
      is 0), the data sender can always have one DATA chunk in flight to 
      the receiver if allowed by cwnd 


    My understanding from the above is, trying to send the Data chunk only once 
for probing, and if that chunk is not acknowledged (or acknowledged with rwnd 
as zero), then even when the Rtx timer expires for that Data chunk, the data chunk 
will be queued in the local Q (as the rwnd is zero, and already one data chunk 
has been sent for probing...) and will not be retransimtted..... 
    I dont know whether my understanding is right......  Just getting a feeling that 
Iam going wrong somewhere....( might be ur understanding is right.....) 
Somebody please clarify..... 


Thanks 
-arun 
  


Ashraf wrote: 


Hi, 
    My understanding is ( pls correct me if I am wromg),  each time when the timer expires, the path /assoc retrans counter will be 

incremented and sender keeps on probing  for change in window size  by sending next fragment ( in ur caase fragment B ) and  ending 


up in T3 timer expiry each time. Will the Assoc will be  terminated once this count reached the threshold value.?" 


Thanks, 
ashraf 


Arun Prasad wrote: 


> Hi Ashraf, 
>     ya RTT timer will be started, but upon its expiry, we come to the same problem.. 
> because we cant send the retransmitted packets due to rwnd shortage...... 
> 
> Thanks 
> -arun 
> 
> Ashraf wrote: 
> 
> > Hi arun, 
> >     I think  for the new data chunk ( which we are sending  to probe for change in window), we start RTT timer and  give 
> > appropraite treatment, when it expires. 
> > 
> > thanks, 
> > ashraf 
> > 
> > Arun Prasad wrote: 
> > 
> > > Hi Ashraf, 
> > >     Thanks for the rsp.... 
> > > ya, I agree that partial-delivery will solve the problem... But imagine an errornous 
> > > stack ( or SCTP stack which didnt implement the partial delivery API)  is on the 
> > > receiving side.... Then with the below example scenario, the Sender stack will 
> > > get stuck.... ie.,  the association will never tear down (as RTX timer is not running) 
> > > so the Sender side will be hung (with no mistake on its side....) forever..... 
> > > I think this situation should be avoided...... 
> > > 
> > > Thanks 
> > > -arun 
> > > 
> > > Ashraf wrote: 
> > > 
> > > > Arun, 
> > > >     This is what RFC has to say : 
> > > > "If the data receiver  run out of buffer space while still  waiting for  for more fragments 
> > > > to complete the re-assembly, of the message, it should dispatch part of its inbound message 
> > > > through a partial deliver API, freeing some if its buffer space in order receive  rest of 
> > > > the message" 
> > > > 
> > > > guess this should take care of the deadlock  situation you are mentioning here. 
> > > > 
> > > > Thanks, 
> > > > Mohd. Ashraf 
> > > > Arun Prasad wrote: 
> > > > 
> > > > > Hi all, 
> > > > >     The question might be more of  SCTP implementation specific....   According 
> > > > > to the SCTP rfc2960, the data sender should not send any data packets if it doesnt 
> > > > > have enough Send-Windown (Peer Rwnd).... Just take the below example.... 
> > > > > 
> > > > > The Peer's Rwnd (at the Init time ) is 1500 bytes. 
> > > > > The Sender receivies a data buffer of size 6000 from its upper layer to be sent 
> > > > > to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of 
> > > > > size 1500 bytes (because the local mtu size is 1500bytes).... 
> > > > > Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the 
> > > > > Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with 
> > > > > rwnd as 0. The Sender which would have queued the frags B,C,D... will then 
> > > > > send frag B as the probing data to the Peer... The peer again will respond with 
> > > > > a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the 
> > > > >     * Sender cant send any more data packets as the Peer's rwnd is zero.... 
> > > > >     * Peer cant increase its rwnd till it releases the data chunks which it had already 
> > > > >        received,  but to release that it needs to reassemble the frags (ABCD) and 
> > > > >       deliver it to the above layer, only then it can release the buffers..... Since its 
> > > > >       waiting for the frags C and D...  this will never release the buffers.... so 
> > > > >       will not increase the rwnd..... 
> > > > > 
> > > > > This situation can be avoided if the Sender doesnt sent Data messages of size 
> > > > > greater than the Peer's rwnd.... But the Rfc doesnt say anything about this..... 
> > > > > Or there should be some timers, to get out of the Deadlock situation..... 
> > > > > 
> > > > > Any suggestions........ 
> > > > > 
> > > > > Thanks 
> > > > > -arun 
> > > > > 
> > > > > _______________________________________________ 
> > > > > tsvwg mailing list 
> > > > > tsvwg@ietf.org 
> > > > > https://www1.ietf.org/mailman/listinfo/tsvwg 
> > > > 
> > > >   ------------------------------------------------------------------------------------------------------------------------ 
> > > >                            Name: Wipro_Disclaimer.txt 
> > > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain) 
> > > >                        Encoding: 7bit 
> > > 
> > > -- 
> > > **************************************************************** 
> > > V.Arun Prasad 
> > > HCL Technologies Ltd. 
> > > 51, Jawaharlal Nehru Road, 
> > > Ekkattuthangal, 
> > > Guindy Industrial Estate, 
> > > Chennai - 600097. 
> > > 
> > > Contact # : 9144 - 2334174 
> > >             9144 - 2334181 
> > >             extn : 233 
> > > **************************************************************** 
> > > 
> > > _______________________________________________ 
> > > tsvwg mailing list 
> > > tsvwg@ietf.org 
> > > https://www1.ietf.org/mailman/listinfo/tsvwg 
> > 
> >   ------------------------------------------------------------------------------------------------------------------------ 
> >                            Name: Wipro_Disclaimer.txt 
> >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain) 
> >                        Encoding: 7bit 
> 
> -- 
> **************************************************************** 
> V.Arun Prasad 
> HCL Technologies Ltd. 
> 51, Jawaharlal Nehru Road, 
> Ekkattuthangal, 
> Guindy Industrial Estate, 
> Chennai - 600097. 
> 
> Contact # : 9144 - 2334174 
>             9144 - 2334181 
>             extn : 233 
> **************************************************************** 


  ------------------------------------------------------------------------------------------------------------------------ 
                           Name: Wipro_Disclaimer.txt 
   Wipro_Disclaimer.txt    Type: Plain Text (text/plain) 
                       Encoding: 7bit

-- 
**************************************************************** 
V.Arun Prasad 
HCL Technologies Ltd. 
51, Jawaharlal Nehru Road, 
Ekkattuthangal, 
Guindy Industrial Estate, 
Chennai - 600097. 


Contact # : 9144 - 2334174 
            9144 - 2334181 
            extn : 233 
**************************************************************** 
  


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 10:23:55 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA03458
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 10:23:55 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA06047
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 10:23:57 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA03565;
	Tue, 5 Feb 2002 09:51:22 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA03535
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 09:51:20 -0500 (EST)
Received: from CNRI.Reston.VA.US (localhost [127.0.0.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA01788;
	Tue, 5 Feb 2002 09:51:18 -0500 (EST)
Message-Id: <200202051451.JAA01788@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
CC: tsvwg@ietf.org
From: Internet-Drafts@ietf.org
Reply-to: Internet-Drafts@ietf.org
Date: Tue, 05 Feb 2002 09:51:18 -0500
Subject: [Tsvwg] I-D ACTION:draft-satran-ndt-framing-00.txt,.pdf
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.


	Title		: NDT-Framing
	Author(s)	: J. Satran
	Filename	: draft-satran-ndt-framing-00.txt,.pdf
	Pages		: 
	Date		: 04-Feb-02
	
During the last year  several attempts have been made to provide a 
generalized framing mechanism for protocols using TCP. The main pur-
pose of those mechanisms is to enable hardware network adapter 
builders to discover fast the next Upper Layer Protocol Data Unit 
boundary and thus minimize the amount of memory required on the 
network adapters in order to handle packet drop or reordering in the 
network.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-satran-ndt-framing-00.txt

To remove yourself from the IETF Announcement list, send a message to 
ietf-announce-request with the word unsubscribe in the body of the message.

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-satran-ndt-framing-00.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-satran-ndt-framing-00.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<20020204134144.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-satran-ndt-framing-00.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-satran-ndt-framing-00.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<20020204134144.I-D@ietf.org>

--OtherAccess--

--NextPart--



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 10:51:31 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA05230
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 10:51:27 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA08520
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 10:51:29 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA04756;
	Tue, 5 Feb 2002 10:02:10 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA04667
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 10:02:04 -0500 (EST)
Received: from wiproecmx1.wipro.com (wiproecmx1.wipro.com [164.164.31.5])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA02398
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 10:01:53 -0500 (EST)
Received: from ecvwall1.wipro.com (ecvwall1.wipro.com [164.164.23.6])
	by wiproecmx1.wipro.com (8.11.3/8.11.3) with SMTP id g15F0OX01547
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 20:30:24 +0530 (IST)
Received: from wipro.com ([127.0.0.1]) by ecmail.mail.wipro.com
          (Netscape Messaging Server 4.15) with ESMTP id GR2F1200.QEF;
          Tue, 5 Feb 2002 20:30:39 +0530 
From: "Mohammed Ashraf" <mohammed.asharaf@wipro.com>
To: Arun Prasad <arun@netlab.hcltech.com>
Cc: Tsvwg <tsvwg@ietf.org>, Randall Stewart <randall@stewart.chicago.il.us>
Message-ID: <43ffab439533.43953343ffab@wipro.com>
Date: Tue, 05 Feb 2002 20:00:38 +0500
X-Mailer: Netscape Webmail
MIME-Version: 1.0
Content-Language: en
Subject: Re: [Tsvwg] sctp Window deadlock
X-Accept-Language: en
Content-Type: multipart/mixed; boundary="--1a5476be40297332"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

This is a multi-part message in MIME format.

----1a5476be40297332
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: 7bit

hi arun,
	If this is the case(if we are not retransmitting ), then there is
another possibility of a deadlock, when the SACK that carries window
change get lost in the n/w.i.e sender didn't receive SACK and  believes
that window is still zero and receiver will assume that, its not getting
anymore data because sender doesn't have  any more data to send(this is
juz because there is no reliablity transmission mechanism built into for
SACK:--)))..

Thanks,
ashraf
Content-Type: multipart/alternative;
	 boundary="------------48E1687CF03AA848973D7833"


--------------48E1687CF03AA848973D7833
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

hi Ashraf,

From Rfc:
However, regardless of the value of rwnd (including if it
      is 0), the data sender can always have one DATA chunk in flight to
      the receiver if allowed by cwnd

    My understanding from the above is, trying to send the Data chunk only once
for probing, and if that chunk is not acknowledged (or acknowledged with rwnd
as zero), then even when the Rtx timer expires for that Data chunk, the data chunk
will be queued in the local Q (as the rwnd is zero, and already one data chunk
has been sent for probing...) and will not be retransimtted.....
    I dont know whether my understanding is right......  Just getting a feeling that
Iam going wrong somewhere....( might be ur understanding is right.....)
Somebody please clarify.....

Thanks
-arun


Ashraf wrote:

> Hi,
>     My understanding is ( pls correct me if I am wromg),  each time when the timer expires, the path /assoc retrans counter will be
>
> incremented and sender keeps on probing  for change in window size  by sending next fragment ( in ur caase fragment B ) and  ending
>
> up in T3 timer expiry each time. Will the Assoc will be  terminated once this count reached the threshold value.?"
>
> Thanks,
> ashraf
>
> Arun Prasad wrote:
>
> > Hi Ashraf,
> >     ya RTT timer will be started, but upon its expiry, we come to the same problem..
> > because we cant send the retransmitted packets due to rwnd shortage......
> >
> > Thanks
> > -arun
> >
> > Ashraf wrote:
> >
> > > Hi arun,
> > >     I think  for the new data chunk ( which we are sending  to probe for change in window), we start RTT timer and  give
> > > appropraite treatment, when it expires.
> > >
> > > thanks,
> > > ashraf
> > >
> > > Arun Prasad wrote:
> > >
> > > > Hi Ashraf,
> > > >     Thanks for the rsp....
> > > > ya, I agree that partial-delivery will solve the problem... But imagine an errornous
> > > > stack ( or SCTP stack which didnt implement the partial delivery API)  is on the
> > > > receiving side.... Then with the below example scenario, the Sender stack will
> > > > get stuck.... ie.,  the association will never tear down (as RTX timer is not running)
> > > > so the Sender side will be hung (with no mistake on its side....) forever.....
> > > > I think this situation should be avoided......
> > > >
> > > > Thanks
> > > > -arun
> > > >
> > > > Ashraf wrote:
> > > >
> > > > > Arun,
> > > > >     This is what RFC has to say :
> > > > > "If the data receiver  run out of buffer space while still  waiting for  for more fragments
> > > > > to complete the re-assembly, of the message, it should dispatch part of its inbound message
> > > > > through a partial deliver API, freeing some if its buffer space in order receive  rest of
> > > > > the message"
> > > > >
> > > > > guess this should take care of the deadlock  situation you are mentioning here.
> > > > >
> > > > > Thanks,
> > > > > Mohd. Ashraf
> > > > > Arun Prasad wrote:
> > > > >
> > > > > > Hi all,
> > > > > >     The question might be more of  SCTP implementation specific....   According
> > > > > > to the SCTP rfc2960, the data sender should not send any data packets if it doesnt
> > > > > > have enough Send-Windown (Peer Rwnd).... Just take the below example....
> > > > > >
> > > > > > The Peer's Rwnd (at the Init time ) is 1500 bytes.
> > > > > > The Sender receivies a data buffer of size 6000 from its upper layer to be sent
> > > > > > to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of
> > > > > > size 1500 bytes (because the local mtu size is 1500bytes)....
> > > > > > Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the
> > > > > > Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with
> > > > > > rwnd as 0. The Sender which would have queued the frags B,C,D... will then
> > > > > > send frag B as the probing data to the Peer... The peer again will respond with
> > > > > > a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the
> > > > > >     * Sender cant send any more data packets as the Peer's rwnd is zero....
> > > > > >     * Peer cant increase its rwnd till it releases the data chunks which it had already
> > > > > >        received,  but to release that it needs to reassemble the frags (ABCD) and
> > > > > >       deliver it to the above layer, only then it can release the buffers..... Since its
> > > > > >       waiting for the frags C and D...  this will never release the buffers.... so
> > > > > >       will not increase the rwnd.....
> > > > > >
> > > > > > This situation can be avoided if the Sender doesnt sent Data messages of size
> > > > > > greater than the Peer's rwnd.... But the Rfc doesnt say anything about this.....
> > > > > > Or there should be some timers, to get out of the Deadlock situation.....
> > > > > >
> > > > > > Any suggestions........
> > > > > >
> > > > > > Thanks
> > > > > > -arun
> > > > > >
> > > > > > _______________________________________________
> > > > > > tsvwg mailing list
> > > > > > tsvwg@ietf.org
> > > > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > > >
> > > > >   ------------------------------------------------------------------------------------------------------------------------
> > > > >                            Name: Wipro_Disclaimer.txt
> > > > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > > > >                        Encoding: 7bit
> > > >
> > > > --
> > > > ****************************************************************
> > > > V.Arun Prasad
> > > > HCL Technologies Ltd.
> > > > 51, Jawaharlal Nehru Road,
> > > > Ekkattuthangal,
> > > > Guindy Industrial Estate,
> > > > Chennai - 600097.
> > > >
> > > > Contact # : 9144 - 2334174
> > > >             9144 - 2334181
> > > >             extn : 233
> > > > ****************************************************************
> > > >
> > > > _______________________________________________
> > > > tsvwg mailing list
> > > > tsvwg@ietf.org
> > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > >
> > >   ------------------------------------------------------------------------------------------------------------------------
> > >                            Name: Wipro_Disclaimer.txt
> > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > >                        Encoding: 7bit
> >
> > --
> > ****************************************************************
> > V.Arun Prasad
> > HCL Technologies Ltd.
> > 51, Jawaharlal Nehru Road,
> > Ekkattuthangal,
> > Guindy Industrial Estate,
> > Chennai - 600097.
> >
> > Contact # : 9144 - 2334174
> >             9144 - 2334181
> >             extn : 233
> > ****************************************************************
>
>   ------------------------------------------------------------------------------------------------------------------------
>                            Name: Wipro_Disclaimer.txt
>    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
>                        Encoding: 7bit

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************


--------------48E1687CF03AA848973D7833
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
hi Ashraf,
<p>From Rfc:
<br><font color="#3333FF">However, regardless of the value of rwnd (including
if it</font>
<br><font color="#3333FF">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; is 0), the data
sender can always have <b>one DATA chunk in flight</b> to</font>
<br><font color="#3333FF">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; the receiver if
allowed by cwnd</font>
<p>&nbsp;&nbsp;&nbsp; My understanding from the above is, trying to send
the Data chunk only once
<br>for probing, and if that chunk is not acknowledged (or acknowledged
with rwnd
<br>as zero), then even when the Rtx timer expires for that Data chunk,
the data chunk
<br>will be queued in the local Q (as the rwnd is zero, and already one
data chunk
<br>has been sent for probing...) and will not be retransimtted.....
<br>&nbsp;&nbsp;&nbsp; I dont know whether my understanding is right......&nbsp;
Just getting a feeling that
<br>Iam going wrong somewhere....( might be ur understanding is right.....)
<br>Somebody please clarify.....
<p>Thanks
<br>-arun
<br>&nbsp;
<p>Ashraf wrote:
<blockquote TYPE=CITE>Hi,
<br>&nbsp;&nbsp;&nbsp; My understanding is ( pls correct me if I am wromg),&nbsp;
each time when the timer expires, the path /assoc retrans counter will
be
<p>incremented and sender keeps on probing&nbsp; for change in window size&nbsp;
by sending next fragment ( in ur caase fragment B ) and&nbsp; ending
<p>up in T3 timer expiry each time. Will the Assoc will be&nbsp; terminated
once this count reached the threshold value.?"
<p>Thanks,
<br>ashraf
<p>Arun Prasad wrote:
<p>> Hi Ashraf,
<br>>&nbsp;&nbsp;&nbsp;&nbsp; ya RTT timer will be started, but upon its
expiry, we come to the same problem..
<br>> because we cant send the retransmitted packets due to rwnd shortage......
<br>>
<br>> Thanks
<br>> -arun
<br>>
<br>> Ashraf wrote:
<br>>
<br>> > Hi arun,
<br>> >&nbsp;&nbsp;&nbsp;&nbsp; I think&nbsp; for the new data chunk (
which we are sending&nbsp; to probe for change in window), we start RTT
timer and&nbsp; give
<br>> > appropraite treatment, when it expires.
<br>> >
<br>> > thanks,
<br>> > ashraf
<br>> >
<br>> > Arun Prasad wrote:
<br>> >
<br>> > > Hi Ashraf,
<br>> > >&nbsp;&nbsp;&nbsp;&nbsp; Thanks for the rsp....
<br>> > > ya, I agree that partial-delivery will solve the problem... But
imagine an errornous
<br>> > > stack ( or SCTP stack which didnt implement the partial delivery
API)&nbsp; is on the
<br>> > > receiving side.... Then with the below example scenario, the
Sender stack will
<br>> > > get stuck.... ie.,&nbsp; the association will never tear down
(as RTX timer is not running)
<br>> > > so the Sender side will be hung (with no mistake on its side....)
forever.....
<br>> > > I think this situation should be avoided......
<br>> > >
<br>> > > Thanks
<br>> > > -arun
<br>> > >
<br>> > > Ashraf wrote:
<br>> > >
<br>> > > > Arun,
<br>> > > >&nbsp;&nbsp;&nbsp;&nbsp; This is what RFC has to say :
<br>> > > > "If the data receiver&nbsp; run out of buffer space while still&nbsp;
waiting for&nbsp; for more fragments
<br>> > > > to complete the re-assembly, of the message, it should dispatch
part of its inbound message
<br>> > > > through a partial deliver API, freeing some if its buffer space
in order receive&nbsp; rest of
<br>> > > > the message"
<br>> > > >
<br>> > > > guess this should take care of the deadlock&nbsp; situation
you are mentioning here.
<br>> > > >
<br>> > > > Thanks,
<br>> > > > Mohd. Ashraf
<br>> > > > Arun Prasad wrote:
<br>> > > >
<br>> > > > > Hi all,
<br>> > > > >&nbsp;&nbsp;&nbsp;&nbsp; The question might be more of&nbsp;
SCTP implementation specific....&nbsp;&nbsp; According
<br>> > > > > to the SCTP rfc2960, the data sender should not send any
data packets if it doesnt
<br>> > > > > have enough Send-Windown (Peer Rwnd).... Just take the below
example....
<br>> > > > >
<br>> > > > > The Peer's Rwnd (at the Init time ) is 1500 bytes.
<br>> > > > > The Sender receivies a data buffer of size 6000 from its
upper layer to be sent
<br>> > > > > to the Peer.... The Sender will then fragment the packets
to fragments A,B,C,D of
<br>> > > > > size 1500 bytes (because the local mtu size is 1500bytes)....
<br>> > > > > Now the Sender sends the frag A to the Peer... Upon receiving
the frag A, the
<br>> > > > > Peer's Rwnd will be zero.... This is informed to the Sender
by sending a SACK with
<br>> > > > > rwnd as 0. The Sender which would have queued the frags B,C,D...
will then
<br>> > > > > send frag B as the probing data to the Peer... The peer again
will respond with
<br>> > > > > a SACK of rwnd 0 .&nbsp; Now the deadlock situation occurs,&nbsp;
as the
<br>> > > > >&nbsp;&nbsp;&nbsp;&nbsp; * Sender cant send any more data
packets as the Peer's rwnd is zero....
<br>> > > > >&nbsp;&nbsp;&nbsp;&nbsp; * Peer cant increase its rwnd till
it releases the data chunks which it had already
<br>> > > > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; received,&nbsp;
but to release that it needs to reassemble the frags (ABCD) and
<br>> > > > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; deliver it to the above
layer, only then it can release the buffers..... Since its
<br>> > > > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; waiting for the frags
C and D...&nbsp; this will never release the buffers.... so
<br>> > > > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; will not increase the
rwnd.....
<br>> > > > >
<br>> > > > > This situation can be avoided if the Sender doesnt sent Data
messages of size
<br>> > > > > greater than the Peer's rwnd.... But the Rfc doesnt say anything
about this.....
<br>> > > > > Or there should be some timers, to get out of the Deadlock
situation.....
<br>> > > > >
<br>> > > > > Any suggestions........
<br>> > > > >
<br>> > > > > Thanks
<br>> > > > > -arun
<br>> > > > >
<br>> > > > > _______________________________________________
<br>> > > > > tsvwg mailing list
<br>> > > > > tsvwg@ietf.org
<br>> > > > > <a href="https://www1.ietf.org/mailman/listinfo/tsvwg">https://www1.ietf.org/mailman/listinfo/tsvwg</a>
<br>> > > >
<br>> > > >&nbsp;&nbsp; ------------------------------------------------------------------------------------------------------------------------
<br>> > > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
Name: Wipro_Disclaimer.txt
<br>> > > >&nbsp;&nbsp;&nbsp; Wipro_Disclaimer.txt&nbsp;&nbsp;&nbsp; Type:
Plain Text (text/plain)
<br>> > > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
Encoding: 7bit
<br>> > >
<br>> > > --
<br>> > > ****************************************************************
<br>> > > V.Arun Prasad
<br>> > > HCL Technologies Ltd.
<br>> > > 51, Jawaharlal Nehru Road,
<br>> > > Ekkattuthangal,
<br>> > > Guindy Industrial Estate,
<br>> > > Chennai - 600097.
<br>> > >
<br>> > > Contact # : 9144 - 2334174
<br>> > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
9144 - 2334181
<br>> > >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
extn : 233
<br>> > > ****************************************************************
<br>> > >
<br>> > > _______________________________________________
<br>> > > tsvwg mailing list
<br>> > > tsvwg@ietf.org
<br>> > > <a href="https://www1.ietf.org/mailman/listinfo/tsvwg">https://www1.ietf.org/mailman/listinfo/tsvwg</a>
<br>> >
<br>> >&nbsp;&nbsp; ------------------------------------------------------------------------------------------------------------------------
<br>> >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
Name: Wipro_Disclaimer.txt
<br>> >&nbsp;&nbsp;&nbsp; Wipro_Disclaimer.txt&nbsp;&nbsp;&nbsp; Type:
Plain Text (text/plain)
<br>> >&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
Encoding: 7bit
<br>>
<br>> --
<br>> ****************************************************************
<br>> V.Arun Prasad
<br>> HCL Technologies Ltd.
<br>> 51, Jawaharlal Nehru Road,
<br>> Ekkattuthangal,
<br>> Guindy Industrial Estate,
<br>> Chennai - 600097.
<br>>
<br>> Contact # : 9144 - 2334174
<br>>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
9144 - 2334181
<br>>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
extn : 233
<br>> ****************************************************************
<p>&nbsp; ------------------------------------------------------------------------------------------------------------------------
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
Name: Wipro_Disclaimer.txt
<br>&nbsp;&nbsp; Wipro_Disclaimer.txt&nbsp;&nbsp;&nbsp; Type: Plain Text
(text/plain)
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
Encoding: 7bit</blockquote>

<p>--
<br>****************************************************************
<br>V.Arun Prasad
<br>HCL Technologies Ltd.
<br>51, Jawaharlal Nehru Road,
<br>Ekkattuthangal,
<br>Guindy Industrial Estate,
<br>Chennai - 600097.
<p>Contact # : 9144 - 2334174
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
9144 - 2334181
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
extn : 233
<br>****************************************************************
<br>&nbsp;</html>

--------------48E1687CF03AA848973D7833--


----1a5476be40297332
Content-Type: text/plain;
	name="Wipro_Disclaimer.txt"
Content-Disposition: attachment;
	filename="Wipro_Disclaimer.txt"
Content-Transfer-Encoding: 7bit

**************************Disclaimer************************************
      


Information contained in this E-MAIL being proprietary to Wipro Limited
is 'privileged' and 'confidential' and intended for use only by the
individual or entity to which it is addressed. You are notified that any
use, copying or dissemination of the information contained in the E-MAIL
in any manner whatsoever is strictly prohibited.



 ********************************************************************

----1a5476be40297332--


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 11:31:39 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA06956
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 11:31:34 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA11815
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 11:31:37 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA08468;
	Tue, 5 Feb 2002 10:50:57 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA08440
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 10:50:56 -0500 (EST)
Received: from sj-msg-core-2.cisco.com (sj-msg-core-2.cisco.com [171.69.24.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA05180
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 10:50:50 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-2.cisco.com (8.11.3/8.9.1) with ESMTP id g15FoMb25796;
	Tue, 5 Feb 2002 07:50:23 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id AAZ82475;
	Tue, 5 Feb 2002 07:50:21 -0800 (PST)
Message-ID: <3C5FFF3C.F3BF56D5@stewart.chicago.il.us>
Date: Tue, 05 Feb 2002 09:50:20 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: arun@netlab.hcltech.com, mohammed.asharaf@wipro.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] sctp Window deadlock
References: <58AEDE27D48F884285007ABF827FF26308EE78@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Ivan:

I agree with all your assesement except I would NOT count
the T3 against anything towards failure (no matter if
HB's are running or not) if we are doing widow-probes.

I run special code for this when we get in the window-probe
mode... And no threshold managment is done... 

This is the same thing that happens in TCP (if my understanding
is correct).. i.e. once the sender clamps the rwnd=0 you
will continue to probe until it opens up.. as long as
you get a SACK back that is...

R


Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>     Hi all!
> 
>     Well, I think that the point here is that once the T3 timer expires on an address, all the TSNs that were outstanding on that address are automatically marked for retransmission, i.e, they are not outstanding TSNs any longer...
> 
>     In this situation, as you are allowed to have at least one TSN outstanding no matter the value of rwnd, you will send the oldest TSN marked for retransmission and restart the T3 again (if possible using another different address than the first time)...
> 
>     After several failures, you will abort the association... It might take time (RTO.Max is one minute), but at the end you get rid of the association...
> 
>     However... If the path heartbeat mechanism is enabled, every received HEARTBEAT ACK chunk will erase the retransmission counter... Thus we are in the same problem... But this is because the receiver is not behaving in the expected way...
> 
>     BR Iván Arias Rodríguez
> 
> -----Original Message-----
> From: ext Arun Prasad [mailto:arun@netlab.hcltech.com]
> Sent: 05. February 2002 16:37
> To: Ashraf
> Cc: Tsvwg; Randall Stewart
> Subject: Re: [Tsvwg] sctp Window deadlock
> 
> hi Ashraf,
> 
> >From Rfc:
> However, regardless of the value of rwnd (including if it
>       is 0), the data sender can always have one DATA chunk in flight to
>       the receiver if allowed by cwnd
> 
>     My understanding from the above is, trying to send the Data chunk only once
> for probing, and if that chunk is not acknowledged (or acknowledged with rwnd
> as zero), then even when the Rtx timer expires for that Data chunk, the data chunk
> will be queued in the local Q (as the rwnd is zero, and already one data chunk
> has been sent for probing...) and will not be retransimtted.....
>     I dont know whether my understanding is right......  Just getting a feeling that
> Iam going wrong somewhere....( might be ur understanding is right.....)
> Somebody please clarify.....
> 
> Thanks
> -arun
> 
> 
> Ashraf wrote:
> 
> Hi,
>     My understanding is ( pls correct me if I am wromg),  each time when the timer expires, the path /assoc retrans counter will be
> 
> incremented and sender keeps on probing  for change in window size  by sending next fragment ( in ur caase fragment B ) and  ending
> 
> up in T3 timer expiry each time. Will the Assoc will be  terminated once this count reached the threshold value.?"
> 
> Thanks,
> ashraf
> 
> Arun Prasad wrote:
> 
> > Hi Ashraf,
> >     ya RTT timer will be started, but upon its expiry, we come to the same problem..
> > because we cant send the retransmitted packets due to rwnd shortage......
> >
> > Thanks
> > -arun
> >
> > Ashraf wrote:
> >
> > > Hi arun,
> > >     I think  for the new data chunk ( which we are sending  to probe for change in window), we start RTT timer and  give
> > > appropraite treatment, when it expires.
> > >
> > > thanks,
> > > ashraf
> > >
> > > Arun Prasad wrote:
> > >
> > > > Hi Ashraf,
> > > >     Thanks for the rsp....
> > > > ya, I agree that partial-delivery will solve the problem... But imagine an errornous
> > > > stack ( or SCTP stack which didnt implement the partial delivery API)  is on the
> > > > receiving side.... Then with the below example scenario, the Sender stack will
> > > > get stuck.... ie.,  the association will never tear down (as RTX timer is not running)
> > > > so the Sender side will be hung (with no mistake on its side....) forever.....
> > > > I think this situation should be avoided......
> > > >
> > > > Thanks
> > > > -arun
> > > >
> > > > Ashraf wrote:
> > > >
> > > > > Arun,
> > > > >     This is what RFC has to say :
> > > > > "If the data receiver  run out of buffer space while still  waiting for  for more fragments
> > > > > to complete the re-assembly, of the message, it should dispatch part of its inbound message
> > > > > through a partial deliver API, freeing some if its buffer space in order receive  rest of
> > > > > the message"
> > > > >
> > > > > guess this should take care of the deadlock  situation you are mentioning here.
> > > > >
> > > > > Thanks,
> > > > > Mohd. Ashraf
> > > > > Arun Prasad wrote:
> > > > >
> > > > > > Hi all,
> > > > > >     The question might be more of  SCTP implementation specific....   According
> > > > > > to the SCTP rfc2960, the data sender should not send any data packets if it doesnt
> > > > > > have enough Send-Windown (Peer Rwnd).... Just take the below example....
> > > > > >
> > > > > > The Peer's Rwnd (at the Init time ) is 1500 bytes.
> > > > > > The Sender receivies a data buffer of size 6000 from its upper layer to be sent
> > > > > > to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of
> > > > > > size 1500 bytes (because the local mtu size is 1500bytes)....
> > > > > > Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the
> > > > > > Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with
> > > > > > rwnd as 0. The Sender which would have queued the frags B,C,D... will then
> > > > > > send frag B as the probing data to the Peer... The peer again will respond with
> > > > > > a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the
> > > > > >     * Sender cant send any more data packets as the Peer's rwnd is zero....
> > > > > >     * Peer cant increase its rwnd till it releases the data chunks which it had already
> > > > > >        received,  but to release that it needs to reassemble the frags (ABCD) and
> > > > > >       deliver it to the above layer, only then it can release the buffers..... Since its
> > > > > >       waiting for the frags C and D...  this will never release the buffers.... so
> > > > > >       will not increase the rwnd.....
> > > > > >
> > > > > > This situation can be avoided if the Sender doesnt sent Data messages of size
> > > > > > greater than the Peer's rwnd.... But the Rfc doesnt say anything about this.....
> > > > > > Or there should be some timers, to get out of the Deadlock situation.....
> > > > > >
> > > > > > Any suggestions........
> > > > > >
> > > > > > Thanks
> > > > > > -arun
> > > > > >
> > > > > > _______________________________________________
> > > > > > tsvwg mailing list
> > > > > > tsvwg@ietf.org
> > > > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > > >
> > > > >   ------------------------------------------------------------------------------------------------------------------------
> > > > >                            Name: Wipro_Disclaimer.txt
> > > > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > > > >                        Encoding: 7bit
> > > >
> > > > --
> > > > ****************************************************************
> > > > V.Arun Prasad
> > > > HCL Technologies Ltd.
> > > > 51, Jawaharlal Nehru Road,
> > > > Ekkattuthangal,
> > > > Guindy Industrial Estate,
> > > > Chennai - 600097.
> > > >
> > > > Contact # : 9144 - 2334174
> > > >             9144 - 2334181
> > > >             extn : 233
> > > > ****************************************************************
> > > >
> > > > _______________________________________________
> > > > tsvwg mailing list
> > > > tsvwg@ietf.org
> > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > >
> > >   ------------------------------------------------------------------------------------------------------------------------
> > >                            Name: Wipro_Disclaimer.txt
> > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > >                        Encoding: 7bit
> >
> > --
> > ****************************************************************
> > V.Arun Prasad
> > HCL Technologies Ltd.
> > 51, Jawaharlal Nehru Road,
> > Ekkattuthangal,
> > Guindy Industrial Estate,
> > Chennai - 600097.
> >
> > Contact # : 9144 - 2334174
> >             9144 - 2334181
> >             extn : 233
> > ****************************************************************
> 
>   ------------------------------------------------------------------------------------------------------------------------
>                            Name: Wipro_Disclaimer.txt
>    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
>                        Encoding: 7bit
> 
> --
> ****************************************************************
> V.Arun Prasad
> HCL Technologies Ltd.
> 51, Jawaharlal Nehru Road,
> Ekkattuthangal,
> Guindy Industrial Estate,
> Chennai - 600097.
> 
> Contact # : 9144 - 2334174
>             9144 - 2334181
>             extn : 233
> ****************************************************************
> 

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 11:41:54 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA07418
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 11:41:54 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA13090
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 11:41:56 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA08094;
	Tue, 5 Feb 2002 10:46:03 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA08063
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 10:46:01 -0500 (EST)
Received: from sj-msg-core-2.cisco.com (sj-msg-core-2.cisco.com [171.69.24.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA04873
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 10:45:57 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-2.cisco.com (8.11.3/8.9.1) with ESMTP id g15Fimb22109;
	Tue, 5 Feb 2002 07:44:48 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id AAZ82370;
	Tue, 5 Feb 2002 07:44:46 -0800 (PST)
Message-ID: <3C5FFDEE.93CD4573@stewart.chicago.il.us>
Date: Tue, 05 Feb 2002 09:44:46 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Arun Prasad <arun@netlab.hcltech.com>
CC: Ashraf <mohammed.asharaf@wipro.com>, Tsvwg <tsvwg@ietf.org>,
        sachin <sachins@netlab.hcltech.com>
Subject: Re: [Tsvwg] sctp Window deadlock
References: <NEBBJGDMMLHHCIKHGBEJMEGGCPAA.dotis@sanlight.net> <3C5F7C39.B86F4150@netlab.hcltech.com> <3C5F8128.CABB3467@wipro.com> <3C5F95BA.923DEE3F@netlab.hcltech.com> <3C5FC314.D15DF4F9@stewart.chicago.il.us> <3C5FD19D.3D0308A8@netlab.hcltech.com> <3C5FD920.43D82001@stewart.chicago.il.us> <3C5FEBEB.A79EB5AD@netlab.hcltech.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

You always can probe with one chunk. However when the sack
comes back saying rwnd is still 0.. you are still in the
probe mode. So at the next expiration of the T3 you will
resend the probe chunk.... until the rwnd is raised or
you give up :-)

I don't see that any failure threshold will enter into
this picture...

R

Arun Prasad wrote:
> 
> Hi Randell,
>     I think we misunderstood something.....  The scenario we are considering will
> not send the data packet to the network at all.... When  the Peer- rwnd is zero,
> one data packet is sent for probing,  then if the endpoint still receives the SACK
> with rwnd as 0, then the stack will not send any packets to the network, instead
> will Queue the data chunks in the local Queue and will be waiting for the
> Peer-rwnd to increase (I think only one probing data chunk is allowed when
> the Peer-rwnd is zero...) . So the solution u are mentioning ( the
> MAX-ASSOC-RETRANS counter ) will not help because the counter will
> be increamented only when the data chunk is sent to the network.....
> Hope Iam right.......
> 
> Thanks
> -arun
> 
> Randall Stewart wrote:
> 
> > Arun:
> >
> > Interesting idea... but this will not help you. The lifetime
> > as specified can only effect packets that have NOT been
> > transmitted.
> >
> > As I said.. some implementations MAY take and implement an
> > overall number of times a chunk may be transmitted.. and if
> > you keep getting it retran'd and no ack.. you get an abort :>
> >
> > The u-sctp draft had something more like what you are considering
> > below.. but that is not currently an active document in the IETF.
> >
> > There are some of us still playing with it and you can find the
> > last updated document on the web site
> >
> > http://www.sctp.org
> >
> > if you are interested...
> >
> > R
> >
> > Arun Prasad wrote:
> > >
> > > Hi Randall,
> > >     Considering your 4th point about logging....
> > >         I think the below scenario can be better handled
> > > if the implementation of "Life_Time" (The user specifying for how long the particular
> > > packet is to be tried to send) parameter in the SEND primitive  is present.
> > > So once the timer (Life_time timer) expires, the user can be indicated that the
> > > data cannot be sent (again for each each data request, there should be a timer
> > > might be costly...)
> > > This way atleast we may come to know that something wrong is going on in the
> > > stack, and the user may choose to send a "Shutdown" primitive...
> > > ( I guess all the optional parameters also has some basic purpose.......)
> > >
> > > Thanks
> > > -arun
> > >
> > > Randall Stewart wrote:
> > >
> > > > Arun:
> > > >
> > > > Some comments:
> > > >
> > > > Arun Prasad wrote:
> > > > >
> > > > > Hi Ashraf,
> > > > >     Thanks for the rsp....
> > > > > ya, I agree that partial-delivery will solve the problem... But imagine an errornous
> > > > > stack ( or SCTP stack which didnt implement the partial delivery API)  is on the
> > > > > receiving side.... Then with the below example scenario, the Sender stack will
> > > > > get stuck.... ie.,  the association will never tear down (as RTX timer is not running)
> > > > > so the Sender side will be hung (with no mistake on its side....) forever.....
> > > > > I think this situation should be avoided......
> > > >
> > > > I will address this comment below ...
> > > >
> > > > >
> > > > > Thanks
> > > > > -arun
> > > > >
> > > > > Ashraf wrote:
> > > > >
> > > > > > Arun,
> > > > > >     This is what RFC has to say :
> > > > > > "If the data receiver  run out of buffer space while still  waiting for  for more fragments
> > > > > > to complete the re-assembly, of the message, it should dispatch part of its inbound message
> > > > > > through a partial deliver API, freeing some if its buffer space in order receive  rest of
> > > > > > the message"
> > > > > >
> > > > > > guess this should take care of the deadlock  situation you are mentioning here.
> > > > > >
> > > > > > Thanks,
> > > > > > Mohd. Ashraf
> > > > > > Arun Prasad wrote:
> > > >
> > > > The above is percisely what we designed the PD-API for. It is
> > > > a part I never really much liked :> but we saw it to be the
> > > > only way out of such a dead-lock.... Also teamed with this
> > > > is the receivers right to reneg. This way in a critical situation
> > > > it can drop some of the last chunks it is holding so it can
> > > > make room for the earlier pieces so it CAN use the PD-API...
> > > >
> > > > Both are necessary to ensure no deadlock can occur..
> > > >
> > > > Now as to the other issue.. what happens if your sender
> > > > is broken or an attacker.. a number of things come to mind
> > > > (as yet I have not had my coffee so there may be more :0)
> > > >
> > > > 1) TCP I would imagine can be in the same situation .. i.e.
> > > >    if it connects to a peer that is intentionally doing something
> > > >    bad.
> > > >
> > > > 2) It really becomes implementation dependant. I know of one
> > > > implementation
> > > >    that will only keep sending the same chunk for so long.. and then
> > > >    it will abort the associaition.
> > > >
> > > > 3) If the guy is really an attacker the problem is more in to the fact
> > > >    that you have entered into the association with him. I.e don't
> > > >    leave the front door open and expect a burgular not to enter...
> > > >
> > > > 4) If it is a broken implementation then it needs to get fixed and I
> > > >    would hope the one stuck would at least have enough logging going
> > > >    on to show that it is broken... (or even in the case 3 I think logs
> > > >    may help) so someone can fix it..
> > > >
> > > > R
> > > >
> > > > > >
> > > > > > > Hi all,
> > > > > > >     The question might be more of  SCTP implementation specific....   According
> > > > > > > to the SCTP rfc2960, the data sender should not send any data packets if it doesnt
> > > > > > > have enough Send-Windown (Peer Rwnd).... Just take the below example....
> > > > > > >
> > > > > > > The Peer's Rwnd (at the Init time ) is 1500 bytes.
> > > > > > > The Sender receivies a data buffer of size 6000 from its upper layer to be sent
> > > > > > > to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of
> > > > > > > size 1500 bytes (because the local mtu size is 1500bytes)....
> > > > > > > Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the
> > > > > > > Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with
> > > > > > > rwnd as 0. The Sender which would have queued the frags B,C,D... will then
> > > > > > > send frag B as the probing data to the Peer... The peer again will respond with
> > > > > > > a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the
> > > > > > >     * Sender cant send any more data packets as the Peer's rwnd is zero....
> > > > > > >     * Peer cant increase its rwnd till it releases the data chunks which it had already
> > > > > > >        received,  but to release that it needs to reassemble the frags (ABCD) and
> > > > > > >       deliver it to the above layer, only then it can release the buffers..... Since its
> > > > > > >       waiting for the frags C and D...  this will never release the buffers.... so
> > > > > > >       will not increase the rwnd.....
> > > > > > >
> > > > > > > This situation can be avoided if the Sender doesnt sent Data messages of size
> > > > > > > greater than the Peer's rwnd.... But the Rfc doesnt say anything about this.....
> > > > > > > Or there should be some timers, to get out of the Deadlock situation.....
> > > > > > >
> > > > > > > Any suggestions........
> > > > > > >
> > > > > > > Thanks
> > > > > > > -arun
> > > > > > >
> > > > > > > _______________________________________________
> > > > > > > tsvwg mailing list
> > > > > > > tsvwg@ietf.org
> > > > > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > > > >
> > > > > >   ------------------------------------------------------------------------------------------------------------------------
> > > > > >                            Name: Wipro_Disclaimer.txt
> > > > > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > > > > >                        Encoding: 7bit
> > > > >
> > > > > --
> > > > > ****************************************************************
> > > > > V.Arun Prasad
> > > > > HCL Technologies Ltd.
> > > > > 51, Jawaharlal Nehru Road,
> > > > > Ekkattuthangal,
> > > > > Guindy Industrial Estate,
> > > > > Chennai - 600097.
> > > > >
> > > > > Contact # : 9144 - 2334174
> > > > >             9144 - 2334181
> > > > >             extn : 233
> > > > > ****************************************************************
> > > >
> > > > --
> > > > Randall R. Stewart
> > > > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> > >
> > > --
> > > ****************************************************************
> > > V.Arun Prasad
> > > HCL Technologies Ltd.
> > > 51, Jawaharlal Nehru Road,
> > > Ekkattuthangal,
> > > Guindy Industrial Estate,
> > > Chennai - 600097.
> > >
> > > Contact # : 9144 - 2334174
> > >             9144 - 2334181
> > >             extn : 233
> > > ****************************************************************
> >
> > --
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> --
> ****************************************************************
> V.Arun Prasad
> HCL Technologies Ltd.
> 51, Jawaharlal Nehru Road,
> Ekkattuthangal,
> Guindy Industrial Estate,
> Chennai - 600097.
> 
> Contact # : 9144 - 2334174
>             9144 - 2334181
>             extn : 233
> ****************************************************************
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb  5 12:05:52 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA08575
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 12:05:47 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id MAA17101
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 12:05:50 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA09974;
	Tue, 5 Feb 2002 11:08:00 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA09899
	for <tsvwg@optimus.ietf.org>; Tue, 5 Feb 2002 11:07:55 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA06058
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 11:07:51 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir04nok.ntc.nokia.com (esvir04nokt.ntc.nokia.com [172.21.143.36])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g15G7pZ26890
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 18:07:51 +0200 (EET)
Received: from esebh002.NOE.Nokia.com (unverified) by esvir04nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T58e3e3953aac158f24078@esvir04nok.ntc.nokia.com>;
 Tue, 5 Feb 2002 18:07:44 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh002.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Tue, 5 Feb 2002 18:07:45 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Subject: RE: [Tsvwg] sctp Window deadlock
Content-Type: text/plain;
	charset="iso-8859-1"
Date: Tue, 5 Feb 2002 18:07:44 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF26308EE7A@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] sctp Window deadlock
Thread-Index: AcGuXP8L0AgRfCOOS++cKC7tX6zNRwAAPiLQ
To: <randall@stewart.chicago.il.us>
Cc: <arun@netlab.hcltech.com>, <mohammed.asharaf@wipro.com>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 05 Feb 2002 16:07:45.0035 (UTC) FILETIME=[3F4EF5B0:01C1AE5F]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id LAA09900
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Ups! this was even clarified in the last version of the Implementors Guide... :-O I should take some more coffee...

	So the point here is that the receiving side will keep open the association as long as it wants... So, it is the upper user at the sending side the one in charge of tearing down the association if it is not advancing... In any case, it is the upper user at the receiver side the one that is not taking the data (or the SCTP implementation does not have the partial delivery API, thus taking the risk of being in this situation)...

	BR Iván Arias Rodríguez

> -----Original Message-----
> From: ext Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: 05. February 2002 17:50
> To: Arias-Rodriguez Ivan (NRC/Helsinki)
> Cc: arun@netlab.hcltech.com; mohammed.asharaf@wipro.com; 
> tsvwg@ietf.org
> Subject: Re: [Tsvwg] sctp Window deadlock
> 
> 
> Ivan:
> 
> I agree with all your assesement except I would NOT count
> the T3 against anything towards failure (no matter if
> HB's are running or not) if we are doing widow-probes.
> 
> I run special code for this when we get in the window-probe
> mode... And no threshold managment is done... 
> 
> This is the same thing that happens in TCP (if my understanding
> is correct).. i.e. once the sender clamps the rwnd=0 you
> will continue to probe until it opens up.. as long as
> you get a SACK back that is...
> 
> R
> 
> 
> Ivan.Arias-Rodriguez@nokia.com wrote:
> > 
> >     Hi all!
> > 
> >     Well, I think that the point here is that once the T3 
> timer expires on an address, all the TSNs that were 
> outstanding on that address are automatically marked for 
> retransmission, i.e, they are not outstanding TSNs any longer...
> > 
> >     In this situation, as you are allowed to have at least 
> one TSN outstanding no matter the value of rwnd, you will 
> send the oldest TSN marked for retransmission and restart the 
> T3 again (if possible using another different address than 
> the first time)...
> > 
> >     After several failures, you will abort the 
> association... It might take time (RTO.Max is one minute), 
> but at the end you get rid of the association...
> > 
> >     However... If the path heartbeat mechanism is enabled, 
> every received HEARTBEAT ACK chunk will erase the 
> retransmission counter... Thus we are in the same problem... 
> But this is because the receiver is not behaving in the 
> expected way...
> > 
> >     BR Iván Arias Rodríguez
> > 
> > -----Original Message-----
> > From: ext Arun Prasad [mailto:arun@netlab.hcltech.com]
> > Sent: 05. February 2002 16:37
> > To: Ashraf
> > Cc: Tsvwg; Randall Stewart
> > Subject: Re: [Tsvwg] sctp Window deadlock
> > 
> > hi Ashraf,
> > 
> > >From Rfc:
> > However, regardless of the value of rwnd (including if it
> >       is 0), the data sender can always have one DATA chunk 
> in flight to
> >       the receiver if allowed by cwnd
> > 
> >     My understanding from the above is, trying to send the 
> Data chunk only once
> > for probing, and if that chunk is not acknowledged (or 
> acknowledged with rwnd
> > as zero), then even when the Rtx timer expires for that 
> Data chunk, the data chunk
> > will be queued in the local Q (as the rwnd is zero, and 
> already one data chunk
> > has been sent for probing...) and will not be retransimtted.....
> >     I dont know whether my understanding is right......  
> Just getting a feeling that
> > Iam going wrong somewhere....( might be ur understanding is 
> right.....)
> > Somebody please clarify.....
> > 
> > Thanks
> > -arun
> > 
> > 
> > Ashraf wrote:
> > 
> > Hi,
> >     My understanding is ( pls correct me if I am wromg),  
> each time when the timer expires, the path /assoc retrans 
> counter will be
> > 
> > incremented and sender keeps on probing  for change in 
> window size  by sending next fragment ( in ur caase fragment 
> B ) and  ending
> > 
> > up in T3 timer expiry each time. Will the Assoc will be  
> terminated once this count reached the threshold value.?"
> > 
> > Thanks,
> > ashraf
> > 
> > Arun Prasad wrote:
> > 
> > > Hi Ashraf,
> > >     ya RTT timer will be started, but upon its expiry, we 
> come to the same problem..
> > > because we cant send the retransmitted packets due to 
> rwnd shortage......
> > >
> > > Thanks
> > > -arun
> > >
> > > Ashraf wrote:
> > >
> > > > Hi arun,
> > > >     I think  for the new data chunk ( which we are 
> sending  to probe for change in window), we start RTT timer and  give
> > > > appropraite treatment, when it expires.
> > > >
> > > > thanks,
> > > > ashraf
> > > >
> > > > Arun Prasad wrote:
> > > >
> > > > > Hi Ashraf,
> > > > >     Thanks for the rsp....
> > > > > ya, I agree that partial-delivery will solve the 
> problem... But imagine an errornous
> > > > > stack ( or SCTP stack which didnt implement the 
> partial delivery API)  is on the
> > > > > receiving side.... Then with the below example 
> scenario, the Sender stack will
> > > > > get stuck.... ie.,  the association will never tear 
> down (as RTX timer is not running)
> > > > > so the Sender side will be hung (with no mistake on 
> its side....) forever.....
> > > > > I think this situation should be avoided......
> > > > >
> > > > > Thanks
> > > > > -arun
> > > > >
> > > > > Ashraf wrote:
> > > > >
> > > > > > Arun,
> > > > > >     This is what RFC has to say :
> > > > > > "If the data receiver  run out of buffer space 
> while still  waiting for  for more fragments
> > > > > > to complete the re-assembly, of the message, it 
> should dispatch part of its inbound message
> > > > > > through a partial deliver API, freeing some if its 
> buffer space in order receive  rest of
> > > > > > the message"
> > > > > >
> > > > > > guess this should take care of the deadlock  
> situation you are mentioning here.
> > > > > >
> > > > > > Thanks,
> > > > > > Mohd. Ashraf
> > > > > > Arun Prasad wrote:
> > > > > >
> > > > > > > Hi all,
> > > > > > >     The question might be more of  SCTP 
> implementation specific....   According
> > > > > > > to the SCTP rfc2960, the data sender should not 
> send any data packets if it doesnt
> > > > > > > have enough Send-Windown (Peer Rwnd).... Just 
> take the below example....
> > > > > > >
> > > > > > > The Peer's Rwnd (at the Init time ) is 1500 bytes.
> > > > > > > The Sender receivies a data buffer of size 6000 
> from its upper layer to be sent
> > > > > > > to the Peer.... The Sender will then fragment the 
> packets to fragments A,B,C,D of
> > > > > > > size 1500 bytes (because the local mtu size is 
> 1500bytes)....
> > > > > > > Now the Sender sends the frag A to the Peer... 
> Upon receiving the frag A, the
> > > > > > > Peer's Rwnd will be zero.... This is informed to 
> the Sender by sending a SACK with
> > > > > > > rwnd as 0. The Sender which would have queued the 
> frags B,C,D... will then
> > > > > > > send frag B as the probing data to the Peer... 
> The peer again will respond with
> > > > > > > a SACK of rwnd 0 .  Now the deadlock situation 
> occurs,  as the
> > > > > > >     * Sender cant send any more data packets as 
> the Peer's rwnd is zero....
> > > > > > >     * Peer cant increase its rwnd till it 
> releases the data chunks which it had already
> > > > > > >        received,  but to release that it needs to 
> reassemble the frags (ABCD) and
> > > > > > >       deliver it to the above layer, only then it 
> can release the buffers..... Since its
> > > > > > >       waiting for the frags C and D...  this will 
> never release the buffers.... so
> > > > > > >       will not increase the rwnd.....
> > > > > > >
> > > > > > > This situation can be avoided if the Sender 
> doesnt sent Data messages of size
> > > > > > > greater than the Peer's rwnd.... But the Rfc 
> doesnt say anything about this.....
> > > > > > > Or there should be some timers, to get out of the 
> Deadlock situation.....
> > > > > > >
> > > > > > > Any suggestions........
> > > > > > >
> > > > > > > Thanks
> > > > > > > -arun
> > > > > > >
> > > > > > > _______________________________________________
> > > > > > > tsvwg mailing list
> > > > > > > tsvwg@ietf.org
> > > > > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > > > >
> > > > > >   
> --------------------------------------------------------------
> ----------------------------------------------------------
> > > > > >                            Name: Wipro_Disclaimer.txt
> > > > > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > > > > >                        Encoding: 7bit
> > > > >
> > > > > --
> > > > > 
> ****************************************************************
> > > > > V.Arun Prasad
> > > > > HCL Technologies Ltd.
> > > > > 51, Jawaharlal Nehru Road,
> > > > > Ekkattuthangal,
> > > > > Guindy Industrial Estate,
> > > > > Chennai - 600097.
> > > > >
> > > > > Contact # : 9144 - 2334174
> > > > >             9144 - 2334181
> > > > >             extn : 233
> > > > > 
> ****************************************************************
> > > > >
> > > > > _______________________________________________
> > > > > tsvwg mailing list
> > > > > tsvwg@ietf.org
> > > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > >
> > > >   
> --------------------------------------------------------------
> ----------------------------------------------------------
> > > >                            Name: Wipro_Disclaimer.txt
> > > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > > >                        Encoding: 7bit
> > >
> > > --
> > > ****************************************************************
> > > V.Arun Prasad
> > > HCL Technologies Ltd.
> > > 51, Jawaharlal Nehru Road,
> > > Ekkattuthangal,
> > > Guindy Industrial Estate,
> > > Chennai - 600097.
> > >
> > > Contact # : 9144 - 2334174
> > >             9144 - 2334181
> > >             extn : 233
> > > ****************************************************************
> > 
> >   
> --------------------------------------------------------------
> ----------------------------------------------------------
> >                            Name: Wipro_Disclaimer.txt
> >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> >                        Encoding: 7bit
> > 
> > --
> > ****************************************************************
> > V.Arun Prasad
> > HCL Technologies Ltd.
> > 51, Jawaharlal Nehru Road,
> > Ekkattuthangal,
> > Guindy Industrial Estate,
> > Chennai - 600097.
> > 
> > Contact # : 9144 - 2334174
> >             9144 - 2334181
> >             extn : 233
> > ****************************************************************
> > 
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Feb  5 14:45:22 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA16098
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 14:45:22 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA28099
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 14:45:24 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA25467;
	Tue, 5 Feb 2002 14:00:09 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA25438
	for <tsvwg@ns.ietf.org>; Tue, 5 Feb 2002 14:00:07 -0500 (EST)
Received: from c003.snv.cp.net (c003-h015.c003.snv.cp.net [209.228.32.229])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id OAA13838
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 14:00:04 -0500 (EST)
Received: (cpmta 12302 invoked from network); 5 Feb 2002 10:59:35 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.229) with SMTP; 5 Feb 2002 10:59:35 -0800
X-Sent: 5 Feb 2002 18:59:35 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <ips@ece.cmu.edu>
Cc: "Tsvwg" <tsvwg@ietf.org>
Date: Tue, 5 Feb 2002 11:01:22 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJCEHACPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
In-Reply-To: <OF76D5DBB7.4C704991-ONC2256B57.005D8A10@telaviv.ibm.com>
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] RE: iSCSI: No Framing
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Julian,

Discussion regarding the concept of framing using FIM within iSCSI is a
valid topic at this time.  I disagree strongly with your statement about the
market not becoming fractured.  With this topic to be decided shortly, now
is my opportunity to speak.

A generalized scheme for implementing DDP is relevant as opposed to one that
is proprietary and application specific.  Yes, I know it is using the four
letter word SCTP.  There is about to be a document released illustrating DDP
using SCTP.  It removes the application specific nature of this endeavor and
does not impact existing TCP implementations or network layering.  SCTP
provides a far superior solution when attempting Direct Data Placement, the
rational for FIM within iSCSI.  iSCSI APIs need not change to introduce SCTP
as an alternative transport if implementing this DDP feature in the future.
To prevent fracturing of the market and creating standards competition
within IETF, FIM should be removed from the iSCSI draft.  Transition to SCTP
for this feature and do not add layers to TCP.

Doug

David,

Dough is continuing his rumblings - whether they are relevant to us or not.
I wonder if there is some administrative measure you can take to save us
some bandwidth or we should install our own
"Otis filters".

Julo

Julian,

Implementing a scheme for doing direct data placement by means of an
application specific, undocumented or proprietary layer below TCP will
fracture the market.  This will result in an unacceptable outcome having
hardware created for each application, limited by various vendors making as
yet unknown claims, if implementing the DDP feature using TCP.  In the
future, should this feature become desired, iSCSI can adopt a generic method
that is possible through the use of SCTP.  An eventual adoption of SCTP
would also simplify much of the existing complexity found with the
multi-connection TCP standard now proposed for iSCSI.  Use of SCTP would
allow common hardware for many protocols desiring Direct Data Placement
without modification of the SCTP transport as it can deliver objects
unordered.  A shim providing the DDP feature could ensure objects are
disclosed in the order sent if desired, independent of the reception at the
shim while not modifying SCTP or adding layers beneath this transport as
would be required for TCP.

Only when constrained to conventional memory allocation, does DDP become
beneficial.  The concept of placing a layer beneath TCP that structures an
array to hold each individual packet based on application specific
structures is fundamentally flawed.  This presumes the layer beneath TCP
knows before hand the content of the byte stream prior to TCP composing the
stream.  This lower layer will also need to create exceptions for handling
duplicates, for placement information not aligned with the TCP segment, and
for application informational errors.  SCTP can easily avoid these
exceptional cases making such development less disruptive.  In addition, an
SCTP stack that does not offer DDP in hardware need not change the API
behavior to the application nor modify any layer functionality.

The ability of SCTP to support many different higher level protocols comes
from the capability of SCTP to deliver objects unordered.  TCP's restriction
on ordered delivery mandates that any direct data placement is done prior to
TCP through the addition of a new and application unique layer between IP
and TCP.  The saving caveat has been that disclosure to TCP of these packets
are sequential but the application interface to TCP will change as a result
of this modification in some undefined way to associate the steering
information with the desired buffers.  This change impacts the interface
between this new layer, as well as the application and TCP.  The desire to
keep this DDP scheme for TCP proprietary will ensure a fractured market in
terms of hardware, OS services, and perhaps interchange not to mention that
each application then also needs unique hardware.  The security risks
involved in creating this new and highly complex layer below TCP is yet
unknown as details are lacking.  In essence, an attempt to implement DDP by
creating a new layer beneath TCP in a proprietary fashion is in direct
competition with the far better practice of using SCTP to implement DDP.

Doug


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Feb  5 14:54:06 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA16443
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 14:54:05 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA28357
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 14:54:08 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA26050;
	Tue, 5 Feb 2002 14:02:54 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA26019
	for <tsvwg@ns.ietf.org>; Tue, 5 Feb 2002 14:02:51 -0500 (EST)
Received: from mgw-x1.nokia.com (mgw-x1.nokia.com [131.228.20.21])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA13995
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 14:02:31 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir05nok.ntc.nokia.com (esvir05nokt.ntc.nokia.com [172.21.143.37])
	by mgw-x1.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g15J1r328904
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 21:01:53 +0200 (EET)
Received: from esebh001.NOE.Nokia.com (unverified) by esvir05nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T58e4835b82ac158f257b6@esvir05nok.ntc.nokia.com> for <tsvwg@ietf.org>;
 Tue, 5 Feb 2002 21:02:15 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh001.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Tue, 5 Feb 2002 21:01:57 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Date: Tue, 5 Feb 2002 21:01:57 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF26308EE7C@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] sctp Window deadlock
Thread-Index: AcGuXP8L0AgRfCOOS++cKC7tX6zNRwAAPiLQAAKDDRA=
To: <tsvwg@ietf.org>
X-OriginalArrivalTime: 05 Feb 2002 19:01:57.0905 (UTC) FILETIME=[95B45810:01C1AE77]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id OAA26020
Subject: [Tsvwg] Comments about the AddIP draft
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi all!

	I have some comments/typos about this draft. Here they go!:

	- The flow limits procedures were taken out from the draft. I read the comment about the Abstract. However, the chunks themselves are called "Address/Stream...". I guess that the "Stream" part should maybe taken off... This affects to Section 3.1, Table 1 (both the ASCONF chunk and the title of the table) and the title of sections 3.1.1 and 3.1.2 (also in the index).

	Also, in that case, shouldn't the chunks be called ACONF and ACONF-ACK?

	- Section 3.1.1, first paragraph. Does TLV stand for Tag-Length-Value or Type-Length-Value (or most probably, both)?

	Another issue. Is there any restriction for the value of the Correlation IDs (apart from being unique inside the ASCONF)? Could they simply be set to 0, 1, 2, 3... in every ASCONF chunk?

	- Section 3.3, Table 3: The Cause Code Values are wrong... There are 0s missing...

	- Section 3.3.2, first paragraph: Shouldn't be the "ASCONF-REQ" chunk that appears there simply "ASCONF"?

	- Section 4.1.1, R2): What is the meaning of "if there is room in the cwnd"? I mean, there is room for a DATA chunk if the user data (but not the whole chunk) is not bigger than cwnd. How is this applied to an ASCONF chunk? I can think about not sending an ASCONF chunk whose total length is not bigger than the cwnd... But it wouldn't silly thinking about being able to send it if cwnd is at least 1 byte... or maybe I'm just silly :-)

	Moreover, imagine we have an ASCONF chunk that could not be sent due to cwnd, and several DATA chunks kept at the sender (I would say the normal situation when we are sending data using the whole cwnd). When a SACK chunk comes, which one has the priority? I mean, should we send as many DATA chunks as we can and then the ASCONF (if there is still room for it), or should we send the ASCONF first?

	- Section 4.3, D0): This happens when an attacker has sent an ASCONF chunk to our peer, knowing the Verification tag, the Serial Number and faking one of our addresses, and our peer sent the ASCONF-ACK to us. If we are going to abort the association, wouldn't it be nice to have an Error Cause to tell the peer about this? Otherwise, the peer will simply receive suddenly an ABORT chunk now knowing why... If there is such an attacker, we cannot do anything against it, but at least the peer would also know about it...

	- Section 4.3.2, first paragraph: IPv6 addresses can be renumbered and that makes that the ASCONF chunk has to carry in this case one of the old IP addresses in order to be able to find the association. Sorry, but I have never liked that much this... :-)

	Couldn't this be avoided simply sending an ASCONF including both an addition (the new address) and a deletion (the old one)? It is true that upon receipt, the datagram will come from the new address and thus it will seem to be an OOTB one at the first sight. But in this cases, we take a look inside the datagram and check if there is an ASCONF chunk, and then we check if the Address Bytes correspond to any of our addresses. As this is already a special case, why don't we look inside the ASCONF chunk and try to find there a Delete IP Address parameter with one of the old addresses (which should be the only address of an association)? Putting this parameter the first one inside the ASCONF chunk would make things mostly the same as they are now. However, it would violate the rule of not erasing the last address. It could also be the second one, as far as the first paragraph is an addition... which would not make things much harder either...

	The procedure now is like this:

		a) If we have an OOTB datagram, check all the other exceptions we already had. If we are not in any of those situations, look for an ASCONF chunk.
		b) If there is an ASCONF chunk check the Address Bytes and check if that address appears in any of your associations.

	The new one would be:

		a) Same as before.
		b) Check that the chunk has at least two parameters.
		c) If so, check that the second parameter is an deletion and check if that address is part of any of our existing associations.

	Or even simpler. We could require in this case that the peer simply sends an ASCONF carrying a single Delete IP Address parameter, and we could in this case add the source address of the SCTP datagram... I don't know, whatever, but... wouldn't the ASCONF chunk (without those Address Bytes) be... more "beautiful"? ;-)

	Or another possibility of change... why not simply adding an IPv4/IPv6 Address parameter when necessary (only in this special "renumbering case") instead of using the Address Bytes? We could force that in this case this parameter should be the first one in the ASCONF chunk.

	In any case, when we are facing renumbering and we receive the ASCONF chunk, wouldn't it be a good idea forcing the ASCONF to be the first one in the datagram (so it is easier to recognize the OOTB packets)?

	I know that this can happen in IPv6, but, can something similar happen in IPv4? If not, why including that Address Type?

	And the last comment about this. Should always the Address Type/Address Bytes fields of the ASCONF chunk be set to one valid address of the association? If not, and it should be set only in the "renumbering case", should we fix them to all 0s when they are meaningless?


	And finally, some minor typos:

	- Section 3.2.5, first paragraph of page 9: "... report an success..."

	- Section 3.2.6, first paragraph: "...the INITand INIT-ACK...", and "...specify a bit int the reserved...".

	- Section 4.1, last paragraph of page 12: I guess that "Each destination address a single RTO estimate" should be "Each destination address has a single RTO estimate".

	- Section 4.1.1, R2), R3), R4) and R7): "A ASCONF" and "... a ASCONF..." might be "an ASCONF".

	- Section 4.1.1, last paragraph of page 13: "... receives a Operational...".

	- Section 4.2, V2): "... in in the INIT ACK...".

	- Section 4.3, D6): "... set a IP header...".

	- Section 4.4, second paragraph: "... appear in a INIT or...".


	Well, that's all, thanks for your time!

	BR Iván Arias Rodríguez

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Feb  5 16:24:59 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA19901
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 16:24:59 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA02766
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 16:25:02 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA00359;
	Tue, 5 Feb 2002 15:30:52 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA00328
	for <tsvwg@ns.ietf.org>; Tue, 5 Feb 2002 15:30:50 -0500 (EST)
Received: from sj-msg-core-2.cisco.com (sj-msg-core-2.cisco.com [171.69.24.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA18126
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 15:30:46 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-2.cisco.com (8.11.3/8.9.1) with ESMTP id g15KUJb11418;
	Tue, 5 Feb 2002 12:30:19 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id AAZ92477;
	Tue, 5 Feb 2002 12:30:18 -0800 (PST)
Message-ID: <3C6040D9.1D5F724D@stewart.chicago.il.us>
Date: Tue, 05 Feb 2002 14:30:17 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] Comments about the AddIP draft
References: <58AEDE27D48F884285007ABF827FF26308EE7C@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Hi all!
> 
>         I have some comments/typos about this draft. Here they go!:
> 
>         - The flow limits procedures were taken out from the draft. I read the comment about the Abstract. However, the chunks themselves are called "Address/Stream...". I guess that the "Stream" part should maybe taken off... This affects to Section 3.1, Table 1 (both the ASCONF chunk and the title of the table) and the title of sections 3.1.1 and 3.1.2 (also in the index).
> 

And one in section 4 as well :0 I will change this in -05...

>         Also, in that case, shouldn't the chunks be called ACONF and ACONF-ACK?
> 

I think ASCONF is easier to say.. so lets leave the S in just for fun :>


>         - Section 3.1.1, first paragraph. Does TLV stand for Tag-Length-Value or Type-Length-Value (or most probably, both)?
> 
Hmm we should use only one term.. but both areactually the same :>

I see we use Type-Length-Value in RFC2960 so I will change -05 to match.




>         Another issue. Is there any restriction for the value of the Correlation IDs (apart from being unique inside the ASCONF)? Could they simply be set to 0, 1, 2, 3... in every ASCONF chunk?
> 

I see no reason why there would be any restrictions.. since you can
only have one out at a time... One could use the same 0-N each
time...

Of course for added security one could even use random values :>

>         - Section 3.3, Table 3: The Cause Code Values are wrong... There are 0s missing...

You would like leading 0's then.. ok I will add them :>

> 
>         - Section 3.3.2, first paragraph: Shouldn't be the "ASCONF-REQ" chunk that appears there simply "ASCONF"?

Yep .. just ASCONF

> 
>         - Section 4.1.1, R2): What is the meaning of "if there is room in the cwnd"? I mean, there is room for a DATA chunk if the user data (but not the whole chunk) is not bigger than cwnd. How is this applied to an ASCONF chunk? I can think about not sending an ASCONF chunk whose total length is not bigger than the cwnd... But it wouldn't silly thinking about being able to send it if cwnd is at least 1 byte... or maybe I'm just silly :-)

I don't think we need to limit ASCONF's based on if they fit in the
cwnd. The idea here was to keep someone from pounding out ASCONF's
when one could not normally send data... Hmm I suppose we could
make it so it must fit in the cwnd.. but it really does not take
up any cwnd space.. i.e. it is not tracked as a "inflight" data since
it is not data. 

I think we can just allow a simple check

if(inflight > cwnd)
	delay sending asconf
else
        send asconf..




> 
>         Moreover, imagine we have an ASCONF chunk that could not be sent due to cwnd, and several DATA chunks kept at the sender (I would say the normal situation when we are sending data using the whole cwnd). When a SACK chunk comes, which one has the priority? I mean, should we send as many DATA chunks as we can and then the ASCONF (if there is still room for it), or should we send the ASCONF first?
> 

Since the ASCONF does not take up any room one would send

ASCONF+DATA--->
DATA->
DATA-->

until the full cwnd was occupied by the data heading out.

Since one can only have one out at a time I think it is reasonable.



>         - Section 4.3, D0): This happens when an attacker has sent an ASCONF chunk to our peer, knowing the Verification tag, the Serial Number and faking one of our addresses, and our peer sent the ASCONF-ACK to us. If we are going to abort the association, wouldn't it be nice to have an Error Cause to tell the peer about this? Otherwise, the peer will simply receive suddenly an ABORT chunk now knowing why... If there is such an attacker, we cannot do anything against it, but at least the peer would also know about it...

Hmm.. yes we could add an error cause something like:

 - Association aborted due to illegal asconf-ack



> 
>         - Section 4.3.2, first paragraph: IPv6 addresses can be renumbered and that makes that the ASCONF chunk has to carry in this case one of the old IP addresses in order to be able to find the association. Sorry, but I have never liked that much this... :-)
> 
>         Couldn't this be avoided simply sending an ASCONF including both an addition (the new address) and a deletion (the old one)? It is true that upon receipt, the datagram will come from the new address and thus it will seem to be an OOTB one at the first sight. But in this cases, we take a look inside the datagram and check if there is an ASCONF chunk, and then we check if the Address Bytes correspond to any of our addresses. As this is already a special case, why don't we look inside the ASCONF chunk and try to find there a Delete IP Address parameter with one of the old addresses (which should be the only address of an association)? Putting this parameter the first one inside the ASCONF chunk would make things mostly the same as they are now. However, it would violate the rule of not erasing the last address. It could also be the second one, as far as the first paragraph is an addition... which would not make things much harder either...
> 

This is a lot of overhead considering there may be many addresses and
you thus have to grub through ALL the addresses. Having a fixed
place to look is much quicker... I don't think I like this idea..
sorry..

Having implemented this it is quite easy.. if you see it is an asconf
and you have not found the assoc, you look in one place.. instead
of having to parse all the asconf parameters...


>         The procedure now is like this:
> 
>                 a) If we have an OOTB datagram, check all the other exceptions we already had. If we are not in any of those situations, look for an ASCONF chunk.
>                 b) If there is an ASCONF chunk check the Address Bytes and check if that address appears in any of your associations.
> 
>         The new one would be:
> 
>                 a) Same as before.
>                 b) Check that the chunk has at least two parameters.
>                 c) If so, check that the second parameter is an deletion and check if that address is part of any of our existing associations.
> 
>         Or even simpler. We could require in this case that the peer simply sends an ASCONF carrying a single Delete IP Address parameter, and we could in this case add the source address of the SCTP datagram... I don't know, whatever, but... wouldn't the ASCONF chunk (without those Address Bytes) be... more "beautiful"? ;-)

Nope. I don't see how the above would work with a single delete.. you 
may well end up deleting the LAST address... No it is far
easier to have one lookup location and look it up... There are
other cases we have seen in our testing (when renumbering occurs) that
the address bytes are really essential... 

An address can become unusable and yet its present for a while. No
delete shows up.. instead you just have the new one added and the
old one hanging on in a IFF_DETACHED state. This kills you unless
you have something like the address bytes.. 

I think that :

A) they are necessary
and
B) they are efficient.. and they are only 16 bytes on a really
   special case thing (ASCONF) that is not that often sent...

> 
>         Or another possibility of change... why not simply adding an IPv4/IPv6 Address parameter when necessary (only in this special "renumbering case") instead of using the Address Bytes? We could force that in this case this parameter should be the first one in the ASCONF chunk.

This is better but I see no advantage to a TLV other than saving
a few bytes on every message.. I don't see how it is worth it.. and even
more so one has to go looking for the TLV instead of looking in a
fixed postion. This requires the code that is worriing about
lookup/vtag verfication to know more about TLV's and things .. more
work in my mind.. I like the fixed location.. sorry that you dont :-0

> 
>         In any case, when we are facing renumbering and we receive the ASCONF chunk, wouldn't it be a good idea forcing the ASCONF to be the first one in the datagram (so it is easier to recognize the OOTB packets)?

I think the ASCONF has to be the first one or you will get an
ABORT back.. You have to be ready for this case (since you could
get re-ordering or something) but still.. probably a good idea..


> 
>         I know that this can happen in IPv6, but, can something similar happen in IPv4? If not, why including that Address Type?

Yes, it can... consider the administrator does:

ifconfig mynet 10.1.1.3 netmask 0xffffff00

Now bring up an sctp association etc... and things
are humming along..

Now my administrator (which may not be me) says .. oh no
I fat fingered it and my address is really 10.1.1.9

ifconfg mynet 10.1.1.9 netmaks 0xffffff00

And boom.. you have the same case.

I have tested with this scenario by the way.. it works very nicely
too..


> 
>         And the last comment about this. Should always the Address Type/Address Bytes fields of the ASCONF chunk be set to one valid address of the association? If not, and it should be set only in the "renumbering case", should we fix them to all 0s when they are meaningless?

I see no reason that it cannot be filled all the time. We do it this
way in our implementation since it makes building the chunk
much easier...

> 
>         And finally, some minor typos:
> 
>         - Section 3.2.5, first paragraph of page 9: "... report an success..."

got it.

> 
>         - Section 3.2.6, first paragraph: "...the INITand INIT-ACK...", and "...specify a bit int the reserved...".

got it

> 
>         - Section 4.1, last paragraph of page 12: I guess that "Each destination paddress a single RTO estimate" should be "Each destination address has a single RTO estimate".

yep.

> 
>         - Section 4.1.1, R2), R3), R4) and R7): "A ASCONF" and "... a ASCONF..." might be "an ASCONF".

ok

> 
>         - Section 4.1.1, last paragraph of page 13: "... receives a Operational...".
> 

got it..

>         - Section 4.2, V2): "... in in the INIT ACK...".
> 
ok
>         - Section 4.3, D6): "... set a IP header...".

yep
> 
>         - Section 4.4, second paragraph: "... appear in a INIT or...".
got it..

> 
>         Well, that's all, thanks for your time!
> 
>         BR Iván Arias Rodríguez
> 
Thanks for your close look Ivan..

R

--
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Feb  5 19:56:14 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA26399
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 19:56:14 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id TAA13382
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 19:56:16 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA12882;
	Tue, 5 Feb 2002 19:44:48 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA12850
	for <tsvwg@ns.ietf.org>; Tue, 5 Feb 2002 19:44:43 -0500 (EST)
Received: from mxic2.us.dg.com (mxic2.isus.emc.com [128.221.31.40])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA26147
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 19:44:41 -0500 (EST)
From: Black_David@emc.com
Received: by mxic2.us.dg.com with Internet Mail Service (5.5.2653.19)
	id <ZRAV08PG>; Tue, 5 Feb 2002 19:39:01 -0500
Message-ID: <277DD60FB639D511AC0400B0D068B71E02937420@CORPMX14>
To: dotis@sanlight.net, ips@ece.cmu.edu
Cc: tsvwg@ietf.org
Date: Tue, 5 Feb 2002 19:44:10 -0500 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: [Tsvwg] RE: iSCSI: No Framing
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Julian,

> Dough is continuing his rumblings - whether they are relevant to us or
not.
> I wonder if there is some administrative measure you can take to save us
> some bandwidth or we should install our own "Otis filters".

Doug is within bounds, but not by much.  He's basically correct that:

> Discussion regarding the concept of framing using FIM within iSCSI is a
> valid topic at this time.  [... snip ...]  With this topic to be decided
> shortly, now is my opportunity to speak.

On the other hand, Doug continues to engage in flawed reasoning that shows
a lack of respect for other members of the WG.  His very next sentence,
for example:

> A generalized scheme for implementing DDP is relevant as opposed to one
that
> is proprietary and application specific.  

The word "proprietary" is an unfortunately typical of Doug:
(1) assuming that standardization of certain components and/or interfaces 
	is/are necessary to the use of FIM,
(2) failing to acknowledge the existence of implementation approaches that
	may not require such components or interfaces, AND
(3) failing to consider the possibility that standardization of such
	components or interfaces may not be necessary to interoperability.
The result is that here and elsewhere when the IPS WG avoids Doug's
invitation to unnecessary standardization, Doug accuses the IPS WG
of working in secret on proprietary solutions.

Needless to say such accusations are not only groundless, but have also
become boring due to excessive repetition.  Doug might get more respect for
his ideas if he showed more respect for the ideas of others.  Nonetheless,
it is fair for Doug to advocate that SCTP is the right long-term solution
and that FIM is not a step in that direction:

> To prevent fracturing of the market and creating standards competition
> within IETF, FIM should be removed from the iSCSI draft.  Transition to
SCTP
> for this feature and do not add layers to TCP.

One's personal filters are ones own concern - I have a responsibility to
pay attention to the contributions of all, but it would help if Doug
would remove the words "proprietary" and "secret" from the vocabulary
he uses when posting to this list.

Thanks,
--David
---------------------------------------------------
David L. Black, Senior Technologist
EMC Corporation, 42 South St., Hopkinton, MA  01748
+1 (508) 249-6449 *NEW*      FAX: +1 (508) 497-8500
black_david@emc.com         Cell: +1 (978) 394-7754
---------------------------------------------------

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Feb  5 22:02:26 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA00012
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 22:02:26 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id WAA20808
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 22:02:29 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA18908;
	Tue, 5 Feb 2002 21:42:45 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA18879
	for <tsvwg@ns.ietf.org>; Tue, 5 Feb 2002 21:42:42 -0500 (EST)
Received: from ps.pxserver.com ([202.164.189.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA29510
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 21:42:34 -0500 (EST)
Received: from localhost ([202.164.189.225]) by ps.pxserver.com with Microsoft SMTPSVC(5.0.2195.2966);
	 Wed, 6 Feb 2002 10:47:29 +0800
X-Sender: johnx3001@yahoo.com
From: John M <johnx3001@yahoo.com>
To: tsvwg@ietf.org
Date: Wed, 06 Feb 2002 10:47:29 +0800
Reply-To: johnx3002@yahoo.com
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 7bit
Message-ID: <PSZDnrasn3c6xpwW35B00000f86@ps.pxserver.com>
X-OriginalArrivalTime: 06 Feb 2002 02:47:29.0783 (UTC) FILETIME=[9E66B070:01C1AEB8]
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] JOIN FOR FREE! ... Learn and Earn
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit


Hello Fellow Networker,

My name is John Miranda. I am now focusing on one
online opportunity. 

I have tried several of these online opportunities 
full of hype promising us thousands of dollars 
every month.

It is a sad fact that many people who are in need
to earn additional income is being victimized by
this fly by night scam artists. Every time I tried
one opportunity after the other, I ended more broke
than ever.

As a result of trying all these opportunities, I
have decided to select on my portfolio the company
which is true to their words. Not the one full of
hype. The only one that has consistently sent me
the monthly check. Even without the hype they have
given me the best compensation plan with their high
% commission. They are providing real service not
the one that simply transfer wealth from the new
signups to the people at the top of the pyramid.
They are not a form of a pyramid scam.

Also when you join, you will have a team of upline
sponsors who will help you succeed every step of
the way. Instead of being left alone, you will be
guided step by step by real people, not those
autoresponders.

Do not believe in do nothing, no recruiting, no
selling schemes. They are scams. You might want to
try it to prove my words. Only you can make your
own success together with the help of your
sponsors.

If you have 2 to 3 hours a day, you will be able to
earn a full income in a few months. And there is
absolutely no limit as how much you can earn. It is
designed to gain momentum after some time.


But I prefer to avoid such statements as "You will
get rich". I read it everywhere and I do not want
to sound like them.

To get your FREE membership ID send email to 
johnx300@yahoo.com  and put "REGISTER ME FOR FREE" 
in the subject and your fullname in the body of your email.



Regards,

John Miranda
johnx300@yahoo.com 


Note: 	You don't need to request for removal.
	This is a one time email.
	Your email address will be automatically de-activated 
	in our list if you don't reply in this mail.















_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Tue Feb  5 23:29:35 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA02509
	for <tsvwg-archive@odin.ietf.org>; Tue, 5 Feb 2002 23:29:33 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id XAA24204
	for tsvwg-archive@odin.ietf.org; Tue, 5 Feb 2002 23:29:37 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id XAA24025;
	Tue, 5 Feb 2002 23:22:19 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id XAA23949
	for <tsvwg@ns.ietf.org>; Tue, 5 Feb 2002 23:22:15 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-200.dallas.net [209.44.42.200])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA02367
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 23:22:09 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id WAA29886;
	Tue, 5 Feb 2002 22:21:53 -0600
Date: Tue, 5 Feb 2002 22:21:53 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Black_David@emc.com
Cc: dotis@sanlight.net, ips@ece.cmu.edu, tsvwg@ietf.org
Subject: Re: [Tsvwg] RE: iSCSI: No Framing
Message-ID: <20020205222153.A23964@openss7.org>
Reply-To: bidulock@openss7.org
References: <277DD60FB639D511AC0400B0D068B71E02937420@CORPMX14>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2i
In-Reply-To: <277DD60FB639D511AC0400B0D068B71E02937420@CORPMX14>; from Black_David@emc.com on Tue, Feb 05, 2002 at 07:44:10PM -0500
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

David,

On Tue, 05 Feb 2002, Black_David@emc.com wrote:

> The result is that here and elsewhere when the IPS WG avoids Doug's
> invitation to unnecessary standardization, Doug accuses the IPS WG
> of working in secret on proprietary solutions.

What's the matter with that?  I would say that 80-90% of
those lurking on these working groups are doing exactly that:
working in secret on proprietary solutions; forming trade
secrets to embed in their implementations; and patenting
whatever the can to encumber the competition.  That's just the
way of the world.

As with any other standard body, IETF standards represent a
tenuous balance between interoperability and proprietary
implementation.  Even so, the number of proprietary (and
patented) protocols proliferate the RFC servers as
"Informational" and some even as proposed standards with
Intellectual Property claims.

I find it hard to believe that a tortuous approach that has
only short-term viability is going to have such a deterious
effect in an industry that holds the trade secret and
intellectual property rights in so high an etseem.

(I would invite Doug to re-read the copyright disclaimer
 on all RFCs.)

What will this market be fractured into?  Those with trade
secrets and those without?  Or is this a Beta vs. VHS thing?

At least the draft provides some basis for interoperability
and addresses the extent to which those taking this short-term
blind alley are willing to reveal their plans to each other.

So how will it be fractured?--Into those up the blind alley
and those in the traffic jam?

Please Doug, if there is a technical (not political) basis
on which this draft is somehow flawed, please express it.
You seem so vehemenently opposed to this draft over the
last number of months: surely there is some technical basis
upon which your first aversions were based.

If it is the protocol layer boundary violations that concern
you, what is the technical (rather than preceived market)
disadvantages of such violation?

It is only on the basis of technical or procedural error that
an appeal can be honestly and successfully lodged under the
IETF process anyway and you had certainly better air those
technical objections within the WG first.

Could you provide a summarized list of your technical
objections with proposed solutions so that others in the
WG can comment?  There might be several solutions that might
lead to improvements even at this late a stage.  It is
normally only a technical argument that will lead to concensus
here.  (I doubt that the WG chair could ask for comments on
a question concerning market fragmentation and trade secret
plots.)

--brian

-- 
Brian F. G. Bidulock
bidulock@openss7.org
http://www.openss7.org/

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Feb  6 00:08:04 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA03510
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 00:08:04 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id AAA26333
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 00:08:06 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id XAA25504;
	Tue, 5 Feb 2002 23:59:20 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id XAA25477
	for <tsvwg@ns.ietf.org>; Tue, 5 Feb 2002 23:59:18 -0500 (EST)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA03359
	for <tsvwg@ietf.org>; Tue, 5 Feb 2002 23:59:12 -0500 (EST)
Received: from netlab.hcltech.com (arun [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id KAA17296;
	Wed, 6 Feb 2002 10:50:16 +0530
Message-ID: <3C60B827.34333756@netlab.hcltech.com>
Date: Wed, 06 Feb 2002 10:29:19 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Randall Stewart <randall@stewart.chicago.il.us>
CC: Ivan.Arias-Rodriguez@nokia.com, mohammed.asharaf@wipro.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] sctp Window deadlock
References: <58AEDE27D48F884285007ABF827FF26308EE78@esebe017.NOE.Nokia.com> <3C5FFF3C.F3BF56D5@stewart.chicago.il.us>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi Randal,
    Yes, in TCP there is a timer called "Persist timer", which is used for the Probing
activity and not the T3 timer... and this has no threshold as u mentioned.....
    But  if this the procedure to be followed, then why its not specified in the Rfc
and the implementation guides (or yet to be included....).  What I thought is,
if the procedure is to be followed then it should be there in the Rfc (atleast
with the MAY condition).......
Or we should go to the tcp implementation for all these conditions which are
missing in the SCTP Rfc  :)

Or did I miss something in the imp guides and Rfc !!!!


Thanks
-arun

Randall Stewart wrote:

> Ivan:
>
> I agree with all your assesement except I would NOT count
> the T3 against anything towards failure (no matter if
> HB's are running or not) if we are doing widow-probes.
>
> I run special code for this when we get in the window-probe
> mode... And no threshold managment is done...
>
> This is the same thing that happens in TCP (if my understanding
> is correct).. i.e. once the sender clamps the rwnd=0 you
> will continue to probe until it opens up.. as long as
> you get a SACK back that is...
>
> R
>
> Ivan.Arias-Rodriguez@nokia.com wrote:
> >
> >     Hi all!
> >
> >     Well, I think that the point here is that once the T3 timer expires on an address, all the TSNs that were outstanding on that address are automatically marked for retransmission, i.e, they are not outstanding TSNs any longer...
> >
> >     In this situation, as you are allowed to have at least one TSN outstanding no matter the value of rwnd, you will send the oldest TSN marked for retransmission and restart the T3 again (if possible using another different address than the first time)...
> >
> >     After several failures, you will abort the association... It might take time (RTO.Max is one minute), but at the end you get rid of the association...
> >
> >     However... If the path heartbeat mechanism is enabled, every received HEARTBEAT ACK chunk will erase the retransmission counter... Thus we are in the same problem... But this is because the receiver is not behaving in the expected way...
> >
> >     BR Iván Arias Rodríguez
> >
> > -----Original Message-----
> > From: ext Arun Prasad [mailto:arun@netlab.hcltech.com]
> > Sent: 05. February 2002 16:37
> > To: Ashraf
> > Cc: Tsvwg; Randall Stewart
> > Subject: Re: [Tsvwg] sctp Window deadlock
> >
> > hi Ashraf,
> >
> > >From Rfc:
> > However, regardless of the value of rwnd (including if it
> >       is 0), the data sender can always have one DATA chunk in flight to
> >       the receiver if allowed by cwnd
> >
> >     My understanding from the above is, trying to send the Data chunk only once
> > for probing, and if that chunk is not acknowledged (or acknowledged with rwnd
> > as zero), then even when the Rtx timer expires for that Data chunk, the data chunk
> > will be queued in the local Q (as the rwnd is zero, and already one data chunk
> > has been sent for probing...) and will not be retransimtted.....
> >     I dont know whether my understanding is right......  Just getting a feeling that
> > Iam going wrong somewhere....( might be ur understanding is right.....)
> > Somebody please clarify.....
> >
> > Thanks
> > -arun
> >
> >
> > Ashraf wrote:
> >
> > Hi,
> >     My understanding is ( pls correct me if I am wromg),  each time when the timer expires, the path /assoc retrans counter will be
> >
> > incremented and sender keeps on probing  for change in window size  by sending next fragment ( in ur caase fragment B ) and  ending
> >
> > up in T3 timer expiry each time. Will the Assoc will be  terminated once this count reached the threshold value.?"
> >
> > Thanks,
> > ashraf
> >
> > Arun Prasad wrote:
> >
> > > Hi Ashraf,
> > >     ya RTT timer will be started, but upon its expiry, we come to the same problem..
> > > because we cant send the retransmitted packets due to rwnd shortage......
> > >
> > > Thanks
> > > -arun
> > >
> > > Ashraf wrote:
> > >
> > > > Hi arun,
> > > >     I think  for the new data chunk ( which we are sending  to probe for change in window), we start RTT timer and  give
> > > > appropraite treatment, when it expires.
> > > >
> > > > thanks,
> > > > ashraf
> > > >
> > > > Arun Prasad wrote:
> > > >
> > > > > Hi Ashraf,
> > > > >     Thanks for the rsp....
> > > > > ya, I agree that partial-delivery will solve the problem... But imagine an errornous
> > > > > stack ( or SCTP stack which didnt implement the partial delivery API)  is on the
> > > > > receiving side.... Then with the below example scenario, the Sender stack will
> > > > > get stuck.... ie.,  the association will never tear down (as RTX timer is not running)
> > > > > so the Sender side will be hung (with no mistake on its side....) forever.....
> > > > > I think this situation should be avoided......
> > > > >
> > > > > Thanks
> > > > > -arun
> > > > >
> > > > > Ashraf wrote:
> > > > >
> > > > > > Arun,
> > > > > >     This is what RFC has to say :
> > > > > > "If the data receiver  run out of buffer space while still  waiting for  for more fragments
> > > > > > to complete the re-assembly, of the message, it should dispatch part of its inbound message
> > > > > > through a partial deliver API, freeing some if its buffer space in order receive  rest of
> > > > > > the message"
> > > > > >
> > > > > > guess this should take care of the deadlock  situation you are mentioning here.
> > > > > >
> > > > > > Thanks,
> > > > > > Mohd. Ashraf
> > > > > > Arun Prasad wrote:
> > > > > >
> > > > > > > Hi all,
> > > > > > >     The question might be more of  SCTP implementation specific....   According
> > > > > > > to the SCTP rfc2960, the data sender should not send any data packets if it doesnt
> > > > > > > have enough Send-Windown (Peer Rwnd).... Just take the below example....
> > > > > > >
> > > > > > > The Peer's Rwnd (at the Init time ) is 1500 bytes.
> > > > > > > The Sender receivies a data buffer of size 6000 from its upper layer to be sent
> > > > > > > to the Peer.... The Sender will then fragment the packets to fragments A,B,C,D of
> > > > > > > size 1500 bytes (because the local mtu size is 1500bytes)....
> > > > > > > Now the Sender sends the frag A to the Peer... Upon receiving the frag A, the
> > > > > > > Peer's Rwnd will be zero.... This is informed to the Sender by sending a SACK with
> > > > > > > rwnd as 0. The Sender which would have queued the frags B,C,D... will then
> > > > > > > send frag B as the probing data to the Peer... The peer again will respond with
> > > > > > > a SACK of rwnd 0 .  Now the deadlock situation occurs,  as the
> > > > > > >     * Sender cant send any more data packets as the Peer's rwnd is zero....
> > > > > > >     * Peer cant increase its rwnd till it releases the data chunks which it had already
> > > > > > >        received,  but to release that it needs to reassemble the frags (ABCD) and
> > > > > > >       deliver it to the above layer, only then it can release the buffers..... Since its
> > > > > > >       waiting for the frags C and D...  this will never release the buffers.... so
> > > > > > >       will not increase the rwnd.....
> > > > > > >
> > > > > > > This situation can be avoided if the Sender doesnt sent Data messages of size
> > > > > > > greater than the Peer's rwnd.... But the Rfc doesnt say anything about this.....
> > > > > > > Or there should be some timers, to get out of the Deadlock situation.....
> > > > > > >
> > > > > > > Any suggestions........
> > > > > > >
> > > > > > > Thanks
> > > > > > > -arun
> > > > > > >
> > > > > > > _______________________________________________
> > > > > > > tsvwg mailing list
> > > > > > > tsvwg@ietf.org
> > > > > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > > > >
> > > > > >   ------------------------------------------------------------------------------------------------------------------------
> > > > > >                            Name: Wipro_Disclaimer.txt
> > > > > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > > > > >                        Encoding: 7bit
> > > > >
> > > > > --
> > > > > ****************************************************************
> > > > > V.Arun Prasad
> > > > > HCL Technologies Ltd.
> > > > > 51, Jawaharlal Nehru Road,
> > > > > Ekkattuthangal,
> > > > > Guindy Industrial Estate,
> > > > > Chennai - 600097.
> > > > >
> > > > > Contact # : 9144 - 2334174
> > > > >             9144 - 2334181
> > > > >             extn : 233
> > > > > ****************************************************************
> > > > >
> > > > > _______________________________________________
> > > > > tsvwg mailing list
> > > > > tsvwg@ietf.org
> > > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > >
> > > >   ------------------------------------------------------------------------------------------------------------------------
> > > >                            Name: Wipro_Disclaimer.txt
> > > >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> > > >                        Encoding: 7bit
> > >
> > > --
> > > ****************************************************************
> > > V.Arun Prasad
> > > HCL Technologies Ltd.
> > > 51, Jawaharlal Nehru Road,
> > > Ekkattuthangal,
> > > Guindy Industrial Estate,
> > > Chennai - 600097.
> > >
> > > Contact # : 9144 - 2334174
> > >             9144 - 2334181
> > >             extn : 233
> > > ****************************************************************
> >
> >   ------------------------------------------------------------------------------------------------------------------------
> >                            Name: Wipro_Disclaimer.txt
> >    Wipro_Disclaimer.txt    Type: Plain Text (text/plain)
> >                        Encoding: 7bit
> >
> > --
> > ****************************************************************
> > V.Arun Prasad
> > HCL Technologies Ltd.
> > 51, Jawaharlal Nehru Road,
> > Ekkattuthangal,
> > Guindy Industrial Estate,
> > Chennai - 600097.
> >
> > Contact # : 9144 - 2334174
> >             9144 - 2334181
> >             extn : 233
> > ****************************************************************
> >
>
> --
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 01:14:08 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA04355
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 01:14:03 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id BAA07608
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 01:14:04 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA28559;
	Wed, 6 Feb 2002 00:59:12 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA28529
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 00:59:10 -0500 (EST)
Received: from c003.snv.cp.net (c003-h000.c003.snv.cp.net [209.228.32.214])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id AAA04179
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 00:59:09 -0500 (EST)
Received: (cpmta 16756 invoked from network); 5 Feb 2002 21:58:35 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.214) with SMTP; 5 Feb 2002 21:58:35 -0800
X-Sent: 6 Feb 2002 05:58:35 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <ips@ece.cmu.edu>
Cc: <tsvwg@ietf.org>
Date: Tue, 5 Feb 2002 22:00:22 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJIEHECPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Importance: Normal
In-Reply-To: <277DD60FB639D511AC0400B0D068B71E02937420@CORPMX14>
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] RE: iSCSI: No Framing
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

David,

> Julian,
>
> > Dough is continuing his rumblings - whether they are relevant to us or
> > not. I wonder if there is some administrative measure you can take to
> > save us some bandwidth or we should install our own "Otis filters".
>
> Doug is within bounds, but not by much.  He's basically correct that:
>
> > Discussion regarding the concept of framing using FIM within iSCSI is a
> > valid topic at this time.  [... snip ...]  With this topic to be decided
> > shortly, now is my opportunity to speak.
>
> On the other hand, Doug continues to engage in flawed reasoning that shows
> a lack of respect for other members of the WG.  His very next sentence,
> for example:
>
> > A generalized scheme for implementing DDP is relevant as opposed to one
> > that is proprietary and application specific.
>
> The word "proprietary" is an unfortunately typical of Doug:
> (1) assuming that standardization of certain components and/or interfaces
> 	is/are necessary to the use of FIM,

FIM makes absolutely no sense following TCP.  Advantages for FIM are
realized if done prior to placement within system memory as typified by the
general understanding this scheme is for NIC vendors.  Whether it is done
using specialized hardware, intelligent adapters, or software, the goal, as
often stated, is to reduce buffering required in the event of packet loss.
TCP is unable to deliver network traffic during such an event, and FIM is a
scheme to allow a layer prior to TCP to interpret "as an application" the
expected byte stream "out of sequence" as a means to steer encapsulated
payloads.  A generalization of this would be described as direct data
placement.

If done using a protocol employing the TCP transport, a layer must be added
ahead of TCP, which would be, without documentation, a "privately owned"
layer.  Privately owned as such details of this layer are not shared
publicly.  This layer may be hardware or an intelligent adapter, or it could
be a software solution within a memory starved CPU.  Without documentation
describing the information exchanged between the application and TCP,
together with the states involved in this process, any efforts to
extrapolate the details of this layer are likely to venture into proprietary
areas.  A means of avoiding this problem would be to document the basics and
thereby make such a process public.  It would also afford an ability for
review.

> (2) failing to acknowledge the existence of implementation approaches that
> 	may not require such components or interfaces, AND

Proprietary does not imply hardware although in this case it is a high
probability that hardware or an intelligent adapter would implement this
layer.  Unfortunately, the use of TCP will also require that actions of this
"synchronization and steering" layer will be unique for each application.
This is a significant problem that is completely overcome through the use of
SCTP.

> (3) failing to consider the possibility that standardization of such
> 	components or interfaces may not be necessary to interoperability.

This is difficult to determine due to the lack of details.  It is also
difficult to judge security risks for the same reason.  At this point, it is
a black box.  As this layer must anticipate the behavior of TCP, it would be
wise to expose techniques used for doing so and this should be done within
the tsvwg for a broader range of expertise.  There is also a requirement for
associations to be made between buffers and application specific structures
communicated in some fashion from the application to this layer.

> The result is that here and elsewhere when the IPS WG avoids Doug's
> invitation to unnecessary standardization, Doug accuses the IPS WG
> of working in secret on proprietary solutions.

The IPS wg has demonstrated a desire to institute a scheme for incorporating
direct data placement as a generalization of this feature allowed through
the use of FIM.  A noble goal.  This desire has manifested itself as drafts
designed to implement framing using Urgent Pointers, Fixed Interval Markers,
Byte Stuffing, to an all out mandated framing of TCP as detected by segment
length.  This "improvement" to TCP was initiated at the pre-BOF for IPS and
has not relented.  I have never invited change to TCP, but rather strongly
urged the wg to consider SCTP as an alternative solution for DDP versus
these "improvements" to TCP.  My continued opposition to these ideas have
not won me any popularity as I still view these concepts as a kludge and a
corruption of network layering.

My reasoning for pressing this issue is not to prevent framing, but to
suggest a better practice for achieving this feature.  Failing that, have
the iSCSI draft document the interface and states involved in this
"synchronization and steering" layer to assist efforts in the creation of a
generalized scheme for DDP using SCTP.  This would ensure migration from
this application specific layer if it were to become used in practice.  The
benefits of FIM would be marginal at this time as targets can use techniques
to avoid the need for relocating buffers without the use of framing.  This
should allow time for deployment of generalized schemes at the clients using
SCTP.  NIC vendors would reap a larger return on their investment when
offering a generalized feature as there would be a greater opportunity for
its use.  DDP could be employed in many protocols such as RPC as one
example, in addition to SCSI without introducing complex layers beneath the
transport.

> Needless to say such accusations are not only groundless, but have also
> become boring due to excessive repetition.  Doug might get more respect
> for his ideas if he showed more respect for the ideas of others.
> Nonetheless, it is fair for Doug to advocate that SCTP is the right
> long-term solution and that FIM is not a step in that direction:
>
> > To prevent fracturing of the market and creating standards competition
> > within IETF, FIM should be removed from the iSCSI draft.  Transition to
> > SCTP for this feature and do not add layers to TCP.
>
> One's personal filters are ones own concern - I have a responsibility to
> pay attention to the contributions of all, but it would help if Doug
> would remove the words "proprietary" and "secret" from the vocabulary
> he uses when posting to this list.

I should have started my response at the bottom and worked my way up.  By
"proprietary", I am referring to private.  If you see private, think
proprietary and vice versa.  I believe I am using this word correctly.  I
have not used the word "secret" except to comment about the NDT header
offered by Julian.  In that header is a "shared secret" field.  My comment
was that CRC does not mask the "shared secrets" or something to that effect.
I apologize if this is becoming boring, but do not think my concerns are
groundless.  It has been a long process, however.  Again, thank you for your
response.

Doug

>
> Thanks,
> --David
> ---------------------------------------------------
> David L. Black, Senior Technologist
> EMC Corporation, 42 South St., Hopkinton, MA  01748
> +1 (508) 249-6449 *NEW*      FAX: +1 (508) 497-8500
> black_david@emc.com         Cell: +1 (978) 394-7754
> ---------------------------------------------------
>


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 05:42:19 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA16162
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 05:42:19 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA22131
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 05:42:22 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA20100;
	Wed, 6 Feb 2002 05:00:45 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA20057
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 05:00:43 -0500 (EST)
Received: from c003.snv.cp.net (c003-h016.c003.snv.cp.net [209.228.32.230])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id FAA15480
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 05:00:35 -0500 (EST)
Received: (cpmta 7550 invoked from network); 6 Feb 2002 02:00:06 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.230) with SMTP; 6 Feb 2002 02:00:06 -0800
X-Sent: 6 Feb 2002 10:00:06 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <tsvwg@ietf.org>
Cc: <ips@ece.cmu.edu>
Subject: RE: [Tsvwg] RE: iSCSI: No Framing
Date: Wed, 6 Feb 2002 02:01:54 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJGEHFCPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Importance: Normal
In-Reply-To: <20020205222153.A23964@openss7.org>
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

> David,
>
> On Tue, 05 Feb 2002, Black_David@emc.com wrote:
>
> > The result is that here and elsewhere when the IPS WG avoids Doug's
> > invitation to unnecessary standardization, Doug accuses the IPS WG
> > of working in secret on proprietary solutions.
>
> What's the matter with that?  I would say that 80-90% of
> those lurking on these working groups are doing exactly that:
> working in secret on proprietary solutions; forming trade
> secrets to embed in their implementations; and patenting
> whatever the can to encumber the competition.  That's just the
> way of the world.
>
> As with any other standard body, IETF standards represent a
> tenuous balance between interoperability and proprietary
> implementation.  Even so, the number of proprietary (and
> patented) protocols proliferate the RFC servers as
> "Informational" and some even as proposed standards with
> Intellectual Property claims.

By placing these concepts into the public forum, that marks the end in which
"invented here" can be claimed and those organizations participating are
expected to volunteer aspects they consider "proprietary."  You are right
that such is often the case and hence the general agreement for disclosure.

> I find it hard to believe that a tortuous approach that has
> only short-term viability is going to have such a deterious
> effect in an industry that holds the trade secret and
> intellectual property rights in so high an etseem.
>
> (I would invite Doug to re-read the copyright disclaimer
>  on all RFCs.)

I am aware of the disclaimer.  I am also aware of the need for review.
There are many other applications clamoring for this "useful" dodge.  It can
be done in a haphazard fashion, or this transition toward Direct Data
Placement can be done with little disruption to existing software and
equipment.  There is already an RDMA group wishing to form as well as other
clustering schemes that desire this ability, in addition to iSCSI.

> What will this market be fractured into?  Those with trade
> secrets and those without?  Or is this a Beta vs. VHS thing?

There will always be trade secrets or proprietary information regarding any
solution offered by a vendor.  Each has their investment to protect.  The
greatest danger regarding something as basic as a network interface comes
from the vendors.  Should a vendor lay claim to a means of accomplishing the
interface to this beneath TCP "synchronization and steering" layer, then
applications will be forced to modify their interface significantly between
systems.  In essence, it would be Beta vs VHS with respect to something that
should remain ubiquitous.

> At least the draft provides some basis for interoperability
> and addresses the extent to which those taking this short-term
> blind alley are willing to reveal their plans to each other.
>
> So how will it be fractured?--Into those up the blind alley
> and those in the traffic jam?

There are already many proprietary schemes in the market and the IETF has
not prevented any vendor from entering that arena whether they are blind
allies or not.  The IETF should take the long view and strive to prevent
unnecessary disruption to existing technology used within IP.  SCTP offers
the necessary features and protections to allow Direct Data Placement to be
deployed without disruption.  A clever state machines running beneath TCP
attempting to anticipate TCP will likely work in the general case, but then
fail completely when confronting a new and untried piece of equipment that
does something unexpected.  In that case, who would be at fault?  You would
have nothing upon which to judge the cause of the failure.  Was it due to
the application not properly monitoring this undefined status coming for the
"synchronization and steering" layer beneath TCP?  Was it the state machine
within the S&S layer?  There is nothing like complexity to allow a large
company to dominate however. : )

> Please Doug, if there is a technical (not political) basis
> on which this draft is somehow flawed, please express it.
> You seem so vehemenently opposed to this draft over the
> last number of months: surely there is some technical basis
> upon which your first aversions were based.

I think that the concept of network layering where functions are concisely
defined is a technical aspect and not a political one.  That does not mean,
there are not politics at play.  When you introduce a layer below the
transport that manipulates a partial data stream on behalf of the
application, you have thrown the concept of layering out the window.  This
goes well beyond techniques already on the fridge such as packet filters or
ALGs within NATs.  Here you would have continuous interchange between this
S&S layer and the application.  I must admit I am not surprised by the
opposition to a ready made solution that preserves network layering while
offering virtually every possible innovation.

The storage industry rightfully considers themselves an industry unto
themselves.  It is also not surprising the zeal in their attacking perceived
problems.  It has been my privilege to work in this industry and networking
for a quarter of a century.  Working within the tape industry, I have also
learned that it is difficult to make a difference in the direction the disk
industry takes.  The disk cadre and their barbs require a relatively thick
skin if you wish to oppose their consensus.  My resource in this effort is
an annoying persistence.  It may take a prolonged period for what may appear
as a minor detail.

> If it is the protocol layer boundary violations that concern
> you, what is the technical (rather than preceived market)
> disadvantages of such violation?

This new layering concept is not without its risks as it devises placement
of information prior to the protections offered by the transport.  It does
so in perhaps a dangerously predictable manner in that the location of the
marking scheme may easily be spoofed later in time.  This S&S layer through
the use of a highly predictable marker is intended to allow "out of
sequence" segments to be placed.  Depending on how a duplication problem is
handled within this S&S layer, evidence of this event may be obscured.
There is also opportunity for information to be diverted by means of the
communication sent to the S&S layer.  As this layer is below the transport,
this link may suffer weaknesses not present at the more robust API for the
transport itself.  If this communication is present at the transport API, it
should be documented to allow applications consistent essential constructs
when supporting this feature.

> It is only on the basis of technical or procedural error that
> an appeal can be honestly and successfully lodged under the
> IETF process anyway and you had certainly better air those
> technical objections within the WG first.

I see the function of the IETF in broad terms.  There is a general
architecture that is being created in addition to individual protocols.
This process should not opt for a proprietary solution (sorry David)
especially making such a Must, Should, or even Recommended within a draft
when the IETF has already provided a technically superior competing
solution.  SCTP is years ahead on this curve, and although the IPS wg is
good, there is no doubt which is the technically superior solution.
Although David has done an admirable job point out the lines I should not
cross, I too feel there have been some lines crossed.

> Could you provide a summarized list of your technical
> objections with proposed solutions so that others in the
> WG can comment?

I think that I have indicated my concerns and these areas are judgment calls
for those responsible perhaps should framing remain within the iSCSI draft.
More than a year ago my opposition to the use of the TCP Urgent Pointer for
framing caused a similar conflict between David and myself and perhaps set
the stage for this friendly give and take.  One might say that necessity is
the mother of invention, but in this case, it is not necessity as SCTP is a
ready made solution for this framing and layering problem.  I think that if
the IPS wg could see past creating new solutions for a problem already
solved, they would be months, if not years, ahead.  It is a minor point,
remove the FIM from the iSCSI draft and forego use of a layer beneath TCP
for the reasons stated.

Doug

> There might be several solutions that might
> lead to improvements even at this late a stage.  It is
> normally only a technical argument that will lead to concensus
> here.  (I doubt that the WG chair could ask for comments on
> a question concerning market fragmentation and trade secret
> plots.)
>
> --brian
>
> --
> Brian F. G. Bidulock
> bidulock@openss7.org
> http://www.openss7.org/


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 06:16:42 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA16906
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 06:16:42 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA24570
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 06:16:45 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA22890;
	Wed, 6 Feb 2002 05:50:36 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA22860
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 05:50:33 -0500 (EST)
Received: from sj-msg-core-3.cisco.com (sj-msg-core-3.cisco.com [171.70.157.152])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA16326
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 05:50:29 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-3.cisco.com (8.11.3/8.9.1) with ESMTP id g16Antj15893;
	Wed, 6 Feb 2002 02:49:55 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABA11713;
	Wed, 6 Feb 2002 02:50:00 -0800 (PST)
Message-ID: <3C610A57.683630A@stewart.chicago.il.us>
Date: Wed, 06 Feb 2002 04:49:59 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Arun Prasad <arun@netlab.hcltech.com>
CC: Ivan.Arias-Rodriguez@nokia.com, mohammed.asharaf@wipro.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] sctp Window deadlock
References: <58AEDE27D48F884285007ABF827FF26308EE78@esebe017.NOE.Nokia.com> <3C5FFF3C.F3BF56D5@stewart.chicago.il.us> <3C60B827.34333756@netlab.hcltech.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Arun Prasad wrote:
> 
> Or did I miss something in the imp guides and Rfc !!!!
> 

Section 2.15 of the implementation guide

---
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 07:26:05 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA18221
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 07:26:04 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA27573
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 07:26:07 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA26273;
	Wed, 6 Feb 2002 06:56:52 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA26125
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 06:56:44 -0500 (EST)
Received: from hrtades9.atea.be (viap111.atea.be [194.78.143.111])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA17820;
	Wed, 6 Feb 2002 06:56:33 -0500 (EST)
Received: from hrtades10.atea.be (siemens.atea.be [139.10.143.141]) by hrtades9.atea.be with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13)
	id DF4YRD42; Wed, 6 Feb 2002 12:55:51 +0100
Received: by siemens.atea.be with Internet Mail Service (5.5.2653.19)
	id <D6SNFDH6>; Wed, 6 Feb 2002 12:56:16 +0100
Message-ID: <E76F715C0429D5118F2100508BB9EDEE0795AC@hrtades7.atea.be>
From: Coene Lode <Lode.Coene@siemens.atea.be>
To: "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)"
	 <sigtran@ietf.org>,
        "'multi6@ops.ietf.org'" <multi6@ops.ietf.org>
Date: Wed, 6 Feb 2002 12:53:46 +0100 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: [Tsvwg] SCTP multihoming issues draft
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Previous week, the draft on SCTP multihoming issues was published.

Apart from some editorial changes concerning author list and missing/wrong
page numbers,
the only major change is a sentence added at the end of paragraph 2.3 SCTP
multihoming and Network Adress Translators(NAT):
"It should be noted that the NAT box becomes a single
    point-of-failure in this case, as ALL the paths of the SCTP
    association have to go through that single NAT box."
Theoretical it is possible to let all the different paths run through
different NAT boxes, but then the NAT boxes would never know exactly when to
teardown/terminate the association as the terminate or abort takes only one
of the paths(and there are other difficulties, such as the heartbeats
...etc...) So in practical terms this sentence should be correct. If you
feel otherwise, please let me know and I'll remove or add something to it to
make everything clear.

The draft can be found at the IETF draft directory: 
Multihoming issues in the Stream Control Transmission Protocol
<draft-coene-sctp-multihome-02.txt>

If by then end of next week, no further issues have been identified, then I
would ask for this draft to be moved to informational.

yours sincerly
Lode Coene


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 10:13:04 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA22536
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 10:13:04 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA06623
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 10:13:06 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA04603;
	Wed, 6 Feb 2002 09:30:56 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA04546
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 09:30:53 -0500 (EST)
Received: from raven.ecs.soton.ac.uk (raven.ecs.soton.ac.uk [152.78.70.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA20919;
	Wed, 6 Feb 2002 09:30:43 -0500 (EST)
Received: from hawk.ecs.soton.ac.uk (hawk.ecs.soton.ac.uk [152.78.68.142])
	by raven.ecs.soton.ac.uk (8.9.3/8.9.3) with ESMTP id OAA15173;
	Wed, 6 Feb 2002 14:30:39 GMT
Received: from login.ecs.soton.ac.uk (IDENT:tjc@login.ecs.soton.ac.uk [152.78.68.149])
	by hawk.ecs.soton.ac.uk (8.9.3/8.9.3) with ESMTP id OAA22751;
	Wed, 6 Feb 2002 14:30:44 GMT
Date: Wed, 6 Feb 2002 14:30:38 +0000 (GMT)
From: Tim Chown <tjc@ecs.soton.ac.uk>
To: Coene Lode <Lode.Coene@siemens.atea.be>
cc: "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)" <sigtran@ietf.org>,
        "'multi6@ops.ietf.org'" <multi6@ops.ietf.org>
In-Reply-To: <E76F715C0429D5118F2100508BB9EDEE0795AC@hrtades7.atea.be>
Message-ID: <Pine.LNX.4.21.0202061413560.7370-100000@login.ecs.soton.ac.uk>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Subject: [Tsvwg] Re: SCTP multihoming issues draft
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

On Wed, 6 Feb 2002, Coene Lode wrote:

> If by then end of next week, no further issues have been identified, then I
> would ask for this draft to be moved to informational.

Does the draft have to only consider two network interfaces for the
multihoming issues?   In IPv6, as you point out, each interface can have
more than one IPv6 address (and addresses of many scopes), so the simplest
multihoming case for IPv6 is a host (endpoint) with one interface, 
recieving two different /64 prefix RAs, and thus having two IPv6 addresses
from two different providers.   The dual interface scenario may be useful 
of course, e.g. when considering wired and air interfaces.

When you say "it is recommended that IP addresses in a multihomed endpoint 
be assigned IP endpoints from different TLA's to ensure against network 
failure", this would be the default case for IPv6.  

For IPv6, delete section 2.3 :-)

In general, more consideration could be given for IPv6, perhaps?  (you 
did cc multi6 :-)   If I've misinterpreted the text, perhaps the title
of the draft needs revision?

Best wishes,
Tim


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 10:15:36 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA22607
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 10:15:36 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA06757
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 10:15:38 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA06287;
	Wed, 6 Feb 2002 10:03:59 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA06255
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 10:03:57 -0500 (EST)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA22060
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 10:03:53 -0500 (EST)
Received: from netlab.hcltech.com (arun [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id UAA02414;
	Wed, 6 Feb 2002 20:55:12 +0530
Message-ID: <3C6145EF.BDCD44AA@netlab.hcltech.com>
Date: Wed, 06 Feb 2002 20:34:15 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Randall Stewart <randall@stewart.chicago.il.us>
CC: Ivan.Arias-Rodriguez@nokia.com, mohammed.asharaf@wipro.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] sctp Window deadlock
References: <58AEDE27D48F884285007ABF827FF26308EE78@esebe017.NOE.Nokia.com> <3C5FFF3C.F3BF56D5@stewart.chicago.il.us> <3C60B827.34333756@netlab.hcltech.com> <3C610A57.683630A@stewart.chicago.il.us>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi Randel,
    Thanks, I missed it (or rather misunderstood it..... )....  I have one more
doubt in that.... will post it in a seperate mail....

Thanks
-arun

Randall Stewart wrote:

> Arun Prasad wrote:
> >
> > Or did I miss something in the imp guides and Rfc !!!!
> >
>
> Section 2.15 of the implementation guide
>
> ---
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 10:55:20 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA24117
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 10:55:20 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA09407
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 10:55:23 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA07332;
	Wed, 6 Feb 2002 10:24:09 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA07301
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 10:24:06 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-200.dallas.net [209.44.42.200])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA22960
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 10:24:02 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id JAA07447;
	Wed, 6 Feb 2002 09:23:54 -0600
Date: Wed, 6 Feb 2002 09:23:54 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Douglas Otis <dotis@sanlight.net>
Cc: tsvwg@ietf.org, ips@ece.cmu.edu
Subject: Re: [Tsvwg] RE: iSCSI: No Framing
Message-ID: <20020206092354.B23964@openss7.org>
Reply-To: bidulock@openss7.org
References: <20020205222153.A23964@openss7.org> <NEBBJGDMMLHHCIKHGBEJGEHFCPAA.dotis@sanlight.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <NEBBJGDMMLHHCIKHGBEJGEHFCPAA.dotis@sanlight.net>; from dotis@sanlight.net on Wed, Feb 06, 2002 at 02:01:54AM -0800
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Douglas,

On Wed, 06 Feb 2002, Douglas Otis wrote:

[snip]
> > Please Doug, if there is a technical (not political) basis
> > on which this draft is somehow flawed, please express it.
> > You seem so vehemenently opposed to this draft over the
> > last number of months: surely there is some technical basis
> > upon which your first aversions were based.
> 
> I think that the concept of network layering where functions are concisely
> defined is a technical aspect and not a political one.  That does not mean,
> there are not politics at play.  When you introduce a layer below the
> transport that manipulates a partial data stream on behalf of the
> application, you have thrown the concept of layering out the window.  This
> goes well beyond techniques already on the fridge such as packet filters or
> ALGs within NATs.  Here you would have continuous interchange between this
> S&S layer and the application.  I must admit I am not surprised by the
> opposition to a ready made solution that preserves network layering while
> offering virtually every possible innovation.
[snip]

What is lost by throwing layering out the window?  Layering is an
artificial categorization of functions at best.  What technical
advantages are lost?

--brian

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 11:22:12 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA24963
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 11:22:11 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA10913
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 11:22:14 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA08526;
	Wed, 6 Feb 2002 10:35:47 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA08493
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 10:35:45 -0500 (EST)
Received: from prue.eim.surrey.ac.uk (IDENT:exim@prue.eim.surrey.ac.uk [131.227.76.5])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA23500
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 10:35:42 -0500 (EST)
Received: from phaestos.ee.surrey.ac.uk ([131.227.88.14] ident=eep1lw)
	by prue.eim.surrey.ac.uk with esmtp (Exim 3.33 #4)
	id 16YU6I-0006ia-00; Wed, 06 Feb 2002 15:35:10 +0000
Date: Wed, 6 Feb 2002 15:34:51 +0000 (GMT)
From: Lloyd Wood <l.wood@eim.surrey.ac.uk>
X-X-Sender: eep1lw@phaestos.ee.surrey.ac.uk
Reply-To: Lloyd Wood <L.Wood@eim.surrey.ac.uk>
To: "Brian F. G. Bidulock" <bidulock@openss7.org>
cc: Douglas Otis <dotis@sanlight.net>, <tsvwg@ietf.org>, <ips@ece.cmu.edu>
Subject: Re: [Tsvwg] RE: iSCSI: No Framing
In-Reply-To: <20020206092354.B23964@openss7.org>
Message-ID: <Pine.SOL.4.43.0202061534040.333-100000@phaestos.ee.surrey.ac.uk>
Organization: speaking for none
X-url: http://www.ee.surrey.ac.uk/Personal/L.Wood/
X-no-archive: yes
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
X-Scanner: exiscan *16YU6I-0006ia-00*dHMzEGl0ovM* (SECM, UniS)
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

On Wed, 6 Feb 2002, Brian F. G. Bidulock wrote:

> What is lost by throwing layering out the window?

modularity, ease of understanding, and portability.

L.

<L.Wood@surrey.ac.uk>PGP<http://www.ee.surrey.ac.uk/Personal/L.Wood/>


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 11:27:52 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA25130
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 11:27:52 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA11212
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 11:27:55 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA10576;
	Wed, 6 Feb 2002 11:12:58 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA10549
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 11:12:56 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-200.dallas.net [209.44.42.200])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA24752
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 11:12:52 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id KAA08511;
	Wed, 6 Feb 2002 10:12:28 -0600
Date: Wed, 6 Feb 2002 10:12:28 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Lloyd Wood <l.wood@eim.surrey.ac.uk>
Cc: Douglas Otis <dotis@sanlight.net>, tsvwg@ietf.org, ips@ece.cmu.edu
Subject: Re: [Tsvwg] RE: iSCSI: No Framing
Message-ID: <20020206101228.D23964@openss7.org>
Reply-To: bidulock@openss7.org
References: <20020206092354.B23964@openss7.org> <Pine.SOL.4.43.0202061534040.333-100000@phaestos.ee.surrey.ac.uk>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2i
In-Reply-To: <Pine.SOL.4.43.0202061534040.333-100000@phaestos.ee.surrey.ac.uk>; from l.wood@eim.surrey.ac.uk on Wed, Feb 06, 2002 at 03:34:51PM +0000
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Lloyd,

On Wed, 06 Feb 2002, Lloyd Wood wrote:

> On Wed, 6 Feb 2002, Brian F. G. Bidulock wrote:
> 
> > What is lost by throwing layering out the window?
> 
> modularity, ease of understanding, and portability.

Modularity can be acheived without layering.  The entire
field of OOAD proves this.

Ease of understanding is acheived through accurate
interface description.  Again OO shows this.

Portability can also be acheived through OO approaches.

The classical layering of communications functions
into functional groupings is not a necessary condition
for any of these benefits.

--brian

-- 
Brian F. G. Bidulock
bidulock@openss7.org
http://www.openss7.org/

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 11:51:14 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA25551
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 11:51:14 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA12524
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 11:51:17 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA12034;
	Wed, 6 Feb 2002 11:39:36 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA12009
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 11:39:34 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-200.dallas.net [209.44.42.200])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA25349
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 11:39:29 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id KAA09097;
	Wed, 6 Feb 2002 10:39:12 -0600
Date: Wed, 6 Feb 2002 10:39:12 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: "ERICKSON,SHAWN (HP-Roseville,ex1)" <shawn_erickson@hp.com>
Cc: "'Lloyd Wood'" <L.Wood@eim.surrey.ac.uk>,
        Douglas Otis <dotis@sanlight.net>, tsvwg@ietf.org, ips@ece.cmu.edu
Subject: Re: [Tsvwg] RE: iSCSI: No Framing
Message-ID: <20020206103912.E23964@openss7.org>
Reply-To: bidulock@openss7.org
References: <499DC368E25AD411B3F100902740AD6508E99F1F@xrose03.rose.hp.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <499DC368E25AD411B3F100902740AD6508E99F1F@xrose03.rose.hp.com>; from shawn_erickson@hp.com on Wed, Feb 06, 2002 at 08:25:31AM -0800
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

ERICKSON,SHAWN,

These too are better acheived through approaches other
than functional grouping.

--brian

On Wed, 06 Feb 2002, ERICKSON,SHAWN (HP-Roseville,ex1) wrote:

> As well as... potentially security holes, debug-ability, interoperability.
> 
> -Shawn
> 
> -------------------------------------------------------
>  Shawn Carl Erickson
> 
> 
> > -----Original Message-----
> > From: Lloyd Wood [mailto:l.wood@eim.surrey.ac.uk]
> > Sent: Wednesday, February 06, 2002 7:35 AM
> > To: Brian F. G. Bidulock
> > Cc: Douglas Otis; tsvwg@ietf.org; ips@ece.cmu.edu
> > Subject: Re: [Tsvwg] RE: iSCSI: No Framing
> > 
> > 
> > On Wed, 6 Feb 2002, Brian F. G. Bidulock wrote:
> > 
> > > What is lost by throwing layering out the window?
> > 
> > modularity, ease of understanding, and portability.
> > 
> > L.
> > 
> > <L.Wood@surrey.ac.uk>PGP<http://www.ee.surrey.ac.uk/Personal/L.Wood/>
> > 

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 12:15:06 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA26013
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 12:15:05 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id MAA15159
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 12:15:09 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA11298;
	Wed, 6 Feb 2002 11:29:45 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA11273
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 11:29:44 -0500 (EST)
Received: from patan.sun.com (patan.Sun.COM [192.18.98.43])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA25154
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 11:29:40 -0500 (EST)
Received: from phys-hanwk16-1.ebay.sun.com ([129.149.1.5])
	by patan.sun.com (8.9.3+Sun/8.9.3) with ESMTP id JAA16494;
	Wed, 6 Feb 2002 09:29:39 -0700 (MST)
Received: from jetsun (jetsun.Central.Sun.COM [129.153.128.60])
	by phys-hanwk16-1.ebay.sun.com (8.10.2+Sun/8.10.2/ENSMAIL,v2.1p1) with SMTP id g16GTcU25106;
	Wed, 6 Feb 2002 08:29:38 -0800 (PST)
Message-Id: <200202061629.g16GTcU25106@phys-hanwk16-1.ebay.sun.com>
Date: Wed, 6 Feb 2002 10:29:38 -0600 (CST)
From: David Robinson <David.Robinson@Sun.COM>
Reply-To: David Robinson <David.Robinson@Sun.COM>
To: ips@ece.cmu.edu, dotis@sanlight.net
Cc: tsvwg@ietf.org
MIME-Version: 1.0
Content-Type: TEXT/plain; charset=us-ascii
Content-MD5: 01K7YCLjDOjG288LW9hYeg==
X-Mailer: dtmail 1.3.0 @(#)CDE Version 1.5_53 SunOS 5.9 sun4u sparc 
Subject: [Tsvwg] RE: iSCSI: No Framing
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

> > The word "proprietary" is an unfortunately typical of Doug:
> > (1) assuming that standardization of certain components and/or interfaces
> > 	is/are necessary to the use of FIM,

> If done using a protocol employing the TCP transport, a layer must be added
> ahead of TCP, which would be, without documentation, a "privately owned"
> layer.  Privately owned as such details of this layer are not shared
> publicly.  This layer may be hardware or an intelligent adapter, or it could
> be a software solution within a memory starved CPU.  Without documentation
> describing the information exchanged between the application and TCP,
> together with the states involved in this process, any efforts to
> extrapolate the details of this layer are likely to venture into proprietary
> areas.  A means of avoiding this problem would be to document the basics and
> thereby make such a process public.  It would also afford an ability for
> review.

What I am failing to understand is how what you describe is any
different than the variablity of the design of existing NIC
cards. Virtually every NIC card needs its own custom driver
for each OS that is supported. Implementing FIM is no different
than the multitude of existing checksumming offload NICs. It
is true that the hardware is using knowledge of the upper layer
protocols, but it is all within the confines of the implementation
details. There is no wire protocol implication here.

In order to "standardize" these techniques the IETF will be
dictating hardware design and/or protocol stack design. While
this might be useful for an industry consortium to make
portability of software easier, it is not a fundemental IETF
issue. In what is essentially an implementation specific communications
between layers of a software stack, how can you standardize
something that will be uniquely different depending on if
the OS uses STREAMS, BSD sockets, Windows XYZ, or custom OS frozbits.
This doesn't make sense for the IETF, you may consider going
to SNIA to standardize implemenation details.

If I am not understanding your point, I suggest that it would
be useful for you to write a couple paragraphs suitable for
inclusion in a standards document that describes how you
would standardize the FIM specification. Without such a concrete
proposal I believe we are all just flailing at amorphous
ideas and making no progress on this issue.

	-David
	


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 12:21:00 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA26193
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 12:21:00 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id MAA15520
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 12:21:04 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA11092;
	Wed, 6 Feb 2002 11:26:15 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA11065
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 11:26:13 -0500 (EST)
Received: from palrel12.hp.com (palrel12.hp.com [156.153.255.237])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA25058
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 11:26:09 -0500 (EST)
Received: from xparelay1.corp.hp.com (xparelay1.corp.hp.com [15.58.136.173])
	by palrel12.hp.com (Postfix) with ESMTP
	id 1CA6A600850; Wed,  6 Feb 2002 08:25:42 -0800 (PST)
Received: from xpabh3.corp.hp.com (xpabh3.corp.hp.com [15.58.136.223])
	by xparelay1.corp.hp.com (Postfix) with ESMTP
	id 43819E0009B; Wed,  6 Feb 2002 08:25:33 -0800 (PST)
Received: by xpabh3.corp.hp.com with Internet Mail Service (5.5.2653.19)
	id <1KHM23NH>; Wed, 6 Feb 2002 08:25:33 -0800
Message-ID: <499DC368E25AD411B3F100902740AD6508E99F1F@xrose03.rose.hp.com>
From: "ERICKSON,SHAWN (HP-Roseville,ex1)" <shawn_erickson@hp.com>
To: "'Lloyd Wood'" <L.Wood@eim.surrey.ac.uk>,
        "Brian F. G. Bidulock" <bidulock@openss7.org>
Cc: Douglas Otis <dotis@sanlight.net>, tsvwg@ietf.org, ips@ece.cmu.edu
Subject: RE: [Tsvwg] RE: iSCSI: No Framing
Date: Wed, 6 Feb 2002 08:25:31 -0800 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

As well as... potentially security holes, debug-ability, interoperability.

-Shawn

-------------------------------------------------------
 Shawn Carl Erickson


> -----Original Message-----
> From: Lloyd Wood [mailto:l.wood@eim.surrey.ac.uk]
> Sent: Wednesday, February 06, 2002 7:35 AM
> To: Brian F. G. Bidulock
> Cc: Douglas Otis; tsvwg@ietf.org; ips@ece.cmu.edu
> Subject: Re: [Tsvwg] RE: iSCSI: No Framing
> 
> 
> On Wed, 6 Feb 2002, Brian F. G. Bidulock wrote:
> 
> > What is lost by throwing layering out the window?
> 
> modularity, ease of understanding, and portability.
> 
> L.
> 
> <L.Wood@surrey.ac.uk>PGP<http://www.ee.surrey.ac.uk/Personal/L.Wood/>
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 12:49:22 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA26881
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 12:49:22 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id MAA16883
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 12:49:24 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA16400;
	Wed, 6 Feb 2002 12:37:10 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA16369
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 12:37:08 -0500 (EST)
Received: from c003.snv.cp.net (c003-h000.c003.snv.cp.net [209.228.32.214])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id MAA26502
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 12:36:59 -0500 (EST)
Received: (cpmta 9141 invoked from network); 6 Feb 2002 09:36:28 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.214) with SMTP; 6 Feb 2002 09:36:28 -0800
X-Sent: 6 Feb 2002 17:36:28 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <tsvwg@ietf.org>
Cc: <ips@ece.cmu.edu>
Subject: RE: [Tsvwg] RE: iSCSI: No Framing
Date: Wed, 6 Feb 2002 09:38:18 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJMEHKCPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Importance: Normal
In-Reply-To: <20020206101228.D23964@openss7.org>
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Brian,

> Lloyd,
>
> On Wed, 06 Feb 2002, Lloyd Wood wrote:
>
> > On Wed, 6 Feb 2002, Brian F. G. Bidulock wrote:
> >
> > > What is lost by throwing layering out the window?
> >
> > modularity, ease of understanding, and portability.
>
> Modularity can be acheived without layering.  The entire
> field of OOAD proves this.
>
> Ease of understanding is acheived through accurate
> interface description.  Again OO shows this.
>
> Portability can also be acheived through OO approaches.
>
> The classical layering of communications functions
> into functional groupings is not a necessary condition
> for any of these benefits.

Unlike a typical play dough program, this stuff is more like working with
concrete.  A portion of each application demanding a direct placement
feature would be included in hardware or intelligent adapters if using FIM
and TCP, or framing and TCP, for the most part.  SCTP allows layering for
this "clean modularity" needed if one is to design an adapter that need not
understand the complexities and structures of every application that desires
this function.  The adapter interface only needs to understand SCTP and not
the application.  By using the unordered delivery mode, a shim would be able
to implement generalized structures for including Direct Data Placement or
even a full implementation of RDMA.  It would be possible for more elaborate
interfaces to built upon this foundation, but at least this establishes the
modularity, ease of understanding, and portability desired, if one is going
to start working in concrete.

Unlike the case with FIM and TCP, this would introduce NIC devices that will
be forever sensitive to even minor changes to structures employed in these
applications and the desire for this feature does not even begin to end at
just one application.

Doug


> --brian
>
> --
> Brian F. G. Bidulock
> bidulock@openss7.org
> http://www.openss7.org/


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 13:05:16 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA27189
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 13:05:15 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA17905
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 13:05:18 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA15331;
	Wed, 6 Feb 2002 12:16:45 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA15306
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 12:16:44 -0500 (EST)
Received: from c003.snv.cp.net (c003-h016.c003.snv.cp.net [209.228.32.230])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id MAA26077
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 12:16:39 -0500 (EST)
Received: (cpmta 12718 invoked from network); 6 Feb 2002 09:16:11 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.230) with SMTP; 6 Feb 2002 09:16:11 -0800
X-Sent: 6 Feb 2002 17:16:11 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <ips@ece.cmu.edu>
Cc: <tsvwg@ietf.org>
Subject: RE: [Tsvwg] RE: iSCSI: No Framing
Date: Wed, 6 Feb 2002 09:18:01 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJEEHKCPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Importance: Normal
In-Reply-To: <200202061629.g16GTcU25106@phys-hanwk16-1.ebay.sun.com>
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

David,

> > > The word "proprietary" is an unfortunately typical of Doug:
> > > (1) assuming that standardization of certain components
> > > and/or interfaces	is/are necessary to the use of FIM,
>
> > If done using a protocol employing the TCP transport, a layer
> > must be added ahead of TCP, which would be, without
> > documentation, a "privately owned" layer.  Privately owned as
> > such details of this layer are not shared publicly.  This
> > layer may be hardware or an intelligent adapter, or it could
> > be a software solution within a memory starved CPU.  Without
> > documentation describing the information exchanged between the
> > application and TCP, together with the states involved in this
> > process, any efforts to extrapolate the details of this layer
> > are likely to venture into proprietary areas.  A means of
> > avoiding this problem would be to document the basics and
> > thereby make such a process public.  It would also afford an
> > ability for review.
>
> What I am failing to understand is how what you describe is any
> different than the variablity of the design of existing NIC
> cards. Virtually every NIC card needs its own custom driver
> for each OS that is supported. Implementing FIM is no different
> than the multitude of existing checksumming offload NICs. It
> is true that the hardware is using knowledge of the upper layer
> protocols, but it is all within the confines of the implementation
> details. There is no wire protocol implication here.

I think you have quickly glossed over the difference.  The interface created
will need to include a portion of the application.  This is very different
than virtually ever NIC card interface.

> In order to "standardize" these techniques the IETF will be
> dictating hardware design and/or protocol stack design. While
> this might be useful for an industry consortium to make
> portability of software easier, it is not a fundemental IETF
> issue. In what is essentially an implementation specific communications
> between layers of a software stack, how can you standardize
> something that will be uniquely different depending on if
> the OS uses STREAMS, BSD sockets, Windows XYZ, or custom OS frozbits.
> This doesn't make sense for the IETF, you may consider going
> to SNIA to standardize implemenation details.

Again, you are touching on the problem.  Why should the application dictate
the NIC interface?  If this were the common practice, every protocol would
need a portion of the application built into every NIC.  Worse yet, you will
then see this complexity repeated for OS STREAMS, BSD, Windows XYZ and OS
Frozbits.  You will also be required to modify your NIC to accomodate the
next wonderful version of iSCSI.

> If I am not understanding your point, I suggest that it would
> be useful for you to write a couple paragraphs suitable for
> inclusion in a standards document that describes how you
> would standardize the FIM specification. Without such a concrete
> proposal I believe we are all just flailing at amorphous
> ideas and making no progress on this issue.

There is such a document.  At this point it is waiting for feedback from
Stephen Bailey and Allyn Romanow of the RDMA for their edits.  It has been
awhile, so I guess they are fighting other problems.


> 	-David
>


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 14:00:09 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA28599
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 14:00:09 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA20077
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 14:00:11 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA18305;
	Wed, 6 Feb 2002 13:14:14 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA18276
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 13:14:12 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-200.dallas.net [209.44.42.200])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA27387
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 13:14:08 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id MAA11746;
	Wed, 6 Feb 2002 12:14:05 -0600
Date: Wed, 6 Feb 2002 12:14:05 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Douglas Otis <dotis@sanlight.net>
Cc: tsvwg@ietf.org, ips@ece.cmu.edu
Subject: Re: [Tsvwg] RE: iSCSI: No Framing
Message-ID: <20020206121405.A10800@openss7.org>
Reply-To: bidulock@openss7.org
References: <20020206101228.D23964@openss7.org> <NEBBJGDMMLHHCIKHGBEJMEHKCPAA.dotis@sanlight.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2i
In-Reply-To: <NEBBJGDMMLHHCIKHGBEJMEHKCPAA.dotis@sanlight.net>; from dotis@sanlight.net on Wed, Feb 06, 2002 at 09:38:18AM -0800
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Douglas,

On Wed, 06 Feb 2002, Douglas Otis wrote:

> Brian,
> 
> > Lloyd,
> >
> > On Wed, 06 Feb 2002, Lloyd Wood wrote:
> >
> > > On Wed, 6 Feb 2002, Brian F. G. Bidulock wrote:
> > >
> > > > What is lost by throwing layering out the window?
> > >
> > > modularity, ease of understanding, and portability.
> >
> > Modularity can be acheived without layering.  The entire
> > field of OOAD proves this.
> >
> > Ease of understanding is acheived through accurate
> > interface description.  Again OO shows this.
> >
> > Portability can also be acheived through OO approaches.
> >
> > The classical layering of communications functions
> > into functional groupings is not a necessary condition
> > for any of these benefits.
> 
> Unlike a typical play dough program, this stuff is more like working with
> concrete.  A portion of each application demanding a direct placement
> feature would be included in hardware or intelligent adapters if using FIM
> and TCP, or framing and TCP, for the most part.  SCTP allows layering for
> this "clean modularity" needed if one is to design an adapter that need not
> understand the complexities and structures of every application that desires
> this function.  The adapter interface only needs to understand SCTP and not
> the application.  By using the unordered delivery mode, a shim would be able
> to implement generalized structures for including Direct Data Placement or
> even a full implementation of RDMA.  It would be possible for more elaborate
> interfaces to built upon this foundation, but at least this establishes the
> modularity, ease of understanding, and portability desired, if one is going
> to start working in concrete.

Not necessarily...  Nowadays entire OS's and applications are embedded on
the adapter.  The adtapter can have full knowledge of all applications
which it supports and provide only higher level application interfaces
to app blades.  This is the preferred architecture for scalability and
redundancy as well as proving a clean separation between applications
and the application engine.

If some want to agree on how this embedding is going to occur with some
vestige of interoperability, what does it matter?  Would not your time be
better spent furthering DDP/SCTP?  It is certainly not the first time that
IETF has had competing approaches: ultimately the market decides.

--brian

> 
> Unlike the case with FIM and TCP, this would introduce NIC devices that will
> be forever sensitive to even minor changes to structures employed in these
> applications and the desire for this feature does not even begin to end at
> just one application.
> 


-- 
Brian F. G. Bidulock
bidulock@openss7.org
http://www.openss7.org/

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 14:21:06 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA29059
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 14:21:05 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA21374
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 14:21:08 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA20826;
	Wed, 6 Feb 2002 14:06:55 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA20791
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 14:06:52 -0500 (EST)
Received: from prue.eim.surrey.ac.uk (IDENT:exim@prue.eim.surrey.ac.uk [131.227.76.5])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA28748
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 14:06:49 -0500 (EST)
Received: from phaestos.ee.surrey.ac.uk ([131.227.88.14] ident=eep1lw)
	by prue.eim.surrey.ac.uk with esmtp (Exim 3.33 #4)
	id 16YXOp-0002iz-00; Wed, 06 Feb 2002 19:06:31 +0000
Date: Wed, 6 Feb 2002 19:06:28 +0000 (GMT)
From: Lloyd Wood <l.wood@eim.surrey.ac.uk>
X-X-Sender: eep1lw@phaestos.ee.surrey.ac.uk
Reply-To: Lloyd Wood <L.Wood@eim.surrey.ac.uk>
To: "Brian F. G. Bidulock" <bidulock@openss7.org>
cc: Douglas Otis <dotis@sanlight.net>, <tsvwg@ietf.org>, <ips@ece.cmu.edu>
Subject: Re: [Tsvwg] RE: iSCSI: No Framing
In-Reply-To: <20020206121405.A10800@openss7.org>
Message-ID: <Pine.SOL.4.43.0202061844500.333-100000@phaestos.ee.surrey.ac.uk>
Organization: speaking for none
X-url: http://www.ee.surrey.ac.uk/Personal/L.Wood/
X-no-archive: yes
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
X-Scanner: exiscan *16YXOp-0002iz-00*3U6dxmuO98E* (SECM, UniS)
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

On Wed, 6 Feb 2002, Brian F. G. Bidulock wrote:
[much oo psychobabble]
> If some want to agree on how this embedding is going to occur with some
> vestige of interoperability, what does it matter?

so why are you here?

[I simply can't be bothered to talk about how design choices that
 might make sense within some closed oo system become meaningless
 outside that system. Solipsistic ontologies have little intrinsic worth.
 Network layering provides basic axioms, and lacks solipsism.]

L.

<L.Wood@surrey.ac.uk>PGP<http://www.ee.surrey.ac.uk/Personal/L.Wood/>




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 14:34:31 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA29437
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 14:34:31 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA22533
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 14:34:34 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA21545;
	Wed, 6 Feb 2002 14:24:41 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA21516
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 14:24:39 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-200.dallas.net [209.44.42.200])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA29154
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 14:24:35 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id NAA13274;
	Wed, 6 Feb 2002 13:24:23 -0600
Date: Wed, 6 Feb 2002 13:24:23 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Lloyd Wood <l.wood@eim.surrey.ac.uk>
Cc: Douglas Otis <dotis@sanlight.net>, tsvwg@ietf.org, ips@ece.cmu.edu
Subject: Re: [Tsvwg] RE: iSCSI: No Framing
Message-ID: <20020206132423.B10800@openss7.org>
Reply-To: bidulock@openss7.org
References: <20020206121405.A10800@openss7.org> <Pine.SOL.4.43.0202061844500.333-100000@phaestos.ee.surrey.ac.uk>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2i
In-Reply-To: <Pine.SOL.4.43.0202061844500.333-100000@phaestos.ee.surrey.ac.uk>; from l.wood@eim.surrey.ac.uk on Wed, Feb 06, 2002 at 07:06:28PM +0000
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Lloyd,

On Wed, 06 Feb 2002, Lloyd Wood wrote:

> On Wed, 6 Feb 2002, Brian F. G. Bidulock wrote:
> > If some want to agree on how this embedding is going to occur with some
> > vestige of interoperability, what does it matter?
> 
> so why are you here?
> 

For SCTP.

--brian

-- 
Brian F. G. Bidulock
bidulock@openss7.org
http://www.openss7.org/

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 14:46:38 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA29696
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 14:46:38 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA23246
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 14:46:41 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA20660;
	Wed, 6 Feb 2002 14:03:56 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA20629
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 14:03:54 -0500 (EST)
Received: from c003.snv.cp.net (c003-h013.c003.snv.cp.net [209.228.32.227])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id OAA28666
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 14:03:50 -0500 (EST)
Received: (cpmta 29599 invoked from network); 6 Feb 2002 11:03:21 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.227) with SMTP; 6 Feb 2002 11:03:21 -0800
X-Sent: 6 Feb 2002 19:03:21 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <tsvwg@ietf.org>
Cc: <ips@ece.cmu.edu>
Subject: RE: [Tsvwg] RE: iSCSI: No Framing
Date: Wed, 6 Feb 2002 11:05:10 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJGEHNCPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Importance: Normal
In-Reply-To: <20020206121405.A10800@openss7.org>
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Brian,

>
> Douglas,
>
> On Wed, 06 Feb 2002, Douglas Otis wrote:
>
> > Brian,
> >
> > > Lloyd,
> > >
> > > On Wed, 06 Feb 2002, Lloyd Wood wrote:
> > >
> > > > On Wed, 6 Feb 2002, Brian F. G. Bidulock wrote:
> > > >
> > > > > What is lost by throwing layering out the window?
> > > >
> > > > modularity, ease of understanding, and portability.
> > >
> > > Modularity can be acheived without layering.  The entire
> > > field of OOAD proves this.
> > >
> > > Ease of understanding is acheived through accurate
> > > interface description.  Again OO shows this.
> > >
> > > Portability can also be acheived through OO approaches.
> > >
> > > The classical layering of communications functions
> > > into functional groupings is not a necessary condition
> > > for any of these benefits.
> >
> > Unlike a typical play dough program, this stuff is more like
> > working with concrete.  A portion of each application demanding
> > a direct placement feature would be included in hardware or
> > intelligent adapters if using FIM and TCP, or framing and TCP,
> > for the most part.  SCTP allows layering for this "clean
> > modularity" needed if one is to design an adapter that need not
> > understand the complexities and structures of every application
> > that desires this function.  The adapter interface only needs to
> > understand SCTP and not the application.  By using the unordered
> > delivery mode, a shim would be able to implement generalized
> > structures for including Direct Data Placement or even a full
> > implementation of RDMA.  It would be possible for more elaborate
> > interfaces to built upon this foundation, but at least this
> > establishes the modularity, ease of understanding, and portability
> > desired, if one is going to start working in concrete.
>
> Not necessarily...  Nowadays entire OS's and applications are embedded on
> the adapter.  The adtapter can have full knowledge of all applications
> which it supports and provide only higher level application interfaces
> to app blades.  This is the preferred architecture for scalability and
> redundancy as well as proving a clean separation between applications
> and the application engine.

The rationale for FIM was to avoid the rather modest use of adapter buffer
space.  Now you are suggesting that the adapter will include the entire OS
and application.  It sounds as if you are describing clustering where I
think RDMA is an attempt to standardize this type of interface.

> If some want to agree on how this embedding is going to occur with some
> vestige of interoperability, what does it matter?  Would not your time be
> better spent furthering DDP/SCTP?  It is certainly not the first time that
> IETF has had competing approaches: ultimately the market decides.

This change in architecture is profound with respect to impact on competing
approaches.  I would not wish to under estimate leverage found from storage
as it can be used to wrest control away from standards into an array of
vendor unique solutions.  This FIM/TCP scheme adds a layer beneath the
transport and either modifies the layer above or adds a new interface to
this added layer.  Either way, the definition of what is included within a
transport is greatly diminished.  You are asking me to turn my back to this
while I see it in my interest to keep this interface open and I see SCTP as
a viable solution for doing so.  Restricting iSCSI to TCP would not prevent
future use of SCTP for this application provided this interface remains
open.  If you are right about what can be included within an adapter, then
justification for FIM/TCP vanishes.

Doug

> --brian


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb  6 21:28:05 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA07650
	for <tsvwg-archive@odin.ietf.org>; Wed, 6 Feb 2002 21:28:04 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id VAA12562
	for tsvwg-archive@odin.ietf.org; Wed, 6 Feb 2002 21:28:08 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA12124;
	Wed, 6 Feb 2002 21:11:16 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA12093
	for <tsvwg@optimus.ietf.org>; Wed, 6 Feb 2002 21:11:13 -0500 (EST)
Received: from d12lmsgate-2.de.ibm.com (d12lmsgate-2.de.ibm.com [195.212.91.200])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA07382
	for <tsvwg@ietf.org>; Wed, 6 Feb 2002 21:11:09 -0500 (EST)
Received: from d12relay01.de.ibm.com (d12relay01.de.ibm.com [9.165.215.22])
	by d12lmsgate-2.de.ibm.com (1.0.0) with ESMTP id DAA54500;
	Thu, 7 Feb 2002 03:10:38 +0100
Received: from d10hubm1.telaviv.ibm.com (d10ml001.telaviv.ibm.com [9.148.216.55])
	by d12relay01.de.ibm.com (8.11.1m3/NCO v5.01) with ESMTP id g172C4w114228;
	Thu, 7 Feb 2002 03:12:04 +0100
To: "Martin, Nick" <Nick.Martin@COMPAQ.com>
Cc: tsvwg@ietf.org, rdma@yahoogroups.com
X-Mailer: Lotus Notes Release 5.0.9  November 16, 2001
Message-ID: <OFF95150BD.4846CBB3-ONC2256B59.000B3D0B@telaviv.ibm.com>
From: "Julian Satran" <Julian_Satran@il.ibm.com>
Date: Thu, 7 Feb 2002 04:08:22 +0200
X-MIMETrack: Serialize by Router on D10ML001/10/M/IBM(Release 5.0.8 |June 18, 2001) at
 07/02/2002 04:12:06
MIME-Version: 1.0
Content-type: text/plain; charset=us-ascii
Subject: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


Martin,

I am glad you like it.
The salt is a random number (a RN generator generates a new one at every
PDU). That is why it is hard to guess.
PDU headers are not random enough. It is essential that it is random and
different in every frame.
For iSCSI this comes a bit late but for a "generalized framing" I think
that this is a good solution with a very low price tag.

Thanks,
Julo





                                                                                             
                    "Martin, Nick"                                                           
                    <Nick.Martin@C       To:     Julian Satran/Haifa/IBM@IBMIL               
                    OMPAQ.com>           cc:                                                 
                                         Subject:     RE: [rdma] a proposal for a different  
                    06-02-02 02:42        No-Data-Touch framer for TCP                       
                                                                                             
                                                                                             
                                                                                             



Julian,

I *really* like this proposal.

Here is what I think I understand about the NDT framing draft
(un-optimized).  The two ends would agree upon a unique 32-bit value to be
used as the framing pattern.  The next 32 bits are random, and the next
32-bits are a digest.  This effectively defines 2**32 sets each containing
2**32 valid 96-bit frame markers.  Each connection will use markers from
one of the 2**32 sets.  If a secret value is added, there become 2**64 sets
each containing 2**32 valid 96-bit frame markers.  Entities other than the
intended recipient are less likely to know whether the digest is correct.
However, with each frame beginning with the same framing pattern, and a
published standard for how to compute the digest, it still seems fairly
easy to synthesize the "secret" on a long living connection.

I like NDT (optimized) for iSCSI framing.  Except for the salt and shared
secret, this version looks equivalent to having each iSCSI PDU header start
with a framing pattern, and including that pattern in the header digest.
If someone asked me how to do framing for iSCSI, that is what I would have
done.  [Start of PDU marker without salt or secret, turn on digests].  It
is simple and obvious how to do this in either software or hardware, with
or without turning your network stack inside out. [I have had difficulty
understanding why all the more complex proposals were desired.]  As you
point out, statistically the 32-bit framing pattern alone, even if required
to be on a frame boundary, can easily be fooled into making a mistake.
However, with ULP (iSCSI) digests enabled, the probability of an undetected
mistake is much lower.  [I would say acceptable.]  Given that digests are
enabled, I believe the probability of an undetected mistake with the 32-bit
framing pattern included in the digest is about the same as with the
framing pattern excluded from the digest.  Just adding the framing pattern
to iSCSI requires only 32-bits per PDU.  Adding the framing pattern and
salt is only 64-bits.

For an iSCSI software implementation on top of an unmodified [Operating
System Vendor provided] TCP stack, I can not take advantage of out of order
placement and also may not be able to detect TCP/IP packet boundaries.
When discarding a PDU header with a bad digest possibly followed by data of
un-trusted length, I would like to be able to quickly recover/confirm the
next PDU boundary using a framing marker and a digest, rather than
computing the iSCSI digest at each succeeding 4 byte boundary.  [I would
*not* like to have to generate, nor to locate and discard, framing markers
in the middle of otherwise perfectly good PDUs.]

I have a couple of questions about the salt.  Are there any rules for
generating it?  If the rest of the ULP PDU header is included in the
digest, would this not be as effective as the salt?  When a shared secret
or other data value not transmitted with the PDU is included in the digest
calculation, is this as effective as adding salt, so that explicit salt
would not be needed?  If the secret and/or the ULP header can replace the
salt, the frame markers overhead shrinks another 32-bits.

I also appreciate that the un-optimized version of the NDT framer can
easily and efficiently be inserted above or below almost any other protocol
layer by hardware or software.  On the outbound side, it does not need to
be told the length of the next ULP PDU.  It only needs to know when a PDU
boundary (either beginning, ending, or both) has been reached.

Thanks,
Nick

P.S. I did not copy this message the reflector.  You are free to do what
you please with it.
 -----Original Message-----
 From: Julian Satran [mailto:julian_satran@il.ibm.com]
 Sent: Saturday, February 02, 2002 3:49 AM
 To: tsvwg@ietf.org; rdma@yahoogroups.com
 Subject: [rdma] a proposal for a different No-Data-Touch framer for TCP


 Dear Colleagues:

 I've recently submitted a draft including a different "No-Data-Touch"(NDT)
 framer.
 Unlike the current proposed NDT framer that has a "mistake" probability
 that increases asymptotically (to 1) with the stream length the framer I
 am suggesting has a very low and constant mistake probability and is thus
 safe to use even in the absence of alignment or in the presence of
 re-segmentation.

 It does not contain the standard RFC2026 reference not as an indication of
 patent - I have no  patent on it and I do not know about the existence of
 one (nor did I search thoroughly) - but rather as a sign of discomfort for
 it to be used without me being asked about it.

 Until it apperas you can find it also at:

 http://www.haifa.il.ibm.com/satran/ips/draft-satran-NDT-Framing-00.pdf

 http://www.haifa.il.ibm.com/satran/ips/draft-satran-NDT-Framing-00.txt

 Julo
                                   
       Yahoo! Groups Sponsor       
                                   
                                   
                                   
           ADVERTISEMENT           
                                   
                                   
                                   
                                   
                                   


 To unsubscribe from this group, send an email to:
 rdma-unsubscribe@egroups.com



 Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb  7 09:26:35 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA27778
	for <tsvwg-archive@odin.ietf.org>; Thu, 7 Feb 2002 09:26:30 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA22310
	for tsvwg-archive@odin.ietf.org; Thu, 7 Feb 2002 09:26:31 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA21461;
	Thu, 7 Feb 2002 09:07:19 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA21369
	for <tsvwg@optimus.ietf.org>; Thu, 7 Feb 2002 09:07:15 -0500 (EST)
Received: from ztxmail05.ztx.compaq.com (ztxmail05.ztx.compaq.com [161.114.1.209])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA26993
	for <tsvwg@ietf.org>; Thu, 7 Feb 2002 09:07:14 -0500 (EST)
Received: from cceexg11.americas.cpqcorp.net (cceexg11.americas.cpqcorp.net [16.110.250.125])
	by ztxmail05.ztx.compaq.com (Postfix) with ESMTP
	id 567FE3FF3; Thu,  7 Feb 2002 08:06:46 -0600 (CST)
Received: from cceexc18.americas.cpqcorp.net ([16.110.250.64]) by cceexg11.americas.cpqcorp.net with Microsoft SMTPSVC(5.0.2195.2966);
	 Thu, 7 Feb 2002 08:06:46 -0600
X-MimeOLE: Produced By Microsoft Exchange V6.0.4712.0
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="US-ASCII"
Date: Thu, 7 Feb 2002 08:06:45 -0600
Message-ID: <4D027986353D1341ADA4F2A4F8A170760E3184@cceexc18.americas.cpqcorp.net>
Thread-Topic: [rdma] a proposal for a different No-Data-Touch framer for TCP
Thread-Index: AcGvfKfylx8fZnTXRQKzAPb0CorHqAAYDG5A
From: "Culley, Paul" <Paul.Culley@COMPAQ.com>
To: "Julian Satran" <julian_satran@il.ibm.com>
Cc: <tsvwg@ietf.org>, <rdma@yahoogroups.com>
X-OriginalArrivalTime: 07 Feb 2002 14:06:46.0143 (UTC) FILETIME=[AD7F70F0:01C1AFE0]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id JAA21370
Subject: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

I am having a hard time understanding the advantage of the "salt", as
opposed to a longer framing key.  Both would seem to have the same
probability of false recognition.  Neither protects against the extreme
case where an application places a valid frame within the ULP data (I
assume that the application knows the frame key, adds its own salt, that
it can generate a CRC etc.).

I also presume that reasonable implementations will chose the framing
key randomly, perhaps with a bit of selectivity, like not using values
with leading zeros etc.

Also, do you have a probability of false recognition that you are
shooting for?  How does it compare to the probability of other types of
false negative errors associated with TCP with a PDU CRC?

For example, has anyone looked at what happens if a (lightly protected)
TCP sequence number is corrupted, possibly while a packet is also
dropped?

Given a few other rules around out of order placement which also improve
the odds, I personally feel like this issue is being over designed.
Rules like:
 1) don't actively search for frames, unless it appears on a segment
boundary, wait for previous segments to arrive so the correct boundary
can be calculated.  (resegmenting middle boxes kill out of order
placement).
 2) don't place unless its in the TCP RWIN
 3) ULP checks should still be made, and failures could cause the
implementation to drop/wait until framing can be assured. (bigger layer
violation, for more paranoid designers).


Paul R. Culley
Compaq Sr. Fellow
281-514-5543


-----Original Message-----
From: Julian Satran [mailto:julian_satran@il.ibm.com]
Sent: Wednesday, February 06, 2002 8:08 PM
To: Martin, Nick
Cc: tsvwg@ietf.org; rdma@yahoogroups.com
Subject: RE: [rdma] a proposal for a different No-Data-Touch framer for
TCP



Martin,

I am glad you like it.
The salt is a random number (a RN generator generates a new one at every
PDU). That is why it is hard to guess.
PDU headers are not random enough. It is essential that it is random and
different in every frame.
For iSCSI this comes a bit late but for a "generalized framing" I think
that this is a good solution with a very low price tag.

Thanks,
Julo





 

                    "Martin, Nick"

                    <Nick.Martin@C       To:     Julian
Satran/Haifa/IBM@IBMIL               
                    OMPAQ.com>           cc:

                                         Subject:     RE: [rdma] a
proposal for a different  
                    06-02-02 02:42        No-Data-Touch framer for TCP

 

 

 




Julian,

I *really* like this proposal.

Here is what I think I understand about the NDT framing draft
(un-optimized).  The two ends would agree upon a unique 32-bit value to
be
used as the framing pattern.  The next 32 bits are random, and the next
32-bits are a digest.  This effectively defines 2**32 sets each
containing
2**32 valid 96-bit frame markers.  Each connection will use markers from
one of the 2**32 sets.  If a secret value is added, there become 2**64
sets
each containing 2**32 valid 96-bit frame markers.  Entities other than
the
intended recipient are less likely to know whether the digest is
correct.
However, with each frame beginning with the same framing pattern, and a
published standard for how to compute the digest, it still seems fairly
easy to synthesize the "secret" on a long living connection.

I like NDT (optimized) for iSCSI framing.  Except for the salt and
shared
secret, this version looks equivalent to having each iSCSI PDU header
start
with a framing pattern, and including that pattern in the header digest.
If someone asked me how to do framing for iSCSI, that is what I would
have
done.  [Start of PDU marker without salt or secret, turn on digests].
It
is simple and obvious how to do this in either software or hardware,
with
or without turning your network stack inside out. [I have had difficulty
understanding why all the more complex proposals were desired.]  As you
point out, statistically the 32-bit framing pattern alone, even if
required
to be on a frame boundary, can easily be fooled into making a mistake.
However, with ULP (iSCSI) digests enabled, the probability of an
undetected
mistake is much lower.  [I would say acceptable.]  Given that digests
are
enabled, I believe the probability of an undetected mistake with the
32-bit
framing pattern included in the digest is about the same as with the
framing pattern excluded from the digest.  Just adding the framing
pattern
to iSCSI requires only 32-bits per PDU.  Adding the framing pattern and
salt is only 64-bits.

For an iSCSI software implementation on top of an unmodified [Operating
System Vendor provided] TCP stack, I can not take advantage of out of
order
placement and also may not be able to detect TCP/IP packet boundaries.
When discarding a PDU header with a bad digest possibly followed by data
of
un-trusted length, I would like to be able to quickly recover/confirm
the
next PDU boundary using a framing marker and a digest, rather than
computing the iSCSI digest at each succeeding 4 byte boundary.  [I would
*not* like to have to generate, nor to locate and discard, framing
markers
in the middle of otherwise perfectly good PDUs.]

I have a couple of questions about the salt.  Are there any rules for
generating it?  If the rest of the ULP PDU header is included in the
digest, would this not be as effective as the salt?  When a shared
secret
or other data value not transmitted with the PDU is included in the
digest
calculation, is this as effective as adding salt, so that explicit salt
would not be needed?  If the secret and/or the ULP header can replace
the
salt, the frame markers overhead shrinks another 32-bits.

I also appreciate that the un-optimized version of the NDT framer can
easily and efficiently be inserted above or below almost any other
protocol
layer by hardware or software.  On the outbound side, it does not need
to
be told the length of the next ULP PDU.  It only needs to know when a
PDU
boundary (either beginning, ending, or both) has been reached.

Thanks,
Nick

P.S. I did not copy this message the reflector.  You are free to do what
you please with it.
 -----Original Message-----
 From: Julian Satran [mailto:julian_satran@il.ibm.com]
 Sent: Saturday, February 02, 2002 3:49 AM
 To: tsvwg@ietf.org; rdma@yahoogroups.com
 Subject: [rdma] a proposal for a different No-Data-Touch framer for TCP


 Dear Colleagues:

 I've recently submitted a draft including a different
"No-Data-Touch"(NDT)
 framer.
 Unlike the current proposed NDT framer that has a "mistake" probability
 that increases asymptotically (to 1) with the stream length the framer
I
 am suggesting has a very low and constant mistake probability and is
thus
 safe to use even in the absence of alignment or in the presence of
 re-segmentation.

 It does not contain the standard RFC2026 reference not as an indication
of
 patent - I have no  patent on it and I do not know about the existence
of
 one (nor did I search thoroughly) - but rather as a sign of discomfort
for
 it to be used without me being asked about it.

 Until it apperas you can find it also at:

 http://www.haifa.il.ibm.com/satran/ips/draft-satran-NDT-Framing-00.pdf

 http://www.haifa.il.ibm.com/satran/ips/draft-satran-NDT-Framing-00.txt

 Julo
                                   
       Yahoo! Groups Sponsor       
                                   
                                   
                                   
           ADVERTISEMENT           
                                   
                                   
                                   
                                   
                                   


 To unsubscribe from this group, send an email to:
 rdma-unsubscribe@egroups.com



 Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.




------------------------ Yahoo! Groups Sponsor ---------------------~-->
Tiny Wireless Camera under $80!
Order Now! FREE VCR Commander!
Click Here - Only 1 Day Left!
http://us.click.yahoo.com/nuyOHD/7.PDAA/yigFAA/W6uqlB/TM
---------------------------------------------------------------------~->

To unsubscribe from this group, send an email to:
rdma-unsubscribe@egroups.com

 

Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/ 



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Thu Feb  7 15:40:58 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA09214
	for <tsvwg-archive@odin.ietf.org>; Thu, 7 Feb 2002 15:40:58 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA16243
	for tsvwg-archive@odin.ietf.org; Thu, 7 Feb 2002 15:40:59 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA15422;
	Thu, 7 Feb 2002 15:28:52 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA15358
	for <tsvwg@ns.ietf.org>; Thu, 7 Feb 2002 15:28:48 -0500 (EST)
Received: from d12lmsgate-3.de.ibm.com (d12lmsgate-3.de.ibm.com [195.212.91.201])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA08807
	for <tsvwg@ietf.org>; Thu, 7 Feb 2002 15:28:44 -0500 (EST)
Received: from d12relay01.de.ibm.com (d12relay01.de.ibm.com [9.165.215.22])
	by d12lmsgate-3.de.ibm.com (1.0.0) with ESMTP id VAA77464;
	Thu, 7 Feb 2002 21:28:11 +0100
Received: from d10hubm1.telaviv.ibm.com (d10ml001.telaviv.ibm.com [9.148.216.55])
	by d12relay01.de.ibm.com (8.11.1m3/NCO v5.01) with ESMTP id g17KTdA58560;
	Thu, 7 Feb 2002 21:29:39 +0100
To: <tsvwg@ietf.org>, <rdma@yahoogroups.com>
X-Mailer: Lotus Notes Release 5.0.9  November 16, 2001
Message-ID: <OF5D0D4C86.1A657679-ONC2256B59.006FCDAC@telaviv.ibm.com>
From: "Julian Satran" <Julian_Satran@il.ibm.com>
Date: Thu, 7 Feb 2002 22:21:35 +0200
X-MIMETrack: Serialize by Router on D10ML001/10/M/IBM(Release 5.0.8 |June 18, 2001) at
 07/02/2002 22:29:41
MIME-Version: 1.0
Content-type: text/plain; charset=us-ascii
Subject: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org



And in addition even if the frame key is random  it does not change for the
life of the stream - and that makes it impractical.

Julo


                                                                                             
                    "Culley, Paul"                                                           
                    <paul.culley@c       To:     Julian Satran/Haifa/IBM@IBMIL               
                    ompaq.com>           cc:     <tsvwg@ietf.org>, <rdma@yahoogroups.com>    
                                         Subject:     RE: [rdma] a proposal for a different  
                    07-02-02 16:06        No-Data-Touch framer for TCP                       
                                                                                             
                                                                                             
                                                                                             



I am having a hard time understanding the advantage of the "salt", as
opposed to a longer framing key.  Both would seem to have the same
probability of false recognition.  Neither protects against the extreme
case where an application places a valid frame within the ULP data (I
assume that the application knows the frame key, adds its own salt, that
it can generate a CRC etc.).

I also presume that reasonable implementations will chose the framing
key randomly, perhaps with a bit of selectivity, like not using values
with leading zeros etc.

Also, do you have a probability of false recognition that you are
shooting for?  How does it compare to the probability of other types of
false negative errors associated with TCP with a PDU CRC?

For example, has anyone looked at what happens if a (lightly protected)
TCP sequence number is corrupted, possibly while a packet is also
dropped?

Given a few other rules around out of order placement which also improve
the odds, I personally feel like this issue is being over designed.
Rules like:
 1) don't actively search for frames, unless it appears on a segment
boundary, wait for previous segments to arrive so the correct boundary
can be calculated.  (resegmenting middle boxes kill out of order
placement).
 2) don't place unless its in the TCP RWIN
 3) ULP checks should still be made, and failures could cause the
implementation to drop/wait until framing can be assured. (bigger layer
violation, for more paranoid designers).


Paul R. Culley
Compaq Sr. Fellow
281-514-5543


-----Original Message-----
From: Julian Satran [mailto:julian_satran@il.ibm.com]
Sent: Wednesday, February 06, 2002 8:08 PM
To: Martin, Nick
Cc: tsvwg@ietf.org; rdma@yahoogroups.com
Subject: RE: [rdma] a proposal for a different No-Data-Touch framer for
TCP



Martin,

I am glad you like it.
The salt is a random number (a RN generator generates a new one at every
PDU). That is why it is hard to guess.
PDU headers are not random enough. It is essential that it is random and
different in every frame.
For iSCSI this comes a bit late but for a "generalized framing" I think
that this is a good solution with a very low price tag.

Thanks,
Julo







                    "Martin, Nick"

                    <Nick.Martin@C       To:     Julian
Satran/Haifa/IBM@IBMIL
                    OMPAQ.com>           cc:

                                         Subject:     RE: [rdma] a
proposal for a different
                    06-02-02 02:42        No-Data-Touch framer for TCP










Julian,

I *really* like this proposal.

Here is what I think I understand about the NDT framing draft
(un-optimized).  The two ends would agree upon a unique 32-bit value to
be
used as the framing pattern.  The next 32 bits are random, and the next
32-bits are a digest.  This effectively defines 2**32 sets each
containing
2**32 valid 96-bit frame markers.  Each connection will use markers from
one of the 2**32 sets.  If a secret value is added, there become 2**64
sets
each containing 2**32 valid 96-bit frame markers.  Entities other than
the
intended recipient are less likely to know whether the digest is
correct.
However, with each frame beginning with the same framing pattern, and a
published standard for how to compute the digest, it still seems fairly
easy to synthesize the "secret" on a long living connection.

I like NDT (optimized) for iSCSI framing.  Except for the salt and
shared
secret, this version looks equivalent to having each iSCSI PDU header
start
with a framing pattern, and including that pattern in the header digest.
If someone asked me how to do framing for iSCSI, that is what I would
have
done.  [Start of PDU marker without salt or secret, turn on digests].
It
is simple and obvious how to do this in either software or hardware,
with
or without turning your network stack inside out. [I have had difficulty
understanding why all the more complex proposals were desired.]  As you
point out, statistically the 32-bit framing pattern alone, even if
required
to be on a frame boundary, can easily be fooled into making a mistake.
However, with ULP (iSCSI) digests enabled, the probability of an
undetected
mistake is much lower.  [I would say acceptable.]  Given that digests
are
enabled, I believe the probability of an undetected mistake with the
32-bit
framing pattern included in the digest is about the same as with the
framing pattern excluded from the digest.  Just adding the framing
pattern
to iSCSI requires only 32-bits per PDU.  Adding the framing pattern and
salt is only 64-bits.

For an iSCSI software implementation on top of an unmodified [Operating
System Vendor provided] TCP stack, I can not take advantage of out of
order
placement and also may not be able to detect TCP/IP packet boundaries.
When discarding a PDU header with a bad digest possibly followed by data
of
un-trusted length, I would like to be able to quickly recover/confirm
the
next PDU boundary using a framing marker and a digest, rather than
computing the iSCSI digest at each succeeding 4 byte boundary.  [I would
*not* like to have to generate, nor to locate and discard, framing
markers
in the middle of otherwise perfectly good PDUs.]

I have a couple of questions about the salt.  Are there any rules for
generating it?  If the rest of the ULP PDU header is included in the
digest, would this not be as effective as the salt?  When a shared
secret
or other data value not transmitted with the PDU is included in the
digest
calculation, is this as effective as adding salt, so that explicit salt
would not be needed?  If the secret and/or the ULP header can replace
the
salt, the frame markers overhead shrinks another 32-bits.

I also appreciate that the un-optimized version of the NDT framer can
easily and efficiently be inserted above or below almost any other
protocol
layer by hardware or software.  On the outbound side, it does not need
to
be told the length of the next ULP PDU.  It only needs to know when a
PDU
boundary (either beginning, ending, or both) has been reached.

Thanks,
Nick

P.S. I did not copy this message the reflector.  You are free to do what
you please with it.
 -----Original Message-----
 From: Julian Satran [mailto:julian_satran@il.ibm.com]
 Sent: Saturday, February 02, 2002 3:49 AM
 To: tsvwg@ietf.org; rdma@yahoogroups.com
 Subject: [rdma] a proposal for a different No-Data-Touch framer for TCP


 Dear Colleagues:

 I've recently submitted a draft including a different
"No-Data-Touch"(NDT)
 framer.
 Unlike the current proposed NDT framer that has a "mistake" probability
 that increases asymptotically (to 1) with the stream length the framer
I
 am suggesting has a very low and constant mistake probability and is
thus
 safe to use even in the absence of alignment or in the presence of
 re-segmentation.

 It does not contain the standard RFC2026 reference not as an indication
of
 patent - I have no  patent on it and I do not know about the existence
of
 one (nor did I search thoroughly) - but rather as a sign of discomfort
for
 it to be used without me being asked about it.

 Until it apperas you can find it also at:

 http://www.haifa.il.ibm.com/satran/ips/draft-satran-NDT-Framing-00.pdf

 http://www.haifa.il.ibm.com/satran/ips/draft-satran-NDT-Framing-00.txt

 Julo

       Yahoo! Groups Sponsor



           ADVERTISEMENT







 To unsubscribe from this group, send an email to:
 rdma-unsubscribe@egroups.com



 Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.





To unsubscribe from this group, send an email to:
rdma-unsubscribe@egroups.com



Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/



------------------------ Yahoo! Groups Sponsor ---------------------~-->
Sponsored by VeriSign - The Value of Trust
Secure all your Web servers now - with a proven 5-part
strategy. The FREE Server Security Guide shows you how.
http://us.click.yahoo.com/iWSNbC/VdiDAA/yigFAA/W6uqlB/TM
---------------------------------------------------------------------~->

To unsubscribe from this group, send an email to:
rdma-unsubscribe@egroups.com



Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/









_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Thu Feb  7 16:08:16 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA09954
	for <tsvwg-archive@odin.ietf.org>; Thu, 7 Feb 2002 16:08:16 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA17380
	for tsvwg-archive@odin.ietf.org; Thu, 7 Feb 2002 16:08:17 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA16611;
	Thu, 7 Feb 2002 15:54:50 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA16580
	for <tsvwg@ns.ietf.org>; Thu, 7 Feb 2002 15:54:48 -0500 (EST)
Received: from prue.eim.surrey.ac.uk (IDENT:exim@prue.eim.surrey.ac.uk [131.227.76.5])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA09574
	for <tsvwg@ietf.org>; Thu, 7 Feb 2002 15:54:46 -0500 (EST)
Received: from phaestos.ee.surrey.ac.uk ([131.227.88.14] ident=eep1lw)
	by prue.eim.surrey.ac.uk with esmtp (Exim 3.33 #4)
	id 16YvYx-0007Nr-00; Thu, 07 Feb 2002 20:54:35 +0000
Date: Thu, 7 Feb 2002 20:54:32 +0000 (GMT)
From: Lloyd Wood <l.wood@eim.surrey.ac.uk>
X-X-Sender: eep1lw@phaestos.ee.surrey.ac.uk
Reply-To: Lloyd Wood <L.Wood@eim.surrey.ac.uk>
To: Julian Satran <Julian_Satran@il.ibm.com>
cc: tsvwg <tsvwg@ietf.org>, rdma <rdma@yahoogroups.com>
Subject: Re: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch
 framer for TCP
In-Reply-To: <OF8BB94B11.B145A3E1-ONC2256B59.006FE94A@telaviv.ibm.com>
Message-ID: <Pine.SOL.4.43.0202072053330.2665-100000@phaestos.ee.surrey.ac.uk>
Organization: speaking for none
X-url: http://www.ee.surrey.ac.uk/Personal/L.Wood/
X-no-archive: yes
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
X-Scanner: exiscan *16YvYx-0007Nr-00*d5kUBcz5HCY* (SECM, UniS)
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

On Thu, 7 Feb 2002, Julian Satran wrote:

> The salt is a random number that is changing on every PDU.

you mean a nonce?

L.

<L.Wood@surrey.ac.uk>PGP<http://www.ee.surrey.ac.uk/Personal/L.Wood/>


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Thu Feb  7 16:40:43 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA09225
	for <tsvwg-archive@odin.ietf.org>; Thu, 7 Feb 2002 15:40:59 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA16257
	for tsvwg-archive@odin.ietf.org; Thu, 7 Feb 2002 15:41:00 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA15391;
	Thu, 7 Feb 2002 15:28:51 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA15360
	for <tsvwg@ns.ietf.org>; Thu, 7 Feb 2002 15:28:49 -0500 (EST)
Received: from d12lmsgate-3.de.ibm.com (d12lmsgate-3.de.ibm.com [195.212.91.201])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA08808
	for <tsvwg@ietf.org>; Thu, 7 Feb 2002 15:28:47 -0500 (EST)
Received: from d12relay01.de.ibm.com (d12relay01.de.ibm.com [9.165.215.22])
	by d12lmsgate-3.de.ibm.com (1.0.0) with ESMTP id VAA06216;
	Thu, 7 Feb 2002 21:28:16 +0100
Received: from d10hubm1.telaviv.ibm.com (d10ml001.telaviv.ibm.com [9.148.216.55])
	by d12relay01.de.ibm.com (8.11.1m3/NCO v5.01) with ESMTP id g17KThA19476;
	Thu, 7 Feb 2002 21:29:43 +0100
To: <tsvwg@ietf.org>, <rdma@yahoogroups.com>
X-Mailer: Lotus Notes Release 5.0.9  November 16, 2001
Message-ID: <OF8BB94B11.B145A3E1-ONC2256B59.006FE94A@telaviv.ibm.com>
From: "Julian Satran" <Julian_Satran@il.ibm.com>
Date: Thu, 7 Feb 2002 22:22:39 +0200
X-MIMETrack: Serialize by Router on D10ML001/10/M/IBM(Release 5.0.8 |June 18, 2001) at
 07/02/2002 22:29:45
MIME-Version: 1.0
Content-type: text/plain; charset=us-ascii
Subject: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org



Paul,

The salt is a random number that is changing on every PDU. That changes
radically the false detection probability in face on an infinite stream
form certainty to a value that is acceptable from engineering point of
view. Neither protects against malicious user. However the probability of
random good framer in the data becomes nil.

Julo


                                                                                             
                    "Culley, Paul"                                                           
                    <Paul.Culley@C       To:     Julian Satran/Haifa/IBM@IBMIL               
                    OMPAQ.com>           cc:     <tsvwg@ietf.org>, <rdma@yahoogroups.com>    
                                         Subject:     RE: [rdma] a proposal for a different  
                    07-02-02 16:06        No-Data-Touch framer for TCP                       
                                                                                             
                                                                                             
                                                                                             



I am having a hard time understanding the advantage of the "salt", as
opposed to a longer framing key.  Both would seem to have the same
probability of false recognition.  Neither protects against the extreme
case where an application places a valid frame within the ULP data (I
assume that the application knows the frame key, adds its own salt, that
it can generate a CRC etc.).

I also presume that reasonable implementations will chose the framing
key randomly, perhaps with a bit of selectivity, like not using values
with leading zeros etc.

Also, do you have a probability of false recognition that you are
shooting for?  How does it compare to the probability of other types of
false negative errors associated with TCP with a PDU CRC?

For example, has anyone looked at what happens if a (lightly protected)
TCP sequence number is corrupted, possibly while a packet is also
dropped?

Given a few other rules around out of order placement which also improve
the odds, I personally feel like this issue is being over designed.
Rules like:
 1) don't actively search for frames, unless it appears on a segment
boundary, wait for previous segments to arrive so the correct boundary
can be calculated.  (resegmenting middle boxes kill out of order
placement).
 2) don't place unless its in the TCP RWIN
 3) ULP checks should still be made, and failures could cause the
implementation to drop/wait until framing can be assured. (bigger layer
violation, for more paranoid designers).


Paul R. Culley
Compaq Sr. Fellow
281-514-5543


-----Original Message-----
From: Julian Satran [mailto:julian_satran@il.ibm.com]
Sent: Wednesday, February 06, 2002 8:08 PM
To: Martin, Nick
Cc: tsvwg@ietf.org; rdma@yahoogroups.com
Subject: RE: [rdma] a proposal for a different No-Data-Touch framer for
TCP



Martin,

I am glad you like it.
The salt is a random number (a RN generator generates a new one at every
PDU). That is why it is hard to guess.
PDU headers are not random enough. It is essential that it is random and
different in every frame.
For iSCSI this comes a bit late but for a "generalized framing" I think
that this is a good solution with a very low price tag.

Thanks,
Julo







                    "Martin, Nick"

                    <Nick.Martin@C       To:     Julian
Satran/Haifa/IBM@IBMIL
                    OMPAQ.com>           cc:

                                         Subject:     RE: [rdma] a
proposal for a different
                    06-02-02 02:42        No-Data-Touch framer for TCP










Julian,

I *really* like this proposal.

Here is what I think I understand about the NDT framing draft
(un-optimized).  The two ends would agree upon a unique 32-bit value to
be
used as the framing pattern.  The next 32 bits are random, and the next
32-bits are a digest.  This effectively defines 2**32 sets each
containing
2**32 valid 96-bit frame markers.  Each connection will use markers from
one of the 2**32 sets.  If a secret value is added, there become 2**64
sets
each containing 2**32 valid 96-bit frame markers.  Entities other than
the
intended recipient are less likely to know whether the digest is
correct.
However, with each frame beginning with the same framing pattern, and a
published standard for how to compute the digest, it still seems fairly
easy to synthesize the "secret" on a long living connection.

I like NDT (optimized) for iSCSI framing.  Except for the salt and
shared
secret, this version looks equivalent to having each iSCSI PDU header
start
with a framing pattern, and including that pattern in the header digest.
If someone asked me how to do framing for iSCSI, that is what I would
have
done.  [Start of PDU marker without salt or secret, turn on digests].
It
is simple and obvious how to do this in either software or hardware,
with
or without turning your network stack inside out. [I have had difficulty
understanding why all the more complex proposals were desired.]  As you
point out, statistically the 32-bit framing pattern alone, even if
required
to be on a frame boundary, can easily be fooled into making a mistake.
However, with ULP (iSCSI) digests enabled, the probability of an
undetected
mistake is much lower.  [I would say acceptable.]  Given that digests
are
enabled, I believe the probability of an undetected mistake with the
32-bit
framing pattern included in the digest is about the same as with the
framing pattern excluded from the digest.  Just adding the framing
pattern
to iSCSI requires only 32-bits per PDU.  Adding the framing pattern and
salt is only 64-bits.

For an iSCSI software implementation on top of an unmodified [Operating
System Vendor provided] TCP stack, I can not take advantage of out of
order
placement and also may not be able to detect TCP/IP packet boundaries.
When discarding a PDU header with a bad digest possibly followed by data
of
un-trusted length, I would like to be able to quickly recover/confirm
the
next PDU boundary using a framing marker and a digest, rather than
computing the iSCSI digest at each succeeding 4 byte boundary.  [I would
*not* like to have to generate, nor to locate and discard, framing
markers
in the middle of otherwise perfectly good PDUs.]

I have a couple of questions about the salt.  Are there any rules for
generating it?  If the rest of the ULP PDU header is included in the
digest, would this not be as effective as the salt?  When a shared
secret
or other data value not transmitted with the PDU is included in the
digest
calculation, is this as effective as adding salt, so that explicit salt
would not be needed?  If the secret and/or the ULP header can replace
the
salt, the frame markers overhead shrinks another 32-bits.

I also appreciate that the un-optimized version of the NDT framer can
easily and efficiently be inserted above or below almost any other
protocol
layer by hardware or software.  On the outbound side, it does not need
to
be told the length of the next ULP PDU.  It only needs to know when a
PDU
boundary (either beginning, ending, or both) has been reached.

Thanks,
Nick

P.S. I did not copy this message the reflector.  You are free to do what
you please with it.
 -----Original Message-----
 From: Julian Satran [mailto:julian_satran@il.ibm.com]
 Sent: Saturday, February 02, 2002 3:49 AM
 To: tsvwg@ietf.org; rdma@yahoogroups.com
 Subject: [rdma] a proposal for a different No-Data-Touch framer for TCP


 Dear Colleagues:

 I've recently submitted a draft including a different
"No-Data-Touch"(NDT)
 framer.
 Unlike the current proposed NDT framer that has a "mistake" probability
 that increases asymptotically (to 1) with the stream length the framer
I
 am suggesting has a very low and constant mistake probability and is
thus
 safe to use even in the absence of alignment or in the presence of
 re-segmentation.

 It does not contain the standard RFC2026 reference not as an indication
of
 patent - I have no  patent on it and I do not know about the existence
of
 one (nor did I search thoroughly) - but rather as a sign of discomfort
for
 it to be used without me being asked about it.

 Until it apperas you can find it also at:

 http://www.haifa.il.ibm.com/satran/ips/draft-satran-NDT-Framing-00.pdf

 http://www.haifa.il.ibm.com/satran/ips/draft-satran-NDT-Framing-00.txt

 Julo

       Yahoo! Groups Sponsor



           ADVERTISEMENT







 To unsubscribe from this group, send an email to:
 rdma-unsubscribe@egroups.com



 Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.




------------------------ Yahoo! Groups Sponsor ---------------------~-->
Tiny Wireless Camera under $80!
Order Now! FREE VCR Commander!
Click Here - Only 1 Day Left!
http://us.click.yahoo.com/nuyOHD/7.PDAA/yigFAA/W6uqlB/TM
---------------------------------------------------------------------~->

To unsubscribe from this group, send an email to:
rdma-unsubscribe@egroups.com



Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/









_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Thu Feb  7 22:36:21 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA18909
	for <tsvwg-archive@odin.ietf.org>; Thu, 7 Feb 2002 22:36:21 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id WAA03391
	for tsvwg-archive@odin.ietf.org; Thu, 7 Feb 2002 22:36:24 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA02748;
	Thu, 7 Feb 2002 22:18:15 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA02716
	for <tsvwg@ns.ietf.org>; Thu, 7 Feb 2002 22:18:13 -0500 (EST)
Received: from sandmail.sandburst.com (sandmail.sandburst.com [216.57.132.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA17826
	for <tsvwg@ietf.org>; Thu, 7 Feb 2002 22:18:10 -0500 (EST)
To: "Julian Satran" <julian_satran@il.ibm.com>
Cc: tsvwg@ietf.org, rdma@yahoogroups.com
In-Reply-To: Message from "Julian Satran" <julian_satran@il.ibm.com> 
   of "Thu, 07 Feb 2002 22:22:39 +0200." <OF8BB94B11.B145A3E1-ONC2256B59.006FE94A@telaviv.ibm.com> 
References: <OF8BB94B11.B145A3E1-ONC2256B59.006FE94A@telaviv.ibm.com> 
Date: Thu, 07 Feb 2002 22:17:05 -0500
From: Stephen Bailey <steph@cs.uchicago.edu>
Message-Id: <20020208031740.9E18F4E8E@sandmail.sandburst.com>
Subject: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Julian,

I was also initially excited by your proposal, but I've come to a
similar conclusion to Paul.

The key question is what property is probability in the framing
mechanism providing.  I claim it's for safety.

A safety analysis should assume a worst possible data stream.  But not
worst `impossible'---bits that are knowable are knowable, and bits
that are unknowable can't be used to construct a worst possible
stream).

I do not think your proposal provides better safety than the current
TUF.  In fact, it might be worse.  I agree that both are difficult to
analyze, in any case.

Bad data streams in your proposal would be something like groups of 3
words, with an incrementing first word (the supposed marker), a 0
salt and the corresponding digest.  If you got clever, I bet
you could even interleave so that the digest from the first candidate
is is the marker for the second.  Really clever would be to use the
salt of the first as the marker for the second.  I imagine that a
`good digest' would have the property that you could cover all the
marker possibilities this way in only 2^32 words.

As far as I can tell, ceasing PDU containment validation attempts long
before you've checked many false candidates is of primary importance
to maximizing safety in EITHER approach.  I don't buy your claim that
safety is not a function of the number of tests of false candidates in
your approach.  The probability that must be controlled in either
case, is the chance of actually testing a false candidate.

Steph

> Paul,
> 
> The salt is a random number that is changing on every PDU. That changes
> radically the false detection probability in face on an infinite stream
> form certainty to a value that is acceptable from engineering point of
> view. Neither protects against malicious user. However the probability of
> random good framer in the data becomes nil.
> 
> Julo

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Thu Feb  7 22:50:52 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA19148
	for <tsvwg-archive@odin.ietf.org>; Thu, 7 Feb 2002 22:50:52 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id WAA03677
	for tsvwg-archive@odin.ietf.org; Thu, 7 Feb 2002 22:50:53 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA03469;
	Thu, 7 Feb 2002 22:37:30 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA03438
	for <tsvwg@ns.ietf.org>; Thu, 7 Feb 2002 22:37:28 -0500 (EST)
Received: from david.siemens.com.sg (david.siemens.com.sg [194.138.240.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA18934
	for <tsvwg@ietf.org>; Thu, 7 Feb 2002 22:37:23 -0500 (EST)
X-Envelope-Sender-Is: ExchAdmin@sae.siemens.com.sg (at relayer david.siemens.com.sg)
Received: from mail.siemens.com.sg (mail.siemens.com.sg [194.138.237.71])
	by david.siemens.com.sg (8.10.2/8.10.2) with ESMTP id g183bOe14708
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 11:37:24 +0800 (SGT)
Received: from SGPK001A.sae.siemens.com.sg (sgpk001a.sae.siemens.com.sg [140.231.107.1])
	by mail.siemens.com.sg (8.11.4/8.11.4) with ESMTP id g183bMj23205
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 11:37:23 +0800 (SGT)
Received: by sgpk001a.sae.siemens.com.sg with Internet Mail Service (5.5.2653.19)
	id <13P6T34G>; Fri, 8 Feb 2002 11:30:22 +0800
Message-ID: <1FCAB2AB1194A040B37F3035790C634586A9FE@sgpk001a.sae.siemens.com.sg>
From: Exchange Administrator <ExchAdmin@sae.siemens.com.sg>
To: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Date: Fri, 8 Feb 2002 11:30:21 +0800 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: multipart/mixed;
	boundary="----_=_NextPart_000_01C1B050.F003FCF0"
Subject: [Tsvwg] Recovered Mails From 22 Jan 02
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_000_01C1B050.F003FCF0
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C1B050.F003FCF0"


------_=_NextPart_001_01C1B050.F003FCF0
Content-Type: text/plain;
	charset="iso-8859-1"

 <<DL9HSGAB.txt>> 

------_=_NextPart_001_01C1B050.F003FCF0
Content-Type: text/html;
	charset="iso-8859-1"

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Generator" CONTENT="MS Exchange Server version 5.5.2653.12">
<TITLE>Recovered Mails From 22 Jan 02</TITLE>
</HEAD>
<BODY>

<P><FONT FACE="Arial" SIZE=2 COLOR="#000000"> &lt;&lt;DL9HSGAB.txt&gt;&gt; </FONT>
</P>

</BODY>
</HTML>
------_=_NextPart_001_01C1B050.F003FCF0--

------_=_NextPart_000_01C1B050.F003FCF0
Content-Type: text/plain;
	name="DL9HSGAB.txt"
Content-Disposition: attachment;
	filename="DL9HSGAB.txt"


Date: Mon, 21 Jan 2002 16:07:16 +0100
From: Michela Meo <michela@polito.it>

To: tsvwg@ietf.org
CC: Sally Floyd <floyd@aciri.org>, Michela Meo <michela@polito.it>,
   casetti@polito.it, mellia@polito.it

Subject: [Tsvwg] TCP Smart Framing


Sally Floyd wrote:
>
> Michela -
>
> >To revitalize the discussion on the Smart Framing option, let us
> >summarize
> >the discussion so far:
>
> I have not followed this discussion in complete detail, but I would
> suggest that in summarizing the discussion so far, it would also
> be useful to summarize the relationship of SF with Limited Transmit
> (LT), RFC 3042, which has been Proposed Standard for a year now.

As Sally suggested in her last email, we're listing a few points
that we believe can summarize the relationship between Limited
Transmit and Smart Framing:

* Reaction time: in the small window regime (cwnd <= 4MSS), in a
single-drop scenario, it can take LT between two and three RTTs
before enough dupACKs are available to trigger FR. Conversely,
TCP-SF just needs one RTT before FR can be triggered.

* Small windows: LT does not prevent a timeout if IW=1 or when
TCP restarts with LW=1. TCP-SF always prevents timeouts in single-
loss scenarios, whatever the size of cwnd.

* RTT estimation: TCP-SF's "hidden agenda" includes the refinement
of the RTT estimate (hence possibly shorter RTO) early on in a
connection, thanks to the higher number of RTT samples. LT does not
achieve this.

* no new data: as already pointed out, not enough new data may be
available to keep the ACK clock going, i.e. because the original
file was too short; the position (inside the window)
of the dropped packet may exacerbate this problem. Such shortage
might affect TCP-SF too, though the higher number of segments makes
this
occurence less likely. Of course, if one of TCP-SF's last segments
is dropped, timeout expiration cannot be helped.

As an aside, it's probably a good idea to consider the coexistence of
LT and TCP-SF, whose joint deployment would trigger FR even in
multiple-drop scenarios.

> But instead of using smaller-than-normal packets
> to induce a Fast Retransmit, it might be better to ask the question:

> if you get a single duplicate acknowledgement, and there is no new
> data to send, should you be able to retransmit old data that might
> or might not be lost?  A "yes" answer to this question would take
> care of many of the additional cases where Limited Transmit in its
> current form is not sufficient to prevent a timeout, without
> unnecessarily reducing the packet size, and without increasing the
> number of packets used by all connections when they have small
> windows.

In our opinion, retransmitting old data is dangerously close to
getting rid of timeouts and using single dupACKs as drop
indicators. As Mark Allman himself pointed out, this should
probably be a once-per-connection solution.


> I don't buy it that four small packets presents the same load to
> the network as one larger packet.  Aside from the issue of header
> size, there can be router limitations in terms of packets per second

> processing capacity in addition to the link bandwidth limitation
> in terms of bytes per second.  It is still not clear (to me) which
> limitation, packets per second or bytes per second, is likely to
> dominate some years in the future.

Agreed, the overhead is the price you pay. However, one has to wonder
whether processing capacity is going to be the bottleneck, even in the

near future.


Michela, Claudio and Marco


------_=_NextPart_000_01C1B050.F003FCF0--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Feb  8 00:19:58 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA20468
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 00:19:58 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id AAA06772
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 00:19:57 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA06399;
	Fri, 8 Feb 2002 00:08:25 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA06370
	for <tsvwg@ns.ietf.org>; Fri, 8 Feb 2002 00:08:23 -0500 (EST)
Received: from c003.snv.cp.net (c003-h016.c003.snv.cp.net [209.228.32.230])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id AAA20317
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 00:08:22 -0500 (EST)
Received: (cpmta 12121 invoked from network); 7 Feb 2002 21:07:52 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.230) with SMTP; 7 Feb 2002 21:07:52 -0800
X-Sent: 8 Feb 2002 05:07:52 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <icsc-public@opengroup.org>, "Martin Kirk" <m.kirk@opengroup.org>
Cc: "Tsvwg" <tsvwg@ietf.org>
Date: Thu, 7 Feb 2002 21:09:47 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJOEIJCPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Importance: Normal
In-Reply-To: <5.1.0.14.2.20020208005349.03815008@maildrop.rdg.opengroup.org>
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] RE: ICSC Sockets API Extensions Charter Approved
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Kirk,

There is work establishing a DDP (Direct Data Placement) and perhaps an RDMA
layer or shim for SCTP.  This provides methods for allowing direct placement
with/without tags and would also allow emulation of TCP partial writes.
Please note this allows "TCP/IP stack bypass" while providing congestion
avoidance without introducing new IP protocols.  SCTP is used without
modification for this purpose.

Your on going effort at:
http://www.opengroup.org/icsc/ working toward an API for protocols such as:
http://www.ietf.org/rfc/rfc2327.txt and
http://www.ietf.org/rfc/rfc2326.txt within
http://www.ietf.org/html.charters/mmusic-charter.html

There are already documents addressing Socket Descriptions for SCTP at:
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpsocket-03.txt

How do you envision a collaboration with this ongoing effort?

A DDP/(RDMA)/SCTP draft will be available soon.  Presently this document is
being reviewed by Stephen Bailey [mailto:steph@cs.uchicago.edu] and Allyn
Romanow [mailto:allyn@cisco.com] of the RDMA effort and will be posted to
the IETF and announced on tsvwg:

tsvwg [mailto:tsvwg@ietf.org]
To subscribe with message: subscribe email_address
tsvwg-request [mailto:tsvwg-request@ietf.org

An RDMA reflector is at:
RDMA [mailto:rdma@yahoogroups.com]
http://groups.yahoo.com/group/RDMA/

An SCTP-DDP reflector is at:
SCTP-DDP [mailto:sctp-ddp@yahoogroups.com]
http://groups.yahoo.com/group/SCTP-DDP/

Doug


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 03:56:48 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA01403
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 03:56:48 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id DAA24660
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 03:56:50 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA24323;
	Fri, 8 Feb 2002 03:44:06 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA24295
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 03:44:04 -0500 (EST)
Received: from india.mercurykr.com ([165.133.13.60])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA01260
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 03:43:53 -0500 (EST)
Received: from vijay (Archana [165.133.13.38])
	by india.mercurykr.com (8.8.8+Sun/8.8.8) with SMTP id OAA04813
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 14:05:34 -0600 (GMT)
MIME-Version: 1.0
Message-Id: <3C638FB2.000003.89241@vijay.india.mercurykr.com>
Date: Fri, 8 Feb 2002 14:13:30 +0530 (India Standard Time)
Content-Type: Text/Plain
X-Mailer: IncrediMail 2001 (1500245)
From: "Vijay Kumar Choudhary" <vijay@mercurykr.com>
X-Priority: 3
X-FID: FLAVOR00-NONE-0000-0000-000000000000
To: <tsvwg@ietf.org>
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id DAA24296
Subject: [Tsvwg] Peer RWND = 0
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi All,
  In the RFC, it is written that if the Peer Rwnd becomes ZERO, then we can
have one DATA chunk in the flight (if allowed by cwnd).

   My ques is:
      
          If, in the retransmitting queue there are some unsent chunk and
some chunks which has to be retransmitted.

        Is it necessary to send chunk first, which are marked for
retransmission?? What will happen if we will send the unsent chunk first ???



Lastly, Is there any function to get the MTU of the local interface (in
solaris)??


Regards,
Vijay

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 04:56:43 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA02223
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 04:56:43 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id EAA27300
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 04:56:45 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA26823;
	Fri, 8 Feb 2002 04:39:01 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id EAA26742
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 04:38:57 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA02072
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 04:38:50 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir03nok.nokia.com (esvir03nokt.ntc.nokia.com [172.21.143.35])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g189cuZ02919
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 11:38:56 +0200 (EET)
Received: from esebh003.NOE.Nokia.com (unverified) by esvir03nok.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T58f1f29c7aac158f23077@esvir03nok.nokia.com>;
 Fri, 8 Feb 2002 11:38:50 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh003.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Fri, 8 Feb 2002 11:38:50 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] Peer RWND = 0
Date: Fri, 8 Feb 2002 11:38:50 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF26308EE90@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] Peer RWND = 0
Thread-Index: AcGwfOSkVCc2WI33QbG1U9Cb8cvwSAABbz2A
To: <vijay@mercurykr.com>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 08 Feb 2002 09:38:50.0778 (UTC) FILETIME=[6A3F27A0:01C1B084]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id EAA26743
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi Vijay,

> -----Original Message-----
> From: ext Vijay Kumar Choudhary [mailto:vijay@mercurykr.com]
> Sent: 08. February 2002 10:44
> To: tsvwg@ietf.org
> Subject: [Tsvwg] Peer RWND = 0
> 
> 
> Hi All,
>   In the RFC, it is written that if the Peer Rwnd becomes 
> ZERO, then we can
> have one DATA chunk in the flight (if allowed by cwnd).
> 
>    My ques is:
>       
>           If, in the retransmitting queue there are some 
> unsent chunk and
> some chunks which has to be retransmitted.
> 
>         Is it necessary to send chunk first, which are marked for
> retransmission?? What will happen if we will send the unsent 
> chunk first ???

	Don't know where exactly, but somewhere in 2960 it says that if you have some TSNs to send, you have to send first the oldest one... So in this case, the chunks marked for retransmission must go first.

> Lastly, Is there any function to get the MTU of the local 
> interface (in
> solaris)??

	I think there is one. I think that it is one of those getsockopt and the like... :-/

	BR Iván Arias Rodríguez

> Regards,
> Vijay
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 05:37:00 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA02699
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 05:37:00 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA29256
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 05:37:02 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA28155;
	Fri, 8 Feb 2002 05:09:52 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA28124
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 05:09:50 -0500 (EST)
Received: from india.mercurykr.com ([165.133.13.60])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA02419
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 05:09:23 -0500 (EST)
Received: from vijay (Archana [165.133.13.38])
	by india.mercurykr.com (8.8.8+Sun/8.8.8) with SMTP id PAA05877;
	Fri, 8 Feb 2002 15:30:13 -0600 (GMT)
MIME-Version: 1.0
Message-Id: <3C63A387.000005.89241@vijay.india.mercurykr.com>
Date: Fri, 8 Feb 2002 15:38:07 +0530 (India Standard Time)
Content-Type: Text/Plain
X-Mailer: IncrediMail 2001 (1500245)
From: "Vijay Kumar Choudhary" <vijay@mercurykr.com>
References: <58AEDE27D48F884285007ABF827FF26308EE90@esebe017.NOE.Nokia.com>
X-Priority: 3
X-FID: FLAVOR00-NONE-0000-0000-000000000000
To: <tsvwg@ietf.org>, <Ivan.Arias-Rodriguez@nokia.com>
Subject: RE: [Tsvwg] Peer RWND = 0
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id FAA28125
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi Ivan,

  Thanx for ur quick response.In RFC, the following line is written

C) When the time comes for the sender to transmit, before sending new
      DATA chunks, the sender MUST first transmit any outstanding DATA
      chunks which are marked for retransmission (limited by the current
      cwnd).

On page 68.

Does the above line mean that the chunks marked for retransmission must go
first when the peer rwnd is ZERO??? (I  think no....above line points to the
cwnd and not to the peer rwnd......may be i m wrong)

What will happen if we will send the unsent  chunk first ???




Regards,
Vijay





-------Original Message-------
From: Ivan.Arias-Rodriguez@nokia.com
Date: Saturday, February 09, 2002 02:32:07 AM
To: vijay@mercurykr.com; tsvwg@ietf.org
Subject: RE: [Tsvwg] Peer RWND Hi Vijay,
> -----Original Message-----
> From: ext Vijay Kumar Choudhary [mailto:vijay@mercurykr.com]
> Sent: 08. February 2002 10:44
> To: tsvwg@ietf.org
> Subject: [Tsvwg] Peer RWND > 
> 
> Hi All,
> In the RFC, it is written that if the Peer Rwnd becomes 
> ZERO, then we can
> have one DATA chunk in the flight (if allowed by cwnd).
> 
> My ques is:
> 
> If, in the retransmitting queue there are some 
> unsent chunk and
> some chunks which has to be retransmitted.
> 
> Is it necessary to send chunk first, which are marked for
> retransmission?? What will happen if we will send the unsent 
> chunk first ???
Don't know where exactly, but somewhere in 2960 it says that if you have
some TSNs to send, you have to send first the oldest one... So in this case,
the chunks marked for retransmission must go first.
> Lastly, Is there any function to get the MTU of the local 
> interface (in
> solaris)??
I think there is one. I think that it is one of those getsockopt and the
like... :-/
BR Iván Arias Rodríguez
> Regards,
> Vijay
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 05:37:42 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA02719
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 05:37:42 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA29273
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 05:37:45 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA28421;
	Fri, 8 Feb 2002 05:21:20 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA28393
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 05:21:18 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA02506
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 05:21:14 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir04nok.ntc.nokia.com (esvir04nokt.ntc.nokia.com [172.21.143.36])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g18ALLZ05134
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 12:21:21 +0200 (EET)
Received: from esebh001.NOE.Nokia.com (unverified) by esvir04nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T58f2197623ac158f24078@esvir04nok.ntc.nokia.com>;
 Fri, 8 Feb 2002 12:21:16 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh001.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Fri, 8 Feb 2002 12:21:15 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] Peer RWND = 0
Date: Fri, 8 Feb 2002 12:21:15 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF263257A27@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] Peer RWND = 0
Thread-Index: AcGwiNRTZ5bhL4a0SuSWGPpY0aACIQAAAqXw
To: <vijay@mercurykr.com>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 08 Feb 2002 10:21:15.0831 (UTC) FILETIME=[57376C70:01C1B08A]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id FAA28394
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

> Hi Ivan,

	Hi again Vijay,

>   Thanx for ur quick response.In RFC, the following line is written
> 
> C) When the time comes for the sender to transmit, before sending new
>       DATA chunks, the sender MUST first transmit any outstanding DATA
>       chunks which are marked for retransmission (limited by 
> the current
>       cwnd).
> 
> On page 68.
> 
> Does the above line mean that the chunks marked for 
> retransmission must go
> first when the peer rwnd is ZERO??? (I  think no....above 
> line points to the
> cwnd and not to the peer rwnd......may be i m wrong)

	It does not matter in which situation you are. When time to send something comes, you must always send first the lowest TSNs. If you only have unsent chunks, so you must send first the unsent chunk with the lowest TSN, and then the next one and so on (this basically means that you send new TSNs in order...). If you also have some chunks marked for  retransmission, then you must send first the oldest TSN marked for retransmission (and then the next one and so on) before you are able to send any unsent chunk (note that the newest chunk marked for retransmission must always be older than the unsent chunk with the lowest TSN).

	Note that the paragraph tells you what to send (supposing you can send something), not when to send.

> What will happen if we will send the unsent  chunk first ???

	You would be violating the protocol. Some bad guys could be interested in breaking some other rules so their data transmissions can go faster :-). But I don't see the point about not following this rule. If you don't do in that way, the receiver will have holes in its incoming queue and won't be able to deliver anything to the upper user... :-/

	BR Iván Arias Rodríguez

	

> 
> Regards,
> Vijay
> 
> 
> 
> 
> 
> -------Original Message-------
> From: Ivan.Arias-Rodriguez@nokia.com
> Date: Saturday, February 09, 2002 02:32:07 AM
> To: vijay@mercurykr.com; tsvwg@ietf.org
> Subject: RE: [Tsvwg] Peer RWND Hi Vijay,
> > -----Original Message-----
> > From: ext Vijay Kumar Choudhary [mailto:vijay@mercurykr.com]
> > Sent: 08. February 2002 10:44
> > To: tsvwg@ietf.org
> > Subject: [Tsvwg] Peer RWND > 
> > 
> > Hi All,
> > In the RFC, it is written that if the Peer Rwnd becomes 
> > ZERO, then we can
> > have one DATA chunk in the flight (if allowed by cwnd).
> > 
> > My ques is:
> > 
> > If, in the retransmitting queue there are some 
> > unsent chunk and
> > some chunks which has to be retransmitted.
> > 
> > Is it necessary to send chunk first, which are marked for
> > retransmission?? What will happen if we will send the unsent 
> > chunk first ???
> Don't know where exactly, but somewhere in 2960 it says that 
> if you have
> some TSNs to send, you have to send first the oldest one... 
> So in this case,
> the chunks marked for retransmission must go first.
> > Lastly, Is there any function to get the MTU of the local 
> > interface (in
> > solaris)??
> I think there is one. I think that it is one of those 
> getsockopt and the
> like... :-/
> BR Iván Arias Rodríguez
> > Regards,
> > Vijay
> > 
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> > 
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 05:38:26 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA02735
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 05:38:26 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA29306
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 05:38:28 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA28355;
	Fri, 8 Feb 2002 05:20:39 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA28330
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 05:20:37 -0500 (EST)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA02502
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 05:20:33 -0500 (EST)
Received: from netlab.hcltech.com (arun [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id QAA25152;
	Fri, 8 Feb 2002 16:11:36 +0530
Message-ID: <3C63A676.BE215983@netlab.hcltech.com>
Date: Fri, 08 Feb 2002 15:50:39 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Vijay Kumar Choudhary <vijay@mercurykr.com>
CC: tsvwg@ietf.org, Ivan.Arias-Rodriguez@nokia.com
Subject: Re: [Tsvwg] Peer RWND = 0
References: <58AEDE27D48F884285007ABF827FF26308EE90@esebe017.NOE.Nokia.com> <3C63A387.000005.89241@vijay.india.mercurykr.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit



Vijay Kumar Choudhary wrote:

> Hi Ivan,
>
>   Thanx for ur quick response.In RFC, the following line is written
>
> C) When the time comes for the sender to transmit, before sending new
>       DATA chunks, the sender MUST first transmit any outstanding DATA
>       chunks which are marked for retransmission (limited by the current
>       cwnd).
>
> On page 68.
>
> Does the above line mean that the chunks marked for retransmission must go
> first when the peer rwnd is ZERO??? (I  think no....above line points to the
> cwnd and not to the peer rwnd......may be i m wrong)
>

    I think ur guess inside the brackets is right...

>
> What will happen if we will send the unsent  chunk first ???
>

        It will create unnecessary gaps in the Receiver side... No point in sending
new data chunks to the peer when the old ones havent reached there.... (because
the ones u are sending also will not reach the Application, as it is out of sequence...)
Am I answering ur Question????


>
> Regards,
> Vijay
>

Thanks
-arun

>
> -------Original Message-------
> From: Ivan.Arias-Rodriguez@nokia.com
> Date: Saturday, February 09, 2002 02:32:07 AM
> To: vijay@mercurykr.com; tsvwg@ietf.org
> Subject: RE: [Tsvwg] Peer RWND Hi Vijay,
> > -----Original Message-----
> > From: ext Vijay Kumar Choudhary [mailto:vijay@mercurykr.com]
> > Sent: 08. February 2002 10:44
> > To: tsvwg@ietf.org
> > Subject: [Tsvwg] Peer RWND >
> >
> > Hi All,
> > In the RFC, it is written that if the Peer Rwnd becomes
> > ZERO, then we can
> > have one DATA chunk in the flight (if allowed by cwnd).
> >
> > My ques is:
> >
> > If, in the retransmitting queue there are some
> > unsent chunk and
> > some chunks which has to be retransmitted.
> >
> > Is it necessary to send chunk first, which are marked for
> > retransmission?? What will happen if we will send the unsent
> > chunk first ???
> Don't know where exactly, but somewhere in 2960 it says that if you have
> some TSNs to send, you have to send first the oldest one... So in this case,
> the chunks marked for retransmission must go first.
> > Lastly, Is there any function to get the MTU of the local
> > interface (in
> > solaris)??
> I think there is one. I think that it is one of those getsockopt and the
> like... :-/
> BR Iván Arias Rodríguez
> > Regards,
> > Vijay
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> >
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 05:47:55 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA03000
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 05:47:55 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA29525
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 05:47:57 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA28568;
	Fri, 8 Feb 2002 05:23:34 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA28514
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 05:23:30 -0500 (EST)
Received: from netcore.fi (netcore.fi [193.94.160.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA02516;
	Fri, 8 Feb 2002 05:23:25 -0500 (EST)
Received: from localhost (pekkas@localhost)
	by netcore.fi (8.11.6/8.11.6) with ESMTP id g18AMm706105;
	Fri, 8 Feb 2002 12:22:48 +0200
Date: Fri, 8 Feb 2002 12:22:47 +0200 (EET)
From: Pekka Savola <pekkas@netcore.fi>
To: Coene Lode <Lode.Coene@siemens.atea.be>
cc: "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)" <sigtran@ietf.org>,
        "'multi6@ops.ietf.org'" <multi6@ops.ietf.org>
In-Reply-To: <E76F715C0429D5118F2100508BB9EDEE0795AC@hrtades7.atea.be>
Message-ID: <Pine.LNX.4.44.0202081211000.4989-100000@netcore.fi>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Subject: [Tsvwg] Re: SCTP multihoming issues draft
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

On Wed, 6 Feb 2002, Coene Lode wrote:
> Previous week, the draft on SCTP multihoming issues was published.
> 
> The draft can be found at the IETF draft directory: 
> Multihoming issues in the Stream Control Transmission Protocol
> <draft-coene-sctp-multihome-02.txt>
> 
> If by then end of next week, no further issues have been identified, then I
> would ask for this draft to be moved to informational.

This is not a multi6 issue as such, as SCTP does not really provide site 
multihoming (however, SCTP may enable site multihoming with multiple 
addresses per node, so this may be in the scope of wg even though it most 
probably doesn't meet the requirements in the requirements document).

But anyway, I read the draft on a flight.  A few comments:

(first a few editorial nits)

Abstract

    This document describes issues of the Stream Control Transmission
    Protocol (SCTP)[RFC2960] in regard to multihoming on the
    Internet. It explores cases where through situations in the
    internet, single points-of-failure can occur even when using
    multihoming and what the impact is of multihoming on the host
    routing tables.

==> what are "host routing tables" ?

1 Introduction


    SCTP[RFC2960] is a transport protocol that uses multihoming. This   
    draft is a attempt at identifying the issues that may arise in the  
    layers below SCTP when multihoming is used by SCTP. Some issues are
    already being addresses in various other WG and this document will
    try to highlight them. If the solutions would resolve the issues
    presented in this document then the problem presented is no longer
    an issue. This document will also try to gauge the effectiveness of
    the present multihoming architecture.

==> s/a attempt/an attempt/, s/WG/WGs/

2.1 Interaction with routing

    Under the assumption that every IP address will have a different,
    seperate paths towards the remote endpoint, (this is the
    responsibility of the routing protocols or of manual configuration)

==> s/paths/path/, s/endpoint,/endpoint/

    It might be usefull to explore ways where no routing tables are

==> s/usefull/useful/

==> on a generic note: it seems to me there is a need for routing tables 
on hosts only if one wants to optimize the paths: one is used by default, 
and if it fails, just (try to) switch to the next one, right?

2.3 SCTP multihoming and Network Adress Translators(NAT)

==> this very same problem will come up with single-homed NAT's too (need
to be able to do ALG translation), right?  Can be deleted.

2.4 SCTP multihoming and IPsec

 A much better
    implementation approach is to simply use groups of addresses instead
    of single ones.

==> note: all SCTP+IPSEC implementations would have to do this to be
compatible, it appears.

3 Security considerations

    As such the use of multihoming does not provide security risks. The
    solutions needed for allowing multihoming may provide security
    risks.

==> this seems like a circular argument: SCTP is such a solution.

==> consider the security of introducing new addresses to the association 
and capturing traffic: granted, this is an SCTP concern but is highlighted 
by multihoming.  This cannot be handwaved by IPSEC unless IPSEC is 
required for all SCTP.

-- 
Pekka Savola                 "Tell me of difficulties surmounted,
Netcore Oy                   not those you stumble over and fall"
Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 06:24:03 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA03583
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 06:24:03 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA01053
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 06:24:04 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA29758;
	Fri, 8 Feb 2002 05:57:44 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA29724
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 05:57:42 -0500 (EST)
Received: from netcore.fi (netcore.fi [193.94.160.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA03273
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 05:57:38 -0500 (EST)
Received: from localhost (pekkas@localhost)
	by netcore.fi (8.11.6/8.11.6) with ESMTP id g18Av9Z06356;
	Fri, 8 Feb 2002 12:57:09 +0200
Date: Fri, 8 Feb 2002 12:57:09 +0200 (EET)
From: Pekka Savola <pekkas@netcore.fi>
To: lln@cdt.luth.se
cc: tsvwg@ietf.org
In-Reply-To: <200201281302.IAA23637@ietf.org>
Message-ID: <Pine.LNX.4.44.0202081247460.4989-100000@netcore.fi>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Subject: [Tsvwg] Re: I-D ACTION:draft-ietf-tsvwg-udp-lite-00.txt
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Not on the list, so please keep Cc:.

On Mon, 28 Jan 2002 Internet-Drafts@ietf.org wrote:
> 	Title		: The UDP Lite Protocol
> 	Author(s)	: L. Larzon, M. Degermark
> 	Filename	: draft-ietf-tsvwg-udp-lite-00.txt
> 	Pages		: 8
> 	Date		: 25-Jan-02

Just a note: there is no "total length" field in IPv6 header (consider
e.g. a case where there is an extension header before UDP-lite), but it
this is discussed in RFC2460, and hopefully this is enough.  The draft
seems a bit "IPv4" -centric though.

A few typos/etc.:

   insensitive part.  For a link layer that supports partial error
   detection, the Coverage field in the UDP Lite header MAY be used as a
   
==> Checksum Coverage

   use. Most strong encryption transforms today exhibit this behaviour, 
   for good reason.  It might be possible to develop encryption

==> 'for a good reason' ?

-- 
Pekka Savola                 "Tell me of difficulties surmounted,
Netcore Oy                   not those you stumble over and fall"
Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 06:34:53 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA03700
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 06:34:53 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA01638
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 06:34:54 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA00527;
	Fri, 8 Feb 2002 06:11:53 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA00498
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 06:11:51 -0500 (EST)
Received: from wiproecmx1.wipro.com (wiproecmx1.wipro.com [164.164.31.5])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA03391
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 06:11:44 -0500 (EST)
Received: from ecvwall1.wipro.com (ecvwall1.wipro.com [164.164.23.6])
	by wiproecmx1.wipro.com (8.11.3/8.11.3) with SMTP id g18BA9X20518
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 16:40:10 +0530 (IST)
Received: from wipro.com ([192.168.178.83]) by
          ecmail.mail.wipro.com (Netscape Messaging Server 4.15) with
          ESMTP id GR7OEI00.BAY; Fri, 8 Feb 2002 16:41:06 +0530 
Message-ID: <3C63B2C8.5A78A65B@wipro.com>
Date: Fri, 08 Feb 2002 16:43:12 +0530
From: Ashraf <mohammed.asharaf@wipro.com>
Organization: Wipro
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.4.2-2 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Vijay Kumar Choudhary <vijay@mercurykr.com>
CC: tsvwg@ietf.org, Ivan.Arias-Rodriguez@nokia.com
Subject: Re: [Tsvwg] Peer RWND = 0
References: <58AEDE27D48F884285007ABF827FF26308EE90@esebe017.NOE.Nokia.com> <3C63A387.000005.89241@vijay.india.mercurykr.com>
Content-Type: multipart/mixed;
	boundary="----=_NextPartTM-000-738024dd-1c81-11d6-a942-00b0d0d06be8"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


This is a multi-part message in MIME format.

------=_NextPartTM-000-738024dd-1c81-11d6-a942-00b0d0d06be8
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit

hi,
I would like to know when you will be in such situation?( some bytes marked for
retransmission as well as RWND = 0)
bcoz, when your timer times out, and when you mark chunk for retransmission,
RWND opens up ( if its zero ) becasue you will  add the bytes marked for
retransmission to the RWND

Thnx,
ashraf
Vijay Kumar Choudhary wrote:

> Hi Ivan,
>
>   Thanx for ur quick response.In RFC, the following line is written
>
> C) When the time comes for the sender to transmit, before sending new
>       DATA chunks, the sender MUST first transmit any outstanding DATA
>       chunks which are marked for retransmission (limited by the current
>       cwnd).
>
> On page 68.
>
> Does the above line mean that the chunks marked for retransmission must go
> first when the peer rwnd is ZERO??? (I  think no....above line points to the
> cwnd and not to the peer rwnd......may be i m wrong)
>
> What will happen if we will send the unsent  chunk first ???
>
> Regards,
> Vijay
>
> -------Original Message-------
> From: Ivan.Arias-Rodriguez@nokia.com
> Date: Saturday, February 09, 2002 02:32:07 AM
> To: vijay@mercurykr.com; tsvwg@ietf.org
> Subject: RE: [Tsvwg] Peer RWND Hi Vijay,
> > -----Original Message-----
> > From: ext Vijay Kumar Choudhary [mailto:vijay@mercurykr.com]
> > Sent: 08. February 2002 10:44
> > To: tsvwg@ietf.org
> > Subject: [Tsvwg] Peer RWND >
> >
> > Hi All,
> > In the RFC, it is written that if the Peer Rwnd becomes
> > ZERO, then we can
> > have one DATA chunk in the flight (if allowed by cwnd).
> >
> > My ques is:
> >
> > If, in the retransmitting queue there are some
> > unsent chunk and
> > some chunks which has to be retransmitted.
> >
> > Is it necessary to send chunk first, which are marked for
> > retransmission?? What will happen if we will send the unsent
> > chunk first ???
> Don't know where exactly, but somewhere in 2960 it says that if you have
> some TSNs to send, you have to send first the oldest one... So in this case,
> the chunks marked for retransmission must go first.
> > Lastly, Is there any function to get the MTU of the local
> > interface (in
> > solaris)??
> I think there is one. I think that it is one of those getsockopt and the
> like... :-/
> BR Iván Arias Rodríguez
> > Regards,
> > Vijay
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> >
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg


------=_NextPartTM-000-738024dd-1c81-11d6-a942-00b0d0d06be8
Content-Type: text/plain;
	name="Wipro_Disclaimer.txt"
Content-Disposition: attachment;
	filename="Wipro_Disclaimer.txt"
Content-Transfer-Encoding: 7bit

**************************Disclaimer************************************
      


Information contained in this E-MAIL being proprietary to Wipro Limited
is 'privileged' and 'confidential' and intended for use only by the
individual or entity to which it is addressed. You are notified that any
use, copying or dissemination of the information contained in the E-MAIL
in any manner whatsoever is strictly prohibited.



 ********************************************************************

------=_NextPartTM-000-738024dd-1c81-11d6-a942-00b0d0d06be8--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 07:12:51 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA04209
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 07:12:51 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA03159
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 07:12:52 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA01997;
	Fri, 8 Feb 2002 06:47:26 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA01969
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 06:47:24 -0500 (EST)
Received: from irmgard.exp-math.uni-essen.de (irmgard.exp-math.uni-essen.de [132.252.150.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA03848
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 06:47:22 -0500 (EST)
Received: from there (biene.exp-math.uni-essen.de [132.252.150.215])
	by irmgard.exp-math.uni-essen.de (8.12.2/8.12.2) with SMTP id g18BlHj9034602;
	Fri, 8 Feb 2002 12:47:18 +0100
Message-Id: <200202081147.g18BlHj9034602@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain;
  charset="iso-8859-1"
From: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
Organization: University Essen
To: Tim Chown <tjc@ecs.soton.ac.uk>, Coene Lode <Lode.Coene@siemens.atea.be>
Subject: Re: [Tsvwg] Re: SCTP multihoming issues draft
Date: Fri, 8 Feb 2002 12:46:56 +0100
X-Mailer: KMail [version 1.3.1]
Cc: "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)" <sigtran@ietf.org>
References: <Pine.LNX.4.21.0202061413560.7370-100000@login.ecs.soton.ac.uk>
In-Reply-To: <Pine.LNX.4.21.0202061413560.7370-100000@login.ecs.soton.ac.uk>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi, Tim, and others,

please see the comments below...

On Wednesday, 6. February 2002 15:30, Tim Chown wrote:
> On Wed, 6 Feb 2002, Coene Lode wrote:
> > If by then end of next week, no further issues have been identified, then
> > I would ask for this draft to be moved to informational.
>
> Does the draft have to only consider two network interfaces for the
> multihoming issues?   

No, in general it refers to one or more network layer addresses.
It tries to address issues that may come up with redundancy, where 
there are asymmetric path combinations (e.g. there is only one interface 
on one side, the (direct) link of which fails -- i.e. if you have 
several addresses configured on that interface, all these will fail 
for SCTP, if the link fails etc.).

> In IPv6, as you point out, each interface can have
> more than one IPv6 address (and addresses of many scopes), so the simplest
> multihoming case for IPv6 is a host (endpoint) with one interface,
> recieving two different /64 prefix RAs, and thus having two IPv6 addresses
> from two different providers.   The dual interface scenario may be useful
> of course, e.g. when considering wired and air interfaces.
>
> When you say "it is recommended that IP addresses in a multihomed endpoint
> be assigned IP endpoints from different TLA's to ensure against network
> failure", this would be the default case for IPv6.
>
> For IPv6, delete section 2.3 :-)

Yes, I agree that this is not something new for IPv6, but that comment
was meant also for IPv4, where that is not generally true. So the section
should stay there, but maybe something can be added to clarify this.
What would be your suggestion ?

> In general, more consideration could be given for IPv6, perhaps?  

Maybe you could be a bit more descriptive. While we sure have great
consideration for IPv6, what do you feel is missing ?

Thank you for the comments, anyway,

regards,

Andreas

- -- 
- ------------------------------------------------------------------------
Dipl.-Ing. Andreas Jungmaier              |                         _
Computer Networking Technology Group      | ASCII ribbon campaign  ( )
University of Essen                       |  - against HTML email   X
http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8Y7qxXAsLBNOCSsARAi3MAJ9U1SRr1wxLBIQLR0iePqWfe1oCQwCfZE4N
SrhqBBluzPONEoYH5nw+pMI=
=4ZJF
-----END PGP SIGNATURE-----

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 07:33:14 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA04361
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 07:33:14 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA03914
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 07:33:16 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA03229;
	Fri, 8 Feb 2002 07:14:17 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA03177
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 07:14:15 -0500 (EST)
Received: from irmgard.exp-math.uni-essen.de (irmgard.exp-math.uni-essen.de [132.252.150.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA04247;
	Fri, 8 Feb 2002 07:14:12 -0500 (EST)
Received: from there (biene.exp-math.uni-essen.de [132.252.150.215])
	by irmgard.exp-math.uni-essen.de (8.12.2/8.12.2) with SMTP id g18CECj9028264;
	Fri, 8 Feb 2002 13:14:12 +0100
Message-Id: <200202081214.g18CECj9028264@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain;
  charset="iso-8859-1"
From: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
Organization: University Essen
To: Pekka Savola <pekkas@netcore.fi>, Coene Lode <Lode.Coene@siemens.atea.be>
Subject: Re: [Tsvwg] Re: SCTP multihoming issues draft
Date: Fri, 8 Feb 2002 13:13:51 +0100
X-Mailer: KMail [version 1.3.1]
Cc: "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)" <sigtran@ietf.org>
References: <Pine.LNX.4.44.0202081211000.4989-100000@netcore.fi>
In-Reply-To: <Pine.LNX.4.44.0202081211000.4989-100000@netcore.fi>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Pekka, hello all,

please see my comments below:

On Friday, 8. February 2002 11:22, Pekka Savola wrote:
> On Wed, 6 Feb 2002, Coene Lode wrote:
> > Previous week, the draft on SCTP multihoming issues was published.
> > The draft can be found at the IETF draft directory:
> > Multihoming issues in the Stream Control Transmission Protocol
> > <draft-coene-sctp-multihome-02.txt>
> > If by then end of next week, no further issues have been identified, then
> > I would ask for this draft to be moved to informational.
>
> Abstract
>     This document describes issues of the Stream Control Transmission
>     Protocol (SCTP)[RFC2960] in regard to multihoming on the
>     Internet. It explores cases where through situations in the
>     internet, single points-of-failure can occur even when using
>     multihoming and what the impact is of multihoming on the host
>     routing tables.
>
> ==> what are "host routing tables" ?

I suppose that is what you get when you type in "netstat -rn" on a decent
unix machine. Anyway, this is probably confusing: it should probably say
something like "...and how the SCTP multihoming interacts with routing"
since that is what is of concern here: the interface between a transport
protocol and the network layer.

>
>     It might be usefull to explore ways where no routing tables are
>
> ==> on a generic note: it seems to me there is a need for routing tables
> on hosts only if one wants to optimize the paths: one is used by default,
> and if it fails, just (try to) switch to the next one, right?

Any host that has an IP address has some routing tables. Richard Stevens 
has put that nicely in his "TCP/IP Illustrated, Vol.1 - The Protocols", 
chapter 9. That's recommended reading :-)

> 2.3 SCTP multihoming and Network Adress Translators(NAT)
>
> ==> this very same problem will come up with single-homed NAT's too (need
> to be able to do ALG translation), right?  Can be deleted.

What exactly can be deleted ?
Well, actually that section probably has not taken consideration at 
all of single-homed NATs (which in turn is not something that comes to 
mind at once when talking about a NAT). 
This is needed because single-homed _SCTP_ works together with NATs nicely 
(because no multiple IP addresses are carried inside the SCTP packets in the 
association setup phase), but multi-homed SCTP carries these IP addresses 
inside the INIT and INIT ACK chunks, so something needs to be tweaked in
that case.

> 2.4 SCTP multihoming and IPsec
>
>  A much better
>     implementation approach is to simply use groups of addresses instead
>     of single ones.
>
> ==> note: all SCTP+IPSEC implementations would have to do this to be
> compatible, it appears.

Yep. Where I am not sure as to which IPSec implementations support this
(if any). This note should be added, just to be sure no one gets confused.

> 3 Security considerations
>
>     As such the use of multihoming does not provide security risks. The
>     solutions needed for allowing multihoming may provide security
>     risks.
>
> ==> this seems like a circular argument: SCTP is such a solution.

And I agree there, too. That is a true "non-sentence" :-)

> ==> consider the security of introducing new addresses to the association
> and capturing traffic: granted, this is an SCTP concern but is highlighted

No it is not, because SCTP according to all RFC 2960 does not
allow for adding addresses. 

Best regards, and thanks for the comments,

Andreas
- -- 
- ------------------------------------------------------------------------
Dipl.-Ing. Andreas Jungmaier              |                         _
Computer Networking Technology Group      | ASCII ribbon campaign  ( )
University of Essen                       |  - against HTML email   X
http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8Y8D/XAsLBNOCSsARAshrAJ9LuEmq7ElmsqlWtRRl6aHnU9C20gCfQEwf
or9VkpvsEQfjz3agT/rGkdw=
=OFgv
-----END PGP SIGNATURE-----

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 08:31:36 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA05896
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 08:31:36 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA07349
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 08:31:38 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA06526;
	Fri, 8 Feb 2002 08:19:16 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA06496
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 08:19:13 -0500 (EST)
Received: from mgw-x3.nokia.com (mgw-x3.nokia.com [131.228.20.26])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA05546
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 08:19:10 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir01nok.ntc.nokia.com (esvir01nokt.ntc.nokia.com [172.21.143.33])
	by mgw-x3.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g18DJVi24951
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 15:19:31 +0200 (EET)
Received: from esebh001.NOE.Nokia.com (unverified) by esvir01nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T58f2bc546fac158f21081@esvir01nok.ntc.nokia.com>;
 Fri, 8 Feb 2002 15:19:10 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh001.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Fri, 8 Feb 2002 15:19:10 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] Peer RWND = 0
Date: Fri, 8 Feb 2002 15:19:10 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF263257A2E@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] Peer RWND = 0
Thread-Index: AcGwkVrV2vsxfrJYSYmw+HNKiGi53gAEWm2A
To: <mohammed.asharaf@wipro.com>, <vijay@mercurykr.com>
Cc: <tsvwg@ietf.org>
X-OriginalArrivalTime: 08 Feb 2002 13:19:10.0547 (UTC) FILETIME=[31D7D230:01C1B0A3]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id IAA06497
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

> hi,
> I would like to know when you will be in such situation?( 
> some bytes marked for
> retransmission as well as RWND = 0)
> bcoz, when your timer times out, and when you mark chunk for 
> retransmission,
> RWND opens up ( if its zero ) becasue you will  add the bytes 
> marked for
> retransmission to the RWND

	Yes, it is true that it shouldn't be the normal case at all. But, if the peer is shrinking its window, you can end up having a null rwnd even though you don't have anything in flight. This is why in RFC 2960 it says that you are always allowed to have at least one TSN in flight no matter the value of rwnd. If you wouldn't do so you could never know about an enlargement of the rwnd...

	BR Iván Arias Rodríguez

> Thnx,
> ashraf
> Vijay Kumar Choudhary wrote:
> 
> > Hi Ivan,
> >
> >   Thanx for ur quick response.In RFC, the following line is written
> >
> > C) When the time comes for the sender to transmit, before 
> sending new
> >       DATA chunks, the sender MUST first transmit any 
> outstanding DATA
> >       chunks which are marked for retransmission (limited 
> by the current
> >       cwnd).
> >
> > On page 68.
> >
> > Does the above line mean that the chunks marked for 
> retransmission must go
> > first when the peer rwnd is ZERO??? (I  think no....above 
> line points to the
> > cwnd and not to the peer rwnd......may be i m wrong)
> >
> > What will happen if we will send the unsent  chunk first ???
> >
> > Regards,
> > Vijay
> >
> > -------Original Message-------
> > From: Ivan.Arias-Rodriguez@nokia.com
> > Date: Saturday, February 09, 2002 02:32:07 AM
> > To: vijay@mercurykr.com; tsvwg@ietf.org
> > Subject: RE: [Tsvwg] Peer RWND Hi Vijay,
> > > -----Original Message-----
> > > From: ext Vijay Kumar Choudhary [mailto:vijay@mercurykr.com]
> > > Sent: 08. February 2002 10:44
> > > To: tsvwg@ietf.org
> > > Subject: [Tsvwg] Peer RWND >
> > >
> > > Hi All,
> > > In the RFC, it is written that if the Peer Rwnd becomes
> > > ZERO, then we can
> > > have one DATA chunk in the flight (if allowed by cwnd).
> > >
> > > My ques is:
> > >
> > > If, in the retransmitting queue there are some
> > > unsent chunk and
> > > some chunks which has to be retransmitted.
> > >
> > > Is it necessary to send chunk first, which are marked for
> > > retransmission?? What will happen if we will send the unsent
> > > chunk first ???
> > Don't know where exactly, but somewhere in 2960 it says 
> that if you have
> > some TSNs to send, you have to send first the oldest one... 
> So in this case,
> > the chunks marked for retransmission must go first.
> > > Lastly, Is there any function to get the MTU of the local
> > > interface (in
> > > solaris)??
> > I think there is one. I think that it is one of those 
> getsockopt and the
> > like... :-/
> > BR Iván Arias Rodríguez
> > > Regards,
> > > Vijay
> > >
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > >
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 08:37:37 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA06074
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 08:37:37 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA07541
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 08:37:40 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA06332;
	Fri, 8 Feb 2002 08:14:35 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA06220
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 08:14:30 -0500 (EST)
Received: from hrtades9.atea.be (viap111.atea.be [194.78.143.111])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA05459;
	Fri, 8 Feb 2002 08:14:26 -0500 (EST)
Received: from hrtades10.atea.be (siemens.atea.be [139.10.143.141]) by hrtades9.atea.be with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13)
	id DF4YR84T; Fri, 8 Feb 2002 14:13:45 +0100
Received: by siemens.atea.be with Internet Mail Service (5.5.2653.19)
	id <D6SNGPJJ>; Fri, 8 Feb 2002 14:14:09 +0100
Message-ID: <E76F715C0429D5118F2100508BB9EDEE0795B1@hrtades7.atea.be>
From: Coene Lode <Lode.Coene@siemens.atea.be>
To: "'Pekka Savola'" <pekkas@netcore.fi>,
        Coene Lode
	 <Lode.Coene@siemens.atea.be>,
        "Andreas Jungmayer (E-mail)"
	 <ajung@exp-math.uni-essen.de>
Cc: "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)"
	 <sigtran@ietf.org>,
        "'multi6@ops.ietf.org'" <multi6@ops.ietf.org>
Date: Fri, 8 Feb 2002 14:11:34 +0100 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: [Tsvwg] RE: SCTP multihoming issues draft
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

I'll reply to all mailing lists (..crosssposting be dammed...)

> -----Original Message-----
> From: Pekka Savola [mailto:pekkas@netcore.fi]
> Sent: vrijdag 8 februari 2002 11:23
> To: Coene Lode
> Cc: Transport Area wg (E-mail); SIGTRAN (E-mail); 
> 'multi6@ops.ietf.org'
> Subject: Re: SCTP multihoming issues draft
> 
> 
> On Wed, 6 Feb 2002, Coene Lode wrote:
> > Previous week, the draft on SCTP multihoming issues was published.
> > 
> > The draft can be found at the IETF draft directory: 
> > Multihoming issues in the Stream Control Transmission Protocol
> > <draft-coene-sctp-multihome-02.txt>
> > 
> > If by then end of next week, no further issues have been 
> identified, then I
> > would ask for this draft to be moved to informational.
> 
> This is not a multi6 issue as such, as SCTP does not really 
> provide site 
> multihoming (however, SCTP may enable site multihoming with multiple 
> addresses per node, so this may be in the scope of wg even 
> though it most 
> probably doesn't meet the requirements in the requirements document).

Not sure that we are part of the solution space. Am waiting for a release of
the requirements document to see if any of the requirements make live hard
for SCTP multihoming.

> 
> But anyway, I read the draft on a flight.  A few comments:
> 
> (first a few editorial nits)
> 
> Abstract
> 
>     This document describes issues of the Stream Control Transmission
>     Protocol (SCTP)[RFC2960] in regard to multihoming on the
>     Internet. It explores cases where through situations in the
>     internet, single points-of-failure can occur even when using
>     multihoming and what the impact is of multihoming on the host
>     routing tables.
> 
> ==> what are "host routing tables" ?
selecting the correct outgoing link is a routing descision. See also Andreas
comments (with which I agree).
> 
> 1 Introduction
> 
> 
>     SCTP[RFC2960] is a transport protocol that uses 
> multihoming. This   
>     draft is a attempt at identifying the issues that may 
> arise in the  
>     layers below SCTP when multihoming is used by SCTP. Some 
> issues are
>     already being addresses in various other WG and this document will
>     try to highlight them. If the solutions would resolve the issues
>     presented in this document then the problem presented is no longer
>     an issue. This document will also try to gauge the 
> effectiveness of
>     the present multihoming architecture.
> 
> ==> s/a attempt/an attempt/, s/WG/WGs/

Fixed

> 
> 2.1 Interaction with routing
> 
>     Under the assumption that every IP address will have a different,
>     seperate paths towards the remote endpoint, (this is the
>     responsibility of the routing protocols or of manual 
> configuration)
> 
> ==> s/paths/path/, s/endpoint,/endpoint/

fixed

> 
>     It might be usefull to explore ways where no routing tables are
> 
> ==> s/usefull/useful/

fixed
> 
> ==> on a generic note: it seems to me there is a need for 
> routing tables 
> on hosts only if one wants to optimize the paths: one is used 
> by default, 
> and if it fails, just (try to) switch to the next one, right?
> 

See previous remark on routing tables.

> 2.3 SCTP multihoming and Network Adress Translators(NAT)
> 
> ==> this very same problem will come up with single-homed 
> NAT's too (need
> to be able to do ALG translation), right?  Can be deleted.

The single home case is rather straight forward(and works nicely together
with NAT's as Andreas points out), but it is a multihomed SCTP association
which will make live (very) difficult for NAT boxes, esspecially if the NAT
box doesn't want to be the single point of failure. So this paragraph is
rigth to the point.

What this paragraph says(between the lines) is: do not use NAT boxes if you
ever want true multihoming.

> 
> 2.4 SCTP multihoming and IPsec
> 
>  A much better
>     implementation approach is to simply use groups of 
> addresses instead
>     of single ones.
> 
> ==> note: all SCTP+IPSEC implementations would have to do this to be
> compatible, it appears.

We'll add a note(above is a good line).

> 
> 3 Security considerations
> 
>     As such the use of multihoming does not provide security 
> risks. The
>     solutions needed for allowing multihoming may provide security
>     risks.
> 
> ==> this seems like a circular argument: SCTP is such a solution.

I'll remove the first sentence. Then it isn't a circular argument.
The second sentence would then say that SCTP migth provide security risks
when using multihoming(but we don't know them yet)

> 
> ==> consider the security of introducing new addresses to the 
> association 
> and capturing traffic: granted, this is an SCTP concern but 
> is highlighted 
> by multihoming.  This cannot be handwaved by IPSEC unless IPSEC is 
> required for all SCTP.

adding(or removing) addresses to a association is not part of the baseline
SCTP spec. That is a addition in the works (see
http://search.ietf.org/internet-drafts/draft-ietf-tsvwg-addip-sctp-04.txt
This document only talks about the present SCTP capabilities.

> 
> -- 
> Pekka Savola                 "Tell me of difficulties surmounted,
> Netcore Oy                   not those you stumble over and fall"
> Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords
>

Thanks for the remarks. They have been helpfull.

yours sincerly,
Lode Coene 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 10:12:39 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA10231
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 10:12:39 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA12422
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 10:12:42 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA10848;
	Fri, 8 Feb 2002 09:47:27 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA10797
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 09:47:25 -0500 (EST)
Received: from netcore.fi (netcore.fi [193.94.160.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA08685;
	Fri, 8 Feb 2002 09:47:19 -0500 (EST)
Received: from localhost (pekkas@localhost)
	by netcore.fi (8.11.6/8.11.6) with ESMTP id g18EkeC08110;
	Fri, 8 Feb 2002 16:46:40 +0200
Date: Fri, 8 Feb 2002 16:46:40 +0200 (EET)
From: Pekka Savola <pekkas@netcore.fi>
To: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
cc: Coene Lode <Lode.Coene@siemens.atea.be>,
        "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)" <sigtran@ietf.org>
Subject: Re: [Tsvwg] Re: SCTP multihoming issues draft
In-Reply-To: <200202081214.g18CECj9028264@irmgard.exp-math.uni-essen.de>
Message-ID: <Pine.LNX.4.44.0202081632300.7919-100000@netcore.fi>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

On Fri, 8 Feb 2002, Andreas Jungmaier wrote:
> > Abstract
> >     This document describes issues of the Stream Control Transmission
> >     Protocol (SCTP)[RFC2960] in regard to multihoming on the
> >     Internet. It explores cases where through situations in the
> >     internet, single points-of-failure can occur even when using
> >     multihoming and what the impact is of multihoming on the host
> >     routing tables.
> >
> > ==> what are "host routing tables" ?
> 
> I suppose that is what you get when you type in "netstat -rn" on a decent
> unix machine. Anyway, this is probably confusing: it should probably say
> something like "...and how the SCTP multihoming interacts with routing"
> since that is what is of concern here: the interface between a transport
> protocol and the network layer.

That seems more sane, or 'routing tables on hosts'.

> >     It might be usefull to explore ways where no routing tables are
> >
> > ==> on a generic note: it seems to me there is a need for routing tables
> > on hosts only if one wants to optimize the paths: one is used by default,
> > and if it fails, just (try to) switch to the next one, right?
> 
> Any host that has an IP address has some routing tables. Richard Stevens 
> has put that nicely in his "TCP/IP Illustrated, Vol.1 - The Protocols", 
> chapter 9. That's recommended reading :-)

I know my Stevens very well, thank you.

If you read the draft around that place, you'll notice the question was 
about SCTP hosts somehow having to dump routing tables _from a router_ on 
hosts to make sensible decisions.  My belief is that "regular" routing 
table (ie. basically about 4 routes) is enough.

> > 2.3 SCTP multihoming and Network Adress Translators(NAT)
> >
> > ==> this very same problem will come up with single-homed NAT's too (need
> > to be able to do ALG translation), right?  Can be deleted.
> 
> What exactly can be deleted ?

Whole section 2.3.  Or perhaps, some considerations might be  preservable.

> Well, actually that section probably has not taken consideration at 
> all of single-homed NATs (which in turn is not something that comes to 
> mind at once when talking about a NAT). 
>
> This is needed because single-homed _SCTP_ works together with NATs nicely 
> (because no multiple IP addresses are carried inside the SCTP packets in the 
> association setup phase), but multi-homed SCTP carries these IP addresses 
> inside the INIT and INIT ACK chunks, so something needs to be tweaked in
> that case.

If you use IP+SCTP, NAT still has to support rewriting the port numbers 
and IP numbers.  Therefore I gather NAT will have to be modified anyway 
cope with the fact.  This will probably include going into INIT/INIT ACK 
chunks, no?  

Anyway, multihoming through _one_ NAT seems rather stupid.  Via different 
NAT boxes it might even make a little sense.

-- 
Pekka Savola                 "Tell me of difficulties surmounted,
Netcore Oy                   not those you stumble over and fall"
Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 10:25:44 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA10892
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 10:25:44 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA13137
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 10:25:48 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA11878;
	Fri, 8 Feb 2002 10:02:00 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA11825
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 10:01:57 -0500 (EST)
Received: from netcore.fi (netcore.fi [193.94.160.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA09647;
	Fri, 8 Feb 2002 10:01:52 -0500 (EST)
Received: from localhost (pekkas@localhost)
	by netcore.fi (8.11.6/8.11.6) with ESMTP id g18F1ED08259;
	Fri, 8 Feb 2002 17:01:14 +0200
Date: Fri, 8 Feb 2002 17:01:13 +0200 (EET)
From: Pekka Savola <pekkas@netcore.fi>
To: Coene Lode <Lode.Coene@siemens.atea.be>
cc: "Andreas Jungmayer (E-mail)" <ajung@exp-math.uni-essen.de>,
        "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)" <sigtran@ietf.org>,
        "'multi6@ops.ietf.org'" <multi6@ops.ietf.org>
In-Reply-To: <E76F715C0429D5118F2100508BB9EDEE0795B1@hrtades7.atea.be>
Message-ID: <Pine.LNX.4.44.0202081657560.8027-100000@netcore.fi>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Subject: [Tsvwg] RE: SCTP multihoming issues draft
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

On Fri, 8 Feb 2002, Coene Lode wrote:
> >     This document describes issues of the Stream Control Transmission
> >     Protocol (SCTP)[RFC2960] in regard to multihoming on the
> >     Internet. It explores cases where through situations in the
> >     internet, single points-of-failure can occur even when using
> >     multihoming and what the impact is of multihoming on the host
> >     routing tables.
> > 
> > ==> what are "host routing tables" ?
>
> selecting the correct outgoing link is a routing descision. See also Andreas
> comments (with which I agree).

"the routing table on hosts" would be better, perhaps.

> > ==> on a generic note: it seems to me there is a need for routing
> > tables on hosts only if one wants to optimize the paths: one is used
> > by default, and if it fails, just (try to) switch to the next one,
> > right?
> > 
> 
> See previous remark on routing tables.

See my reply to Andreas. (note to others: multi6 was stripped from Cc:).

In short, routing table is naturally required, but the draft talks about 
the need to "download" a fuller table from edge routers on hosts.  IMO 
this is unnecessary.

-- 
Pekka Savola                 "Tell me of difficulties surmounted,
Netcore Oy                   not those you stumble over and fall"
Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 10:42:14 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA11831
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 10:42:14 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA14456
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 10:42:17 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA13260;
	Fri, 8 Feb 2002 10:27:02 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA13225
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 10:26:59 -0500 (EST)
Received: from zcamail04.zca.compaq.com (zcamail04.zca.compaq.com [161.114.32.104])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA10925
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 10:26:55 -0500 (EST)
Received: from cacexg11.americas.cpqcorp.net (cacexg11.americas.cpqcorp.net [16.105.250.94])
	by zcamail04.zca.compaq.com (Postfix) with ESMTP
	id 12EEC8BB; Fri,  8 Feb 2002 07:30:03 -0800 (PST)
Received: from txnexc01.americas.cpqcorp.net ([16.74.7.244]) by cacexg11.americas.cpqcorp.net with Microsoft SMTPSVC(5.0.2195.2966);
	 Fri, 8 Feb 2002 07:26:20 -0800
X-MimeOLE: Produced By Microsoft Exchange V6.0.4712.0
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Date: Fri, 8 Feb 2002 09:26:19 -0600
Message-ID: <717252EC3E37AE4392E2614EA24E9F2B15A08F@txnexc01.americas.cpqcorp.net>
Thread-Topic: ICSC Sockets API Extensions Charter Approved
Thread-Index: AcGwXpqVUZxpLNeMQiqvrk85j/UD1AAVFZbQ
From: "Teisberg, Robert" <robert.teisberg@compaq.com>
To: "Douglas Otis" <dotis@sanlight.net>, <icsc-public@opengroup.org>,
        "Martin Kirk" <m.kirk@opengroup.org>
Cc: "Tsvwg" <tsvwg@ietf.org>
X-OriginalArrivalTime: 08 Feb 2002 15:26:20.0128 (UTC) FILETIME=[F56D7E00:01C1B0B4]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id KAA13226
Subject: [Tsvwg] RE: ICSC Sockets API Extensions Charter Approved
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Comments interspersed.

> 
> There is work establishing a DDP (Direct Data Placement) and 
> perhaps an RDMA
> layer or shim for SCTP.  This provides methods for allowing 
> direct placement
> with/without tags and would also allow emulation of TCP 
> partial writes.
> Please note this allows "TCP/IP stack bypass" while providing 
> congestion
> avoidance without introducing new IP protocols.  SCTP is used without
> modification for this purpose.

There is also work that has been ongoing toward a transport-independent
RDMA protocol that precedes the SCTP work by about a year.  The latter
was the subject of a BOF at the 52nd IETF in Salt Lake City last
December.

A couple of points to avoid misunderstanding: 1) No IETF working group
has yet been chartered to develop RDMA specifications; 2) None of the
proposals within the IETF for RDMA protocols that I've seen requires
changes to transport or lower layer protocols.

> 
> There are already documents addressing Socket Descriptions 
> for SCTP at:
> http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpsocket-03.txt
> 
> How do you envision a collaboration with this ongoing effort?

And lest there be any misunderstanding about how official the IETF work
is at this point, I refer you to
http://www.ietf.org/ietf/1id-guidelines.txt, about midway through the
document:

     Internet-Drafts are draft documents valid for a maximum of six
     months and may be updated, replaced, or obsoleted by other
     documents at any time.  It is inappropriate to use Internet-
     Drafts as reference material or to cite them other than as
     "work in progress."

In other words, it's good to learn from Internet Drafts, and sometimes
standards develop from them, but they aren't standards and in many cases
never will be standards.

> 
> A DDP/(RDMA)/SCTP draft will be available soon.  Presently 
> this document is
> being reviewed by Stephen Bailey [mailto:steph@cs.uchicago.edu] and
Allyn
> Romanow [mailto:allyn@cisco.com] of the RDMA effort and will be posted
to
> the IETF and announced on tsvwg:

As will a set of drafts describing a transport-independent DDP/RDMA
framework.

> An RDMA reflector is at: RDMA [mailto:rdma@yahoogroups.com]
> http://groups.yahoo.com/group/RDMA/

This reflector has had extensive traffic for over a year and is followed
by 260 subscribers.

> An SCTP-DDP reflector is at: SCTP-DDP
[mailto:sctp-ddp@yahoogroups.com]
> http://groups.yahoo.com/group/SCTP-DDP/

This one was founded in January, and as far as I know contains no
substantive traffic except cross-posts to the rdma reflector.

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 11:09:01 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA12885
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 11:09:01 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA16593
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 11:09:03 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA14570;
	Fri, 8 Feb 2002 10:45:58 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA14537
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 10:45:56 -0500 (EST)
Received: from hrtades9.atea.be (viap111.atea.be [194.78.143.111])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA12034;
	Fri, 8 Feb 2002 10:45:51 -0500 (EST)
Received: from hrtades10.atea.be (siemens.atea.be [139.10.143.141]) by hrtades9.atea.be with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13)
	id DF4YR0LN; Fri, 8 Feb 2002 16:45:14 +0100
Received: by siemens.atea.be with Internet Mail Service (5.5.2653.19)
	id <D6SNGSNP>; Fri, 8 Feb 2002 16:45:39 +0100
Message-ID: <E76F715C0429D5118F2100508BB9EDEE0795B2@hrtades7.atea.be>
From: Coene Lode <Lode.Coene@siemens.atea.be>
To: "'Pekka Savola'" <pekkas@netcore.fi>,
        Andreas Jungmaier
	 <ajung@exp-math.uni-essen.de>
Cc: Coene Lode <Lode.Coene@siemens.atea.be>,
        "Transport Area wg (E-mail)"
	 <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)" <sigtran@ietf.org>
Subject: RE: [Tsvwg] Re: SCTP multihoming issues draft
Date: Fri, 8 Feb 2002 16:43:04 +0100 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

> -----Original Message-----
> From: Pekka Savola [mailto:pekkas@netcore.fi]
> Sent: vrijdag 8 februari 2002 15:47
> To: Andreas Jungmaier
> Cc: Coene Lode; Transport Area wg (E-mail); SIGTRAN (E-mail)
> Subject: Re: [Tsvwg] Re: SCTP multihoming issues draft
> 
> 
> On Fri, 8 Feb 2002, Andreas Jungmaier wrote:
> > > Abstract
> > >     This document describes issues of the Stream Control 
> Transmission
> > >     Protocol (SCTP)[RFC2960] in regard to multihoming on the
> > >     Internet. It explores cases where through situations in the
> > >     internet, single points-of-failure can occur even when using
> > >     multihoming and what the impact is of multihoming on the host
> > >     routing tables.
> > >
> > > ==> what are "host routing tables" ?
> > 
> > I suppose that is what you get when you type in "netstat 
> -rn" on a decent
> > unix machine. Anyway, this is probably confusing: it should 
> probably say
> > something like "...and how the SCTP multihoming interacts 
> with routing"
> > since that is what is of concern here: the interface 
> between a transport
> > protocol and the network layer.
> 
> That seems more sane, or 'routing tables on hosts'.

OK, will be updated, no problem

> 
> > >     It might be usefull to explore ways where no routing 
> tables are
> > >
> > > ==> on a generic note: it seems to me there is a need for 
> routing tables
> > > on hosts only if one wants to optimize the paths: one is 
> used by default,
> > > and if it fails, just (try to) switch to the next one, right?
> > 
> > Any host that has an IP address has some routing tables. 
> Richard Stevens 
> > has put that nicely in his "TCP/IP Illustrated, Vol.1 - The 
> Protocols", 
> > chapter 9. That's recommended reading :-)
> 
> I know my Stevens very well, thank you.
> 
> If you read the draft around that place, you'll notice the 
> question was 
> about SCTP hosts somehow having to dump routing tables _from 
> a router_ on 
> hosts to make sensible decisions.  My belief is that 
> "regular" routing 
> table (ie. basically about 4 routes) is enough.

That migh be the case for host A talking with host B for most of the time.
In case of a webserver, you a lot of host X's with different addresses
coming together on a host A. How will that host A make a descision which
outgoing interface/link to choose?
Default routes will be of no help. Maybe educated guesses for loadsharing
purposes may be used but then be sure that for a connection from A to Xi ,
take for every packet the same path! Otherwise, the congestion control
algorithms may go haywire(feedback would be erronous). If that happen, then
we have really entered dangerous territory.
> 
> > > 2.3 SCTP multihoming and Network Adress Translators(NAT)
> > >
> > > ==> this very same problem will come up with single-homed 
> NAT's too (need
> > > to be able to do ALG translation), right?  Can be deleted.
> > 
> > What exactly can be deleted ?
> 
> Whole section 2.3.  Or perhaps, some considerations might be  
> preservable.
> 

To me the section makes a point(a vital one). 

> > Well, actually that section probably has not taken consideration at 
> > all of single-homed NATs (which in turn is not something 
> that comes to 
> > mind at once when talking about a NAT). 
> >
> > This is needed because single-homed _SCTP_ works together 
> with NATs nicely 
> > (because no multiple IP addresses are carried inside the 
> SCTP packets in the 
> > association setup phase), but multi-homed SCTP carries 
> these IP addresses 
> > inside the INIT and INIT ACK chunks, so something needs to 
> be tweaked in
> > that case.
> 
> If you use IP+SCTP, NAT still has to support rewriting the 
> port numbers 
> and IP numbers.  Therefore I gather NAT will have to be 
> modified anyway 
> cope with the fact.  This will probably include going into 
> INIT/INIT ACK 
> chunks, no?  
> 
> Anyway, multihoming through _one_ NAT seems rather stupid.  
> Via different 
> NAT boxes it might even make a little sense.

Well, that would be a really hard one. The INIT goes only through one NAT
box(one path),
the boxes on the other path can only guess on one or more paths of a SCTP
association running through them. That is the really difficult part. The
DATA chunk could be used for that, but what happens if the association is
terminated: the ABORT/TERMINATE also does not pass through all paths. So all
those NAT boxes never know when to kill off  a SCTP path/association. A
watch for heartbeat may help but then it has to guess the heartbeat timer,
which can goes from milliseconds to hours. The basic problem is : NAT boxes
have to staore state for the SCTP association that passes through them.:
they don't know when the association starts(except for the lucky one which
let the INIT through) nor do they know when the association is
terminated(except for the lucky one which let the ABORT/TERMINATE chunk
through, and that might not be the same box which let through the INIT
chunk): because the NAT boxes have to gues what the endpoint is up to, they
will either (a) screw up the association(dropping heartbeats or data) or (b)
screw up the NAT box(got zillions of SCTP paths going through but I don't
know which ones are still active)

conclusion:
NAT boxes are not really helpfull. a conclusion that has be debated over and
over again. 
OK , this section is just a addition to discussion about NAT boxes in
general. 
From here on, the discussion will degenerate in a filosofical discussion
about NAT boxes.
So I stop.

> 
> -- 
> Pekka Savola                 "Tell me of difficulties surmounted,
> Netcore Oy                   not those you stumble over and fall"
> Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords
> 

Yours sincerly,
Lode Coene

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 11:15:41 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA13149
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 11:15:40 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA17074
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 11:15:42 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA14988;
	Fri, 8 Feb 2002 10:52:53 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA14942
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 10:52:50 -0500 (EST)
Received: from hrtades9.atea.be (viap111.atea.be [194.78.143.111])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA12297;
	Fri, 8 Feb 2002 10:52:48 -0500 (EST)
Received: from hrtades10.atea.be (siemens.atea.be [139.10.143.141]) by hrtades9.atea.be with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13)
	id DF4YR0NL; Fri, 8 Feb 2002 16:52:11 +0100
Received: by siemens.atea.be with Internet Mail Service (5.5.2653.19)
	id <D6SNGSRC>; Fri, 8 Feb 2002 16:52:36 +0100
Message-ID: <E76F715C0429D5118F2100508BB9EDEE0795B3@hrtades7.atea.be>
From: Coene Lode <Lode.Coene@siemens.atea.be>
To: "'Pekka Savola'" <pekkas@netcore.fi>,
        Andreas Jungmaier
	 <ajung@exp-math.uni-essen.de>
Cc: Coene Lode <Lode.Coene@siemens.atea.be>,
        "Transport Area wg (E-mail)"
	 <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)" <sigtran@ietf.org>
Subject: RE: [Tsvwg] Re: SCTP multihoming issues draft
Date: Fri, 8 Feb 2002 16:50:01 +0100 
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Sorry, Pekka, Forgot to reply on one of your points.

> > Well, actually that section probably has not taken consideration at 
> > all of single-homed NATs (which in turn is not something 
> that comes to 
> > mind at once when talking about a NAT). 
> >
> > This is needed because single-homed _SCTP_ works together 
> with NATs nicely 
> > (because no multiple IP addresses are carried inside the 
> SCTP packets in the 
> > association setup phase), but multi-homed SCTP carries 
> these IP addresses 
> > inside the INIT and INIT ACK chunks, so something needs to 
> be tweaked in
> > that case.
> 
> If you use IP+SCTP, NAT still has to support rewriting the 
> port numbers 
> and IP numbers.  Therefore I gather NAT will have to be 
> modified anyway 
> cope with the fact.  This will probably include going into 
> INIT/INIT ACK 
> chunks, no?  
> 
you are 100% right. The NAT boxes has to look in the INIT/INIT ACK.(and
probably do some stuff) 

Yours sicnerly,
Lode Coene

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 11:17:33 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA13235
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 11:17:32 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA17158
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 11:17:35 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA16310;
	Fri, 8 Feb 2002 11:05:02 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA16250
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 11:05:00 -0500 (EST)
Received: from netcore.fi (netcore.fi [193.94.160.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA12794;
	Fri, 8 Feb 2002 11:04:56 -0500 (EST)
Received: from localhost (pekkas@localhost)
	by netcore.fi (8.11.6/8.11.6) with ESMTP id g18G4IV08740;
	Fri, 8 Feb 2002 18:04:18 +0200
Date: Fri, 8 Feb 2002 18:04:17 +0200 (EET)
From: Pekka Savola <pekkas@netcore.fi>
To: Coene Lode <Lode.Coene@siemens.atea.be>
cc: Andreas Jungmaier <ajung@exp-math.uni-essen.de>,
        "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)" <sigtran@ietf.org>
Subject: RE: [Tsvwg] Re: SCTP multihoming issues draft
In-Reply-To: <E76F715C0429D5118F2100508BB9EDEE0795B2@hrtades7.atea.be>
Message-ID: <Pine.LNX.4.44.0202081757000.8307-100000@netcore.fi>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

On Fri, 8 Feb 2002, Coene Lode wrote:
> > If you read the draft around that place, you'll notice the question
> > was about SCTP hosts somehow having to dump routing tables _from a
> > router_ on hosts to make sensible decisions.  My belief is that
> > "regular" routing table (ie. basically about 4 routes) is enough.
> 
> That migh be the case for host A talking with host B for most of the time.
> In case of a webserver, you a lot of host X's with different addresses
> coming together on a host A. How will that host A make a descision which
> outgoing interface/link to choose?

I'm assuming the said webserver is multihomed, and inbound traffic is 
distributed by some mechanism (e.g. more specific BGP advertisements to 
some destinations).

> Default routes will be of no help. Maybe educated guesses for loadsharing
> purposes may be used but then be sure that for a connection from A to Xi ,
> take for every packet the same path! Otherwise, the congestion control
> algorithms may go haywire(feedback would be erronous). If that happen, then
> we have really entered dangerous territory.

Please note that I don't see a clear possibility of ensuring completely
aligned inbound/outbound traffic as you can (almost) never be sure how far
the advertisments for the other interface go.

Currently, I believe this is just "solved" by common load-balancing 
by every packet/destination, or the webserver participates in routing 
system, by e.g. receiving more specific routes to some destinations.  I 
think the latter option is rather rare.

But you're treading into a territory that is a general problem with 
multihomed systems already: nothing new to SCTP AFAICS; as such, I don't 
see how SCTP would be any different from current practise.

-- 
Pekka Savola                 "Tell me of difficulties surmounted,
Netcore Oy                   not those you stumble over and fall"
Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 11:43:40 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA14235
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 11:43:39 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA19394
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 11:43:42 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA18356;
	Fri, 8 Feb 2002 11:31:13 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA18295
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 11:31:09 -0500 (EST)
Received: from irmgard.exp-math.uni-essen.de (irmgard.exp-math.uni-essen.de [132.252.150.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA13800;
	Fri, 8 Feb 2002 11:31:04 -0500 (EST)
Received: from there (biene.exp-math.uni-essen.de [132.252.150.215])
	by irmgard.exp-math.uni-essen.de (8.12.2/8.12.2) with SMTP id g18GV5j9035194;
	Fri, 8 Feb 2002 17:31:06 +0100
Message-Id: <200202081631.g18GV5j9035194@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain;
  charset="iso-8859-1"
From: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
Organization: University Essen
To: Pekka Savola <pekkas@netcore.fi>
Subject: Re: [Tsvwg] Re: SCTP multihoming issues draft
Date: Fri, 8 Feb 2002 17:30:44 +0100
X-Mailer: KMail [version 1.3.1]
Cc: Coene Lode <Lode.Coene@siemens.atea.be>,
        "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)" <sigtran@ietf.org>
References: <Pine.LNX.4.44.0202081632300.7919-100000@netcore.fi>
In-Reply-To: <Pine.LNX.4.44.0202081632300.7919-100000@netcore.fi>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Pekka, dear all,

please see my comments inline...

> >
> > Any host that has an IP address has some routing tables. Richard Stevens
> > has put that nicely in his "TCP/IP Illustrated, Vol.1 - The Protocols",
> > chapter 9. That's recommended reading :-)
>
> I know my Stevens very well, thank you.

Sorry, I couldn't resist :)

> If you read the draft around that place, you'll notice the question was
> about SCTP hosts somehow having to dump routing tables _from a router_ on
> hosts to make sensible decisions.  My belief is that "regular" routing
> table (ie. basically about 4 routes) is enough.

We agree on that point. If I want to use SCTP multi-homing sensibly, I must
rely on the fact that my network is administered in a way that *sensibly*
supports this, and that my network admin gives me the right data to install
the right routing tables on my host. 
Any installation of routes from a router basically would make the end-system
participate in a dynamic routing protocol. That is not necessarily needed, 
but, on the other hand, is nothing special, either.

> > > 2.3 SCTP multihoming and Network Adress Translators(NAT)
> > >
> > > ==> this very same problem will come up with single-homed NAT's too
> > > (need to be able to do ALG translation), right?  Can be deleted.
> >
> > What exactly can be deleted ?
> Whole section 2.3.  Or perhaps, some considerations might be  preservable.

No, it might need to be rewritten, but there must be a statement
about SCTP/NAT interaction ! We (and many others) want to know about
it...
 
> If you use IP+SCTP, NAT still has to support rewriting the port numbers
> and IP numbers.  Therefore I gather NAT will have to be modified anyway
> cope with the fact.  This will probably include going into INIT/INIT ACK
> chunks, no?

Yes, you are right: 
A NAT box will do the exact same thing for SCTP as it does for TCP
iff the concerned SCTP association is single-homed.
Only if an SCTP association is multi-homed, the NAT needs to also
dig into the INIT/INIT-ACK fields to modifiy these.
Since a NAT box cannot know a priori whether an association is
single- or multi-homed, it will in any case have to search through
the entire INIT/INIT-ACK chunks to check for addresses.

> Anyway, multihoming through _one_ NAT seems rather stupid.  
Talking from the redundancy viewpoint that sure is correct. 
"NATs are nasty. Don't use them for SCTP." 
Maybe that suffices in the draft ???

> Via different NAT boxes it might even make a little sense.
...and you got yourself a whole lot of other problems...

Well: one NAT box is a single point of failure, many NAT boxes
would build up many,many states -- that seems to be an interesting
problem so far unsolved, isn't it ?

- -- 
- ------------------------------------------------------------------------
Dipl.-Ing. Andreas Jungmaier              |                         _
Computer Networking Technology Group      | ASCII ribbon campaign  ( )
University of Essen                       |  - against HTML email   X
http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8Y/01XAsLBNOCSsARAmrjAJ9krwePkqDwgY8IQoiZgIOi6Ay8MQCfVyJA
8GVP7yMWwwgFX6UiqsXAwro=
=mGIy
-----END PGP SIGNATURE-----

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 12:06:28 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA14884
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 12:06:28 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id MAA22490
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 12:06:30 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA20347;
	Fri, 8 Feb 2002 11:48:25 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA20297
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 11:48:20 -0500 (EST)
Received: from d12lmsgate-2.de.ibm.com (d12lmsgate-2.de.ibm.com [195.212.91.200])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA14464
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 11:48:16 -0500 (EST)
Received: from d12relay01.de.ibm.com (d12relay01.de.ibm.com [9.165.215.22])
	by d12lmsgate-2.de.ibm.com (1.0.0) with ESMTP id RAA54146;
	Fri, 8 Feb 2002 17:47:48 +0100
Received: from d10hubm1.telaviv.ibm.com (d10ml001.telaviv.ibm.com [9.148.216.55])
	by d12relay01.de.ibm.com (8.11.1m3/NCO v5.01) with ESMTP id g18GnGf22232;
	Fri, 8 Feb 2002 17:49:16 +0100
Subject: Re: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch framer for
 TCP
To: Lloyd Wood <L.Wood@eim.surrey.ac.uk>
Cc: Julian Satran <Julian_Satran@il.ibm.com>, rdma <rdma@yahoogroups.com>,
        tsvwg <tsvwg@ietf.org>
X-Mailer: Lotus Notes Release 5.0.9  November 16, 2001
Message-ID: <OFC07BED68.4834B04D-ONC2256B5A.0001E37B@telaviv.ibm.com>
From: "Julian Satran" <Julian_Satran@il.ibm.com>
Date: Fri, 8 Feb 2002 02:58:29 +0200
X-MIMETrack: Serialize by Router on D10ML001/10/M/IBM(Release 5.0.8 |June 18, 2001) at
 08/02/2002 18:49:18
MIME-Version: 1.0
Content-type: text/plain; charset=us-ascii
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

yes - except that unlike a dictionary builder I do not attempt to guarantee
that it is an "only once" word :-)  julo


                                                                                              
                    Lloyd Wood                                                                
                    <l.wood@eim.sur       To:     Julian Satran/Haifa/IBM@IBMIL               
                    rey.ac.uk>            cc:     tsvwg <tsvwg@ietf.org>, rdma                
                                           <rdma@yahoogroups.com>                             
                    07-02-02 22:54        Subject:     Re: [Tsvwg] RE: [rdma] a proposal for  
                    Please respond         a different No-Data-Touch framer for TCP           
                    to Lloyd Wood                                                             
                                                                                              
                                                                                              



On Thu, 7 Feb 2002, Julian Satran wrote:

> The salt is a random number that is changing on every PDU.

you mean a nonce?

L.

<L.Wood@surrey.ac.uk>PGP<http://www.ee.surrey.ac.uk/Personal/L.Wood/>






_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 13:58:49 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA19439
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 13:58:48 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA28843
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 13:58:50 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA27915;
	Fri, 8 Feb 2002 13:33:26 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA27888
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 13:33:25 -0500 (EST)
Received: from c003.snv.cp.net (c003-h004.c003.snv.cp.net [209.228.32.218])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id NAA18679
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 13:33:22 -0500 (EST)
Received: (cpmta 9726 invoked from network); 8 Feb 2002 10:32:53 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.218) with SMTP; 8 Feb 2002 10:32:53 -0800
X-Sent: 8 Feb 2002 18:32:53 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <icsc-public@opengroup.org>
Cc: "Tsvwg" <tsvwg@ietf.org>
Subject: RE: [Tsvwg] RE: ICSC Sockets API Extensions Charter Approved
Date: Fri, 8 Feb 2002 10:34:49 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJKEIPCPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Importance: Normal
In-Reply-To: <717252EC3E37AE4392E2614EA24E9F2B15A08F@txnexc01.americas.cpqcorp.net>
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Robert,

> Comments interspersed.
>
> > There is work establishing a DDP (Direct Data Placement) and
> > perhaps an RDMA layer or shim for SCTP.  This provides methods
> > for allowing direct placement with/without tags and would also
> > allow emulation of TCP partial writes.  Please note this allows
> > "TCP/IP stack bypass" while providing congestion avoidance
> > without introducing new IP protocols.  SCTP is used without
> > modification for this purpose.
>
> There is also work that has been ongoing toward a transport-independent
> RDMA protocol that precedes the SCTP work by about a year.  The latter
> was the subject of a BOF at the 52nd IETF in Salt Lake City last
> December.
>
> A couple of points to avoid misunderstanding: 1) No IETF working group
> has yet been chartered to develop RDMA specifications; 2) None of the
> proposals within the IETF for RDMA protocols that I've seen requires
> changes to transport or lower layer protocols.

There is ongoing work in creating a DDP layer for SCTP as indicated in the
pending announcement of a draft within the tsvwg.  The only IP protocol
suitable for RDMA would be SCTP unless there is interest in something like
UDP or RDP which would incur acceptance issues.

> > There are already documents addressing Socket Descriptions
> > for SCTP at:
> > http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpsocket-03.txt
> >
> > How do you envision a collaboration with this ongoing effort?
>
<snip>
>
> In other words, it's good to learn from Internet Drafts, and sometimes
> standards develop from them, but they aren't standards and in many cases
> never will be standards.

A draft such as this can become an informational RFC.  None the less, it
indicates ongoing efforts.  Unless you wish to exclude SCTP, there should be
interest in this effort.

> > A DDP/(RDMA)/SCTP draft will be available soon.  Presently
> > this document is being reviewed by Stephen Bailey
> > [mailto:steph@cs.uchicago.edu] and Allyn Romanow
> > [mailto:allyn@cisco.com] of the RDMA effort and will be posted
> > to the IETF and announced on tsvwg:
>
> As will a set of drafts describing a transport-independent DDP/RDMA
> framework.
>
> > An RDMA reflector is at: RDMA [mailto:rdma@yahoogroups.com]
> > http://groups.yahoo.com/group/RDMA/
>
> This reflector has had extensive traffic for over a year and is followed
> by 260 subscribers.

That is correct.  RDMA has been the sole focus prior to the BOF.  It is not
clear however if there is interest within this group to focus on SCTP as
anything more than a compatibility point for RDMA.  It would appear there is
a conflict within this group regarding transport.  The consideration of DDP
as related to SCTP would be an example rather than being classified as just
a subset of RDMA (RDMA write) devoid of SCTP features.

> > An SCTP-DDP reflector is at: SCTP-DDP
> [mailto:sctp-ddp@yahoogroups.com]
> > http://groups.yahoo.com/group/SCTP-DDP/
>
> This one was founded in January, and as far as I know contains no
> substantive traffic except cross-posts to the rdma reflector.

This is also correct.  SCTP with DDP is the sole focus for this reflector
but it only contains related cross posts at this time.  Adopting an SCTP-DDP
layer into other protocols such as RPC or entire stacks would be interesting
topics separate from RDMA.  It would also allow a departure from RDMA work
that must conform to InfiniBand or some other unspecified transport to allow
utilization of features unique to SCTP.

If it becomes clear the RDMA group is interested in expanding their charter
to include such topics, then this reflector will likely become moribund.

Doug




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 14:37:41 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA21587
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 14:37:41 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA01320
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 14:37:43 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA00462;
	Fri, 8 Feb 2002 14:24:24 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA00439
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 14:24:22 -0500 (EST)
Received: from zcamail05.zca.compaq.com (zcamail05.zca.compaq.com [161.114.32.105])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA20977
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 14:24:16 -0500 (EST)
Received: from cacexg11.americas.cpqcorp.net (cacexg11.americas.cpqcorp.net [16.105.250.94])
	by zcamail05.zca.compaq.com (Postfix) with ESMTP
	id 9E2B42141; Fri,  8 Feb 2002 11:30:02 -0800 (PST)
Received: from txnexc01.americas.cpqcorp.net ([16.74.7.244]) by cacexg11.americas.cpqcorp.net with Microsoft SMTPSVC(5.0.2195.2966);
	 Fri, 8 Feb 2002 11:23:47 -0800
X-MimeOLE: Produced By Microsoft Exchange V6.0.4712.0
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Subject: RE: [Tsvwg] RE: ICSC Sockets API Extensions Charter Approved
Date: Fri, 8 Feb 2002 13:23:45 -0600
Message-ID: <717252EC3E37AE4392E2614EA24E9F2B15A420@txnexc01.americas.cpqcorp.net>
Thread-Topic: [Tsvwg] RE: ICSC Sockets API Extensions Charter Approved
Thread-Index: AcGwzzDovyjnV35PQD2XzqOSgLL4agABnKtQ
From: "Teisberg, Robert" <robert.teisberg@compaq.com>
To: "Douglas Otis" <dotis@sanlight.net>, <icsc-public@opengroup.org>
Cc: "Tsvwg" <tsvwg@ietf.org>
X-OriginalArrivalTime: 08 Feb 2002 19:23:47.0490 (UTC) FILETIME=[21846820:01C1B0D6]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id OAA00440
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit


> 
> There is ongoing work in creating a DDP layer for SCTP as 
> indicated in the
> pending announcement of a draft within the tsvwg.  The only 
> IP protocol
> suitable for RDMA would be SCTP unless there is interest in 
> something like
> UDP or RDP which would incur acceptance issues.

I'm not going to get into this except to note that not everyone agrees
SCTP is the only suitable protocol.

> 
> A draft such as this can become an informational RFC.  None 
> the less, it
> indicates ongoing efforts.  Unless you wish to exclude SCTP, 
> there should be
> interest in this effort.

I don't wish to exclude SCTP.  On the contrary, I wish to include as
many alternatives as possible.

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 15:41:50 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA24111
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 15:41:50 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA04360
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 15:41:52 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA03288;
	Fri, 8 Feb 2002 15:17:23 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA03259
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 15:17:20 -0500 (EST)
Received: from c003.snv.cp.net (c003-h000.c003.snv.cp.net [209.228.32.214])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id PAA23341
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 15:17:16 -0500 (EST)
Received: (cpmta 13664 invoked from network); 8 Feb 2002 12:16:47 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.214) with SMTP; 8 Feb 2002 12:16:47 -0800
X-Sent: 8 Feb 2002 20:16:47 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <icsc-public@opengroup.org>
Cc: "Tsvwg" <tsvwg@ietf.org>
Subject: RE: [Tsvwg] RE: ICSC Sockets API Extensions Charter Approved
Date: Fri, 8 Feb 2002 12:18:43 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJCEJBCPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Importance: Normal
In-Reply-To: <717252EC3E37AE4392E2614EA24E9F2B15A420@txnexc01.americas.cpqcorp.net>
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Robert,

> > There is ongoing work in creating a DDP layer for SCTP as
> > indicated in the pending announcement of a draft within the
> > tsvwg.  The only IP protocol suitable for RDMA would be
> > SCTP unless there is interest in something like UDP or RDP
> > which would incur acceptance issues.
>
> I'm not going to get into this except to note that not everyone agrees
> SCTP is the only suitable protocol.

There are efforts related to the introduction of an added layer beneath or
within TCP to handle framing in a fashion analogous to a scheme proposed for
iSCSI and as with Julian Satran and Stephen Bailey's ongoing efforts toward
such introduction.  Framing would be a requirement for RDMA if direct
placement were desired.  You should also note such TCP framing schemes are
drafts and are not classified as part of the IP protocol suite.  I believe
the ICSC announcement made reference to this as a TCP/IP stack bypass.

> > A draft such as this can become an informational RFC.  None the
> > less, it indicates ongoing efforts.  Unless you wish to exclude SCTP,
> > there should be interest in this effort.
>
> I don't wish to exclude SCTP.  On the contrary, I wish to include as
> many alternatives as possible.

It is good SCTP is being considered.  As SCTP offers more features, there
may be additional options found when working with SCTP not seen on
alternatives and you will note SCTP does have an impact on the socket API.
A view by RDMA in seeing SCTP as a subset of possible transports is not a
problem.  SCTP offers capabilities beyond alternatives and should receive
additional consideration for applications beyond RDMA as well.  Specialized
adapters for each application seems a poor approach.

Doug




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb  8 15:44:08 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA24143
	for <tsvwg-archive@odin.ietf.org>; Fri, 8 Feb 2002 15:44:08 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA04431
	for tsvwg-archive@odin.ietf.org; Fri, 8 Feb 2002 15:44:11 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA03482;
	Fri, 8 Feb 2002 15:23:12 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA03457
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 15:23:10 -0500 (EST)
Received: from zcamail04.zca.compaq.com (zcamail04.zca.compaq.com [161.114.32.104])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA23623
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 15:23:07 -0500 (EST)
Received: from cacexg12.americas.cpqcorp.net (cacexg12.americas.cpqcorp.net [16.105.250.84])
	by zcamail04.zca.compaq.com (Postfix) with ESMTP
	id 737E09FA; Fri,  8 Feb 2002 12:26:18 -0800 (PST)
Received: from txnexc01.americas.cpqcorp.net ([16.74.7.244]) by cacexg12.americas.cpqcorp.net with Microsoft SMTPSVC(5.0.2195.2966);
	 Fri, 8 Feb 2002 15:22:35 -0500
X-MimeOLE: Produced By Microsoft Exchange V6.0.4712.0
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Subject: RE: [Tsvwg] RE: ICSC Sockets API Extensions Charter Approved
Date: Fri, 8 Feb 2002 14:22:34 -0600
Message-ID: <717252EC3E37AE4392E2614EA24E9F2B15A421@txnexc01.americas.cpqcorp.net>
Thread-Topic: [Tsvwg] RE: ICSC Sockets API Extensions Charter Approved
Thread-Index: AcGw3atvmwX9CmhZQQqB3xxC6SAoNgAAHW7w
From: "Teisberg, Robert" <robert.teisberg@compaq.com>
To: "Douglas Otis" <dotis@sanlight.net>, <icsc-public@opengroup.org>
Cc: "Tsvwg" <tsvwg@ietf.org>
X-OriginalArrivalTime: 08 Feb 2002 20:22:35.0906 (UTC) FILETIME=[589DF620:01C1B0DE]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id PAA03458
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Can we please keep IETF debates out of TOG?

> -----Original Message-----
> From: Douglas Otis [mailto:dotis@sanlight.net]
> Sent: Friday, February 08, 2002 2:19 PM
> To: icsc-public@opengroup.org
> Cc: Tsvwg
> Subject: RE: [Tsvwg] RE: ICSC Sockets API Extensions Charter Approved
> 
> 
> Robert,
> 
> > > There is ongoing work in creating a DDP layer for SCTP as
> > > indicated in the pending announcement of a draft within the
> > > tsvwg.  The only IP protocol suitable for RDMA would be
> > > SCTP unless there is interest in something like UDP or RDP
> > > which would incur acceptance issues.
> >
> > I'm not going to get into this except to note that not 
> everyone agrees
> > SCTP is the only suitable protocol.
> 
> There are efforts related to the introduction of an added 
> layer beneath or
> within TCP to handle framing in a fashion analogous to a 
> scheme proposed for
> iSCSI and as with Julian Satran and Stephen Bailey's ongoing 
> efforts toward
> such introduction.  Framing would be a requirement for RDMA if direct
> placement were desired.  You should also note such TCP 
> framing schemes are
> drafts and are not classified as part of the IP protocol 
> suite.  I believe
> the ICSC announcement made reference to this as a TCP/IP stack bypass.
> 
> > > A draft such as this can become an informational RFC.  None the
> > > less, it indicates ongoing efforts.  Unless you wish to 
> exclude SCTP,
> > > there should be interest in this effort.
> >
> > I don't wish to exclude SCTP.  On the contrary, I wish to include as
> > many alternatives as possible.
> 
> It is good SCTP is being considered.  As SCTP offers more 
> features, there
> may be additional options found when working with SCTP not seen on
> alternatives and you will note SCTP does have an impact on 
> the socket API.
> A view by RDMA in seeing SCTP as a subset of possible 
> transports is not a
> problem.  SCTP offers capabilities beyond alternatives and 
> should receive
> additional consideration for applications beyond RDMA as 
> well.  Specialized
> adapters for each application seems a poor approach.
> 
> Doug
> 
> 
> 
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Feb  9 00:14:19 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA03688
	for <tsvwg-archive@odin.ietf.org>; Sat, 9 Feb 2002 00:14:19 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id AAA25222
	for tsvwg-archive@odin.ietf.org; Sat, 9 Feb 2002 00:14:21 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id XAA24301;
	Fri, 8 Feb 2002 23:54:15 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id XAA24264
	for <tsvwg@optimus.ietf.org>; Fri, 8 Feb 2002 23:54:10 -0500 (EST)
Received: from d12lmsgate-2.de.ibm.com (d12lmsgate-2.de.ibm.com [195.212.91.200])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA03459
	for <tsvwg@ietf.org>; Fri, 8 Feb 2002 23:54:05 -0500 (EST)
Received: from d12relay01.de.ibm.com (d12relay01.de.ibm.com [9.165.215.22])
	by d12lmsgate-2.de.ibm.com (1.0.0) with ESMTP id FAA62032;
	Sat, 9 Feb 2002 05:53:37 +0100
Received: from d10hubm1.telaviv.ibm.com (d10ml001.telaviv.ibm.com [9.148.216.55])
	by d12relay01.de.ibm.com (8.11.1m3/NCO v5.01) with ESMTP id g194t5f30122;
	Sat, 9 Feb 2002 05:55:06 +0100
To: Stephen Bailey <steph@cs.uchicago.edu>
Cc: rdma@yahoogroups.com, tsvwg@ietf.org
MIME-Version: 1.0
X-Mailer: Lotus Notes Release 5.0.9  November 16, 2001
From: "Julian Satran" <Julian_Satran@il.ibm.com>
Message-ID: <OF9AFCF7D9.64AD9B34-ONC2256B5B.000B7873@telaviv.ibm.com>
Date: Sat, 9 Feb 2002 06:55:00 +0200
X-MIMETrack: Serialize by Router on D10ML001/10/M/IBM(Release 5.0.8 |June 18, 2001) at
 09/02/2002 06:55:09,
	Serialize complete at 09/02/2002 06:55:09
Content-Type: multipart/alternative; boundary="=_alternative 000C2B07C2256B5B_="
Subject: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

This is a multipart message in MIME format.
--=_alternative 000C2B07C2256B5B_=
Content-Type: text/plain; charset="us-ascii"

Steph,

The mathematics behind the proposal are simple. You have two random 
numbers "against" a finite-length PDU.
In you original proposal you have single random number (the marker) 
against a quasi infinite stream. The chance to miss is 1.
With any decent digest (e.g. a CRC) you have a small (2**-64) chance to 
miss per PDU word (for every new PDU you have a new pattern).  All the 
tricks like alignment etc do not decrease the cardinality of stream - 
i.e., it is still infinite and then a constant framer is no good.
The cost of the new scheme is also tolerable.


Julo




Stephen Bailey <steph@cs.uchicago.edu>
08-02-02 05:17

 
        To:     Julian Satran/Haifa/IBM@IBMIL
        cc:     tsvwg@ietf.org, rdma@yahoogroups.com
        Subject:        Re: [rdma] a proposal for a different No-Data-Touch framer for TCP

 

Julian,

I was also initially excited by your proposal, but I've come to a
similar conclusion to Paul.

The key question is what property is probability in the framing
mechanism providing.  I claim it's for safety.

A safety analysis should assume a worst possible data stream.  But not
worst `impossible'---bits that are knowable are knowable, and bits
that are unknowable can't be used to construct a worst possible
stream).

I do not think your proposal provides better safety than the current
TUF.  In fact, it might be worse.  I agree that both are difficult to
analyze, in any case.

Bad data streams in your proposal would be something like groups of 3
words, with an incrementing first word (the supposed marker), a 0
salt and the corresponding digest.  If you got clever, I bet
you could even interleave so that the digest from the first candidate
is is the marker for the second.  Really clever would be to use the
salt of the first as the marker for the second.  I imagine that a
`good digest' would have the property that you could cover all the
marker possibilities this way in only 2^32 words.

As far as I can tell, ceasing PDU containment validation attempts long
before you've checked many false candidates is of primary importance
to maximizing safety in EITHER approach.  I don't buy your claim that
safety is not a function of the number of tests of false candidates in
your approach.  The probability that must be controlled in either
case, is the chance of actually testing a false candidate.

Steph

> Paul,
> 
> The salt is a random number that is changing on every PDU. That changes
> radically the false detection probability in face on an infinite stream
> form certainty to a value that is acceptable from engineering point of
> view. Neither protects against malicious user. However the probability 
of
> random good framer in the data becomes nil.
> 
> Julo



--=_alternative 000C2B07C2256B5B_=
Content-Type: text/html; charset="us-ascii"


<br><font size=2 face="sans-serif">Steph,</font>
<br>
<br><font size=2 face="sans-serif">The mathematics behind the proposal are simple. You have two random numbers &quot;against&quot; a finite-length PDU.</font>
<br><font size=2 face="sans-serif">In you original proposal you have single random number (the marker) against a quasi infinite stream. The chance to miss is 1.</font>
<br><font size=2 face="sans-serif">With any decent digest (e.g. a CRC) you have a small (2**-64) chance to miss per PDU word (for every new PDU you have a new pattern). &nbsp;All the tricks like alignment etc do not decrease the cardinality of stream - i.e., it is still infinite and then a constant framer is no good.</font>
<br><font size=2 face="sans-serif">The cost of the new scheme is also tolerable.</font>
<br>
<br>
<br><font size=2 face="sans-serif">Julo</font>
<br>
<br>
<br>
<table width=100%>
<tr valign=top>
<td>
<td><font size=1 face="sans-serif"><b>Stephen Bailey &lt;steph@cs.uchicago.edu&gt;</b></font>
<p><font size=1 face="sans-serif">08-02-02 05:17</font>
<br>
<td><font size=1 face="Arial">&nbsp; &nbsp; &nbsp; &nbsp; </font>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; To: &nbsp; &nbsp; &nbsp; &nbsp;Julian Satran/Haifa/IBM@IBMIL</font>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; cc: &nbsp; &nbsp; &nbsp; &nbsp;tsvwg@ietf.org, rdma@yahoogroups.com</font>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; Subject: &nbsp; &nbsp; &nbsp; &nbsp;Re: [rdma] a proposal for a different No-Data-Touch framer for TCP</font>
<br>
<br><font size=1 face="Arial">&nbsp; &nbsp; &nbsp; &nbsp;</font></table>
<br>
<br><font size=2 face="Courier New">Julian,<br>
<br>
I was also initially excited by your proposal, but I've come to a<br>
similar conclusion to Paul.<br>
<br>
The key question is what property is probability in the framing<br>
mechanism providing. &nbsp;I claim it's for safety.<br>
<br>
A safety analysis should assume a worst possible data stream. &nbsp;But not<br>
worst `impossible'---bits that are knowable are knowable, and bits<br>
that are unknowable can't be used to construct a worst possible<br>
stream).<br>
<br>
I do not think your proposal provides better safety than the current<br>
TUF. &nbsp;In fact, it might be worse. &nbsp;I agree that both are difficult to<br>
analyze, in any case.<br>
<br>
Bad data streams in your proposal would be something like groups of 3<br>
words, with an incrementing first word (the supposed marker), a 0<br>
salt and the corresponding digest. &nbsp;If you got clever, I bet<br>
you could even interleave so that the digest from the first candidate<br>
is is the marker for the second. &nbsp;Really clever would be to use the<br>
salt of the first as the marker for the second. &nbsp;I imagine that a<br>
`good digest' would have the property that you could cover all the<br>
marker possibilities this way in only 2^32 words.<br>
<br>
As far as I can tell, ceasing PDU containment validation attempts long<br>
before you've checked many false candidates is of primary importance<br>
to maximizing safety in EITHER approach. &nbsp;I don't buy your claim that<br>
safety is not a function of the number of tests of false candidates in<br>
your approach. &nbsp;The probability that must be controlled in either<br>
case, is the chance of actually testing a false candidate.<br>
<br>
Steph<br>
<br>
&gt; Paul,<br>
&gt; <br>
&gt; The salt is a random number that is changing on every PDU. That changes<br>
&gt; radically the false detection probability in face on an infinite stream<br>
&gt; form certainty to a value that is acceptable from engineering point of<br>
&gt; view. Neither protects against malicious user. However the probability of<br>
&gt; random good framer in the data becomes nil.<br>
&gt; <br>
&gt; Julo<br>
</font>
<br>
<br>
--=_alternative 000C2B07C2256B5B_=--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Feb  9 10:50:00 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA16499
	for <tsvwg-archive@odin.ietf.org>; Sat, 9 Feb 2002 10:50:00 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA21833
	for tsvwg-archive@odin.ietf.org; Sat, 9 Feb 2002 10:50:02 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA21672;
	Sat, 9 Feb 2002 10:37:04 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA21643
	for <tsvwg@optimus.ietf.org>; Sat, 9 Feb 2002 10:37:02 -0500 (EST)
Received: from ugetmore4less.net (smtp.ugetmore4less.net [64.55.184.155])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA16248
	for <tsvwg@ietf.org>; Sat, 9 Feb 2002 10:36:59 -0500 (EST)
Received: from 64.55.184.155 [64.55.184.155] by ugetmore4less.net
  (SMTPD32-7.05 EVAL) id A1CCDD300FE; Sat, 09 Feb 2002 10:35:40 -0500
From: "Dollar Mountain" <worldofdollars@ugetmore4less.net>
To: tsvwg@ietf.org
Date: Sat, 09 Feb 2002 10:35:41 -0500
X-Mailer: Mach5 Mailer-2.50 PID{6eaffa5e-7932-4682-9044-b0fc00316c94} RC{R01122870742}
MIME-Version: 1.0
Content-Type: multipart/mixed;
	boundary="----000000000000000000000"
Message-Id: <20020209103578.SM01448@64.55.184.155>
Subject: [Tsvwg] It's a simple and easy-to-follow plan and EVERY website will want this service
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

------000000000000000000000
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit

<HTML>
<HEAD>
<META NAME="GENERATOR" Content="Microsoft Visual Studio 6.0">
<TITLE></TITLE>
</HEAD>
<BODY>

<P><!-- $Revision: 1.3 $ $Date: 2001/11/28 19:58:54 $ -->&nbsp; 


<TABLE cellSpacing=0 cellPadding=0 width="100%" border=0>
  
  <TR>
    <TD align=left bgColor=white colSpan=4><a href="http://www.trafficwow.net"><IMG alt="" src="http://www.trafficwow.net/trafficwow.jpg" border=0></a>  </TD></TR></TABLE>
<TABLE cellSpacing=0 cellPadding=0 width="100%" border=0>
  
  <TR>
    <TD width="100%" 
    background=http://www.trafficwow.net/nonav.gif>&nbsp;</TD></TR></TABLE><!--/end new header--><BR>
<TABLE cellSpacing=0 cellPadding=6 width=500 bgColor=#ffffff border=0>
  
  <TR>
    <TD align=left><IMG height=1 alt="" 
      src="http://www.trafficwow.net/spacer_white.gif" width=4 align=left 
      border=0></TD>
    <TD>
      <P><!--DATE-->
      <P align=right><B>TrafficWoW Daily Newsletter February 09, 2002</B> <!--/END DATE--><!--TOP STORY-->
      <P>Disclaimer: ----------- This email is NEVER sent unsolicited. THIS IS 
      NOT SPAM. You are receiving this because you have either answered 
      classified ad, posted to our FFA pages, are on the same opt-in list or 
      have sent me something in the past. If you'd like to be removed, see 
      instructions at the bottom of this message. </P>
      <P>=========================== <STRONG><EM>IN THIS ISSUE</EM></STRONG> 
      =========================== </P>
      <P>
      <TABLE cellPadding=5 width="100%">
        
        <TR>
          <TD class=text2 align=middle><STRONG 
            ><FONT color=#000080 size=4 
            ><U>It's a simple and 
            easy-to-follow plan and EVERY website will want this 
            service.</U></FONT></STRONG></TD>
          <TD>&nbsp;</TD></TR>
        <TR>
          <TD class=text align=middle colSpan=3>
            <DIV><FONT face=Arial color=#808080 
            ><STRONG 
            ></STRONG></FONT>&nbsp;</DIV>
            <DIV><FONT face=Arial 
            ><STRONG>Can you&nbsp;spend 
            just 10 minutes a day sending e-mails to ANY websites on the 
            Internet?</STRONG></FONT></DIV>
            <DIV><FONT face=Arial 
            ><STRONG>iF&nbsp;so&nbsp;this 
            <FONT size=4>UNIQUE</FONT> program will make you 
            a very good income every day!!!<BR 
            ></STRONG></FONT></DIV></TD></TR>
        <TR>
          <TD class=text align=middle colSpan=3><FONT 
            face=Arial color=#808080><STRONG 
            ><U><FONT color=#000080 
            >Its only costs $5 to start up which we refund 
            back to you within 24 hours making this 100% 
            FREE.....</FONT></U></STRONG> <STRONG><U 
            ><FONT color=#0000ff><BR 
            ><BR 
            ></FONT></U></STRONG></FONT></TD></TR>
        <TR>
          <TD class=text align=left colSpan=3><FONT 
            face=Arial color=#000080><STRONG 
            ><U>Here are the BENEFITS to 
            YOU.</U></STRONG></FONT> 
            <UL>
              <LI><FONT face=Arial 
              ><STRONG>Your VERY OWN 
              independent on-line business </STRONG></FONT>
              <LI><FONT face=Arial 
              ><STRONG>A steady stream of 
              $$$$ paid daily to you. </STRONG></FONT>
              <LI><FONT face=Arial 
              ><STRONG>No sales team, 
              equipment or office required. </STRONG></FONT>
              <LI><FONT face=Arial 
              ><STRONG>Full step by step 
              instructions to get you making money today. </STRONG></FONT>
              <LI><FONT face=Arial 
              ><STRONG>No hidden 
              expenses, nothing to download or advertise. </STRONG></FONT>
              <LI><FONT face=Arial 
              ><STRONG>Multiple income 
              streams possible from day one. </STRONG></FONT>
              <LI><FONT face=Arial 
              ><STRONG>Money deposited 
              directly and immediately to YOUR account </STRONG></FONT>
              <LI><FONT face=Arial 
              ><STRONG>Fully automated 
              setup.</STRONG> </FONT></LI></UL>
            <DIV><FONT face=Arial 
            ><STRONG>You can wake up 
            tomorrow morning open your Email and find </STRONG></FONT><FONT 
            face=Arial><STRONG><SPAN 
            class=headingwhite>'Notification for Payment 
            Received" </SPAN>staring back at you. </STRONG></FONT></DIV>
            <DIV><FONT face=Arial 
            ><STRONG>I <I 
            >never</I> get tired of seeing 
            that:))</STRONG></FONT></DIV>
            <DIV><FONT face=Arial 
            ></FONT>&nbsp;</DIV>
            <DIV>
            <TABLE cellPadding=5 width="100%">
              
              <TR>
                <TD class=text align=middle colSpan=3>
                  <DIV align=left><FONT face=Arial 
                  ><A href="https://www.paypal.com/affil/pal=worldofdollars@ugetmore4less.net"><FONT 
                  color=#000080><STRONG 
                  ><U>Sign up 
                  today</A></U></STRONG> -</FONT> <BR><BR 
                  ><STRONG>Press the 
                  link&nbsp;below and send just $5 and you will receive all the 
                  help and infomation you require by e-mail 
                  instantly.</STRONG></FONT></DIV>
                  <DIV align=left><FONT face=Arial 
                  color=#ff0000><STRONG 
                  >We will also send you $5 back just for 
                  signing up a new account so this is 100% 
                  FREE.</STRONG></FONT></DIV></TD></TR>
              <TR>
                <TD class=textr align=middle colSpan=3>
                  <P><STRONG>&nbsp;<FONT color=#000080 
                  >PRESS</FONT></STRONG><FONT color=#000080 
                  > </FONT><A href="https://www.paypal.com/affil/pal=worldofdollars@ugetmore4less.net"><FONT 
                  color=#000080><STRONG 
                  ><FONT size=5 
                  >HERE</FONT></STRONG> </FONT></A><FONT 
                  color=#000080>&nbsp;<STRONG 
                  >TO CONTINUE</STRONG> <BR 
                  ></FONT><BR><STRONG 
                  >FULL 30 DAY MONEY BACK GUARANTEE.<BR 
                  ></STRONG></P>
                  </TD></TR></TABLE></DIV></TD></TR></TABLE>====================================================================== 
     </P>
      <P align=left>Further Transmission to you by the sender of this e-mail may 
      be stopped at no cost to you by clicking here <A 
      href="http://www.trafficwow.net/removeusers.asp?Emails=tsvwg@ietf.org">http://www.trafficwow.net/removeusers.asp?Emails=tsvwg@ietf.org</A></P>
      <P align=left>If your click fails, then copy and paste 
      the link below:<BR><A 
      href="http://www.trafficwow.net/removeusers.asp?Emails=tsvwg@ietf.org"><U>http://www.trafficwow.net/removeusers.asp?Emails=tsvwg@ietf.org</U></A><BR> 
       and paste it into your browser. </P>
      <P align=left><U>Do not reply</U>   to this email for the removal process. The only 
      way to be removed from our mailing lists is to click on the removal link 
      above, else your remove request will fail.</P>
      <P 
      align=left>======================================================================<SPAN 
      class=mainfeatarttxt></P>
      <P>
      <HR width="100%" SIZE=1>

      <P></P>
      
      <p></p>


      <P align=center><SPAN class=mainfeatarttxt><STRONG>© 2002 TrafficWoW.net</STRONG>  </SPAN></P><!--/END BOTTOM BANNER ADVERTISEMENT--></SPAN></TD></TR></TABLE></P>

</BODY>
</HTML>

------000000000000000000000--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Feb  9 11:53:07 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA17248
	for <tsvwg-archive@odin.ietf.org>; Sat, 9 Feb 2002 11:53:07 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA23660
	for tsvwg-archive@odin.ietf.org; Sat, 9 Feb 2002 11:53:11 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA23420;
	Sat, 9 Feb 2002 11:35:52 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA23396
	for <tsvwg@optimus.ietf.org>; Sat, 9 Feb 2002 11:35:51 -0500 (EST)
Received: from sandmail.sandburst.com (sandmail.sandburst.com [216.57.132.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA17130
	for <tsvwg@ietf.org>; Sat, 9 Feb 2002 11:35:47 -0500 (EST)
To: "Julian Satran" <Julian_Satran@il.ibm.com>
Cc: rdma@yahoogroups.com, tsvwg@ietf.org
In-Reply-To: Message from "Julian Satran" <Julian_Satran@il.ibm.com> 
   of "Sat, 09 Feb 2002 06:55:00 +0200." <OF9AFCF7D9.64AD9B34-ONC2256B5B.000B7873@telaviv.ibm.com> 
References: <OF9AFCF7D9.64AD9B34-ONC2256B5B.000B7873@telaviv.ibm.com> 
Date: Sat, 09 Feb 2002 11:35:17 -0500
From: Stephen Bailey <steph@cs.uchicago.edu>
Message-Id: <20020209163519.48BE24E95@sandmail.sandburst.com>
Subject: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

For a proper safety analysis, you can not assume that the the contents
of the the stream (or the PDU) are randomly distributed.  You have to
assume the stream has worst-case contents.  Defining `worst case' is
the hard part.

Steph

> The mathematics behind the proposal are simple. You have two random 
> numbers "against" a finite-length PDU.
> In you original proposal you have single random number (the marker) 
> against a quasi infinite stream. The chance to miss is 1.
> With any decent digest (e.g. a CRC) you have a small (2**-64) chance to 
> miss per PDU word (for every new PDU you have a new pattern).  All the 
> tricks like alignment etc do not decrease the cardinality of stream - 
> i.e., it is still infinite and then a constant framer is no good.
> The cost of the new scheme is also tolerable.

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Feb  9 13:03:27 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA18334
	for <tsvwg-archive@odin.ietf.org>; Sat, 9 Feb 2002 13:03:26 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA27459
	for tsvwg-archive@odin.ietf.org; Sat, 9 Feb 2002 13:03:28 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA26755;
	Sat, 9 Feb 2002 12:51:52 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA26725
	for <tsvwg@optimus.ietf.org>; Sat, 9 Feb 2002 12:51:50 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-200.dallas.net [209.44.42.200])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA18231
	for <tsvwg@ietf.org>; Sat, 9 Feb 2002 12:51:46 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id LAA19998;
	Sat, 9 Feb 2002 11:51:40 -0600
Date: Sat, 9 Feb 2002 11:51:40 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Dollar Mountain <worldofdollars@ugetmore4less.net>
Cc: tsvwg@ietf.org
Subject: Re: [Tsvwg] It's a simple and easy-to-follow plan and EVERY website will want this service
Message-ID: <20020209115140.A9875@openss7.org>
Reply-To: bidulock@openss7.org
References: <20020209103578.SM01448@64.55.184.155>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <20020209103578.SM01448@64.55.184.155>; from worldofdollars@ugetmore4less.net on Sat, Feb 09, 2002 at 10:35:41AM -0500
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Dollar,

On Sat, 09 Feb 2002, Dollar Mountain wrote:

> 
>    
>    
>                             TrafficWoW Daily Newsletter February 09, 2002
>    
>    Disclaimer: ----------- This email is NEVER sent unsolicited. THIS IS
>    NOT SPAM. You are receiving this because you have either answered
>    classified ad, posted to our FFA pages, are on the same opt-in list or
>    have sent me something in the past. If you'd like to be removed, see
>    instructions at the bottom of this message.

You couldn't be more wrong....  Stop attempting to break the
law with impunity.  It can easily be proved here that these
statements above are false.

--brian


>    
>    =========================== IN THIS ISSUE ===========================
>    
>    It's a simple and easy-to-follow plan and EVERY website will want this
>    service.
>    
>    Can you spend just 10 minutes a day sending e-mails to ANY websites on
>    the Internet?
>    iF so this UNIQUE program will make you a very good income every
>    day!!!
>    Its only costs $5 to start up which we refund back to you within 24
>    hours making this 100% FREE..... 
>    Here are the BENEFITS to YOU.
>      * Your VERY OWN independent on-line business 
>      * A steady stream of $$$$ paid daily to you. 
>      * No sales team, equipment or office required. 
>      * Full step by step instructions to get you making money today. 
>      * No hidden expenses, nothing to download or advertise. 
>      * Multiple income streams possible from day one. 
>      * Money deposited directly and immediately to YOUR account 
>      * Fully automated setup.
>        
>    You can wake up tomorrow morning open your Email and find
>    'Notification for Payment Received" staring back at you. 
>    I never get tired of seeing that:))
>    
>    [1]Sign up today -
>    Press the link below and send just $5 and you will receive all the
>    help and infomation you require by e-mail instantly.
>    We will also send you $5 back just for signing up a new account so
>    this is 100% FREE.
>    
>     PRESS [2]HERE  TO CONTINUE
>    FULL 30 DAY MONEY BACK GUARANTEE.
>    ======================================================================
>    
>    Further Transmission to you by the sender of this e-mail may be
>    stopped at no cost to you by clicking here
>    [3]http://www.trafficwow.net/removeusers.asp?Emails=tsvwg@ietf.org
>    
>    If your click fails, then copy and paste the link below:
>    [4]http://www.trafficwow.net/removeusers.asp?Emails=tsvwg@ietf.org
>    and paste it into your browser.
>    
>    Do not reply to this email for the removal process. The only way to be
>    removed from our mailing lists is to click on the removal link above,
>    else your remove request will fail.
>    
>    ======================================================================
>    ______________________________________________________________________
>    
>                            © 2002 TrafficWoW.net
> 
> References
> 
>    Visible links
>    1. https://www.paypal.com/affil/pal=worldofdollars@ugetmore4less.net
>    2. https://www.paypal.com/affil/pal=worldofdollars@ugetmore4less.net
>    3. http://www.trafficwow.net/removeusers.asp?Emails=tsvwg@ietf.org
>    4. http://www.trafficwow.net/removeusers.asp?Emails=tsvwg@ietf.org
> 
>    Hidden links:
>    5. http://www.trafficwow.net/


-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Feb  9 20:25:44 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA21293
	for <tsvwg-archive@odin.ietf.org>; Sat, 9 Feb 2002 20:25:44 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA10103
	for tsvwg-archive@odin.ietf.org; Sat, 9 Feb 2002 20:25:46 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA09793;
	Sat, 9 Feb 2002 20:06:33 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA09762
	for <tsvwg@optimus.ietf.org>; Sat, 9 Feb 2002 20:06:31 -0500 (EST)
Received: from e31.co.us.ibm.com (e31.co.us.ibm.com [32.97.110.129])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA21149
	for <tsvwg@ietf.org>; Sat, 9 Feb 2002 20:06:28 -0500 (EST)
Received: from westrelay02.boulder.ibm.com (westrelay02.boulder.ibm.com [9.99.140.23])
	by e31.co.us.ibm.com (8.9.3/8.9.3) with ESMTP id UAA72276;
	Sat, 9 Feb 2002 20:03:20 -0500
Received: from d03nm014.boulder.ibm.com (avpilot.boulder.ibm.com [9.17.188.135])
	by westrelay02.boulder.ibm.com (8.11.1m3/NCO v5.01) with ESMTP id g1A16SS45034;
	Sat, 9 Feb 2002 18:06:28 -0700
X-Priority: 1 (High)
Importance: Normal
To: Stephen Bailey <steph@cs.uchicago.edu>,
        "Julian Satran" <Julian_Satran@il.ibm.com>
Cc: tsvwg@ietf.org
X-Mailer: Lotus Notes Release 5.0.3 (Intl) 21 March 2000
Message-ID: <OF31443FAC.EDF1393F-ON88256B5B.006A9AA4@boulder.ibm.com>
From: "John Hufferd" <hufferd@us.ibm.com>
Date: Sat, 9 Feb 2002 17:05:37 -0800
X-MIMETrack: Serialize by Router on D03NM014/03/M/IBM(Release 5.0.9 |November 16, 2001) at
 02/09/2002 06:06:28 PM
MIME-Version: 1.0
Content-type: text/plain; charset=us-ascii
Subject: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


Steph, Julian,
The following is an attempt to determine the probability of a false
positive, both with the TUF approach as defined in
draft-ietf-tsvwg-tcp-ulp-frame-01.txt, and the semi draft that Julian
proposed.

Since we are only interested in the case of False Positives given a
re-segmentation occurs, I think it is possible to determine which of the
two approaches is less risky.

Here are the two layouts:

Current TUF has a 64 bit header at the beginning of Segment:
1. Length of 16 bits (Must match up with the physical length of the
segment)
2. Key  48 bits (Key Randomly Chosen by ULP,  -- Known by sender and
receiver, and changed as some undefined interval, if at all)

Julian's Eye-catcher-Salt-CRC approach (which I will call ESC) is 96 bits
1. Eye-catcher - 32 bits (Chosen by the ULP,  could also be chosen like the
Key in TUF, known by both sides)
2. Random number - 32 bits (Generated by ULP, unique to this PDU only)
3. CRC  - 32 bits (Generated by ULP based on 1. & 2.)

Lets first examine the ESC.

Once the Eye-catcher is determined, the Random number is chosen and the CRC
is computed, but to the Framer this is just string of bits passed from the
ULP.  The Receiving Framer (deFramer?) does not know the actual magic
string of bits which it should look for within the incoming payload.
Therefore it needs to do a computation at the start of each segment to make
sure the approprate string is received.

Given any beginning Eye-catcher, randomly chosen or not, there are a number
of 64 bit strings which can be interpreted as a valid salt and CRC.  That
is, for every 32 bit Salt, there is a valid 32 bit CRC.  Therefore, there
is as many valid 64-bit strings, which can follow the Eye-catcher as there
are unique 32-bit Salt numbers.  This means, there are (2**32) valid 8 byte
strings that can follow the Eye-catcher.

So in the case of ESC we have:

 The probability "ex" that a 32-bit Random Eye-catcher (or Key) is
"incorrectly" (do to re-segmentation) located at the beginning of a segment
, and the probability "ey" that one of the 2**32 valid 8 byte strings will
follow that Eye-catcher.  I think that makes it the total probability equal
to:
ex times  ey.

We have "ex" equal to the probability of the Eye-catcher occurring in any
word times the number of chances that might happen.  This means the
probability of any 32 bit value occurring  is 1/(2**32) times 361  where
361 is the number of possible segment points that could have been made in a
1460 byte payload (with a total of 3 header words and one byte of data).
Therefore:

ex = (1/2**32) times 361
ex = 2.3e-10 * 361
or
ex = 8.41e-8

The value of "ey" is equal to the probability that one of the 2**32 valid 8
byte strings will follow the Eye-catcher.  This is 1/2**32 times the number
of chances.  In this case the number of chances equals the number of
possible segment points that could have been made in a 1460 byte payload
(with a total of 3 header words and one byte of data), which is 361.
Therefore:

ey = 2.3e-10 * 361
or
ey = 8.41e-8

Therefore with ESC  we have the total probability of false positive of:

ex*ey = 8.41e-8 * 8.41e-8
or
ex*ey = 7.06 e-15


Now we will examine  the case of the current TUF proposal:

We have a probability "tx" that a 48-bit Random Key (or Eye-catcher) is
"incorrectly" (do to re-segmentation) located at + 2 bytes from the
beginning of a segment , and the probably "ty" that a 16 bit field
preceding it, will be a valid length of the current Segment.

I think that probability is equal to  tx  times  ty.

Now "ty" has 1 chance of occurring in 2**16 -10 (there must be at lease 8
bytes of TUF header + one byte of data, and zero is not valid so subtract
10 from the number of valid lengths)
ty = 1/((2**16)-10)
ty = 1/65526
ty = 1.526e-5

"tx" is the probability of a specific 48 bit string occurring at +2 bytes
into each possible segmentation point.  This is the probability of the
string occurring times the number of possible chances it has to occur.
Since there can be at most 365 TCP payload words in a normal Ethernet Frame
((1500-20-20)/4 =1460/4 = 365 words) then that means there are 365-3 or 362
possible segment point, which could leave at least a 2 word header and 1
byte of data.
Therefore:
tx = (1/2**48) times 362
tx = ((3.55e-15) * 362
tx = 1.286e-12).

Therefore with Current TUF we have the total probability of false positive
of:

tx*ty = 1.286e-12 * 1.526e-5
or
tx =  1.962e-17

Therefore, based on the calculations above, I believe that the Current TUF
approach offers a lower probability of False positive then does the
straight version of ESC.

But remember, Current TUF has a 64 bit header and ESC has a 96 bit header.
If we were to extend the TUF proposal to 96 bits the calculation would look
like:

ty = 1.526e-5 (Same as before)

AND,

tx =l (1/2**80) * 361
tx = 8.27e-25 * 361
tx = 2.986e-22

and
tx * ty = 2.986e-22 * 1.526e-5
tx * ty = 4.557e-27

However, ECS can also use a PDU header as part of the total salt.  The PDU
is not a fully random set of bits, but it does also reduce the probability
of False Positive a great deal more.  (I will let someone else compute that
probability.) The current probability of a false positive with a TUF's two
word header is 1.962e-17, but I think it is safe to guess that the ECS when
factoring in a PDU, would lower its probability a great deal lower then
7.06 e-15 (probably lower then current TUF), and it would then only need a
two word header like TUF.

Summary:
If you compare Current TUF (which has 2 word headers) to ECS (which has 3
word headers) you have a lower probability of False Positives with TUF.  If
you compare Current TUF with ECS that uses the CRC with the Header of the
PDU (so ECS only has a 2 byte Header) there is reason to believe that ECS
would have the lower probability of false positive.

So the questions should be: is  1.962e-17 a low enough probability of false
positives that we can get on with the Framing per the current TUFs
proposal?  Or do we need to add a third word to TUFs, or use the ECS with
PDU CRC.


 .
.
 .
John L. Hufferd
Senior Technical Staff Member (STSM)
IBM/SSG San Jose Ca
Main Office (408) 256-0403, Tie: 276-0403,  eFax: (408) 904-4688
Home Office (408) 997-6136, Cell: (408) 499-9702
Internet address: hufferd@us.ibm.com


Stephen Bailey <steph@cs.uchicago.edu> on 02/09/2002 08:35:17 AM

To:    Julian Satran/Haifa/IBM@IBMIL
cc:    rdma@yahoogroups.com, tsvwg@ietf.org
Subject:    Re: [rdma] a proposal for a different No-Data-Touch framer for
       TCP



For a proper safety analysis, you can not assume that the the contents
of the the stream (or the PDU) are randomly distributed.  You have to
assume the stream has worst-case contents.  Defining `worst case' is
the hard part.

Steph

> The mathematics behind the proposal are simple. You have two random
> numbers "against" a finite-length PDU.
> In you original proposal you have single random number (the marker)
> against a quasi infinite stream. The chance to miss is 1.
> With any decent digest (e.g. a CRC) you have a small (2**-64) chance to
> miss per PDU word (for every new PDU you have a new pattern).  All the
> tricks like alignment etc do not decrease the cardinality of stream -
> i.e., it is still infinite and then a constant framer is no good.
> The cost of the new scheme is also tolerable.

------------------------ Yahoo! Groups Sponsor ---------------------~-->
Sponsored by VeriSign - The Value of Trust
Pinpoint the right security solution for your company - FREE
Guide from industry leader VeriSign gives you all the facts.
http://us.click.yahoo.com/pCuuSA/WdiDAA/yigFAA/W6uqlB/TM
---------------------------------------------------------------------~->

To unsubscribe from this group, send an email to:
rdma-unsubscribe@egroups.com



Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/






_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sun Feb 10 14:37:08 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA09760
	for <tsvwg-archive@odin.ietf.org>; Sun, 10 Feb 2002 14:37:08 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA21880
	for tsvwg-archive@odin.ietf.org; Sun, 10 Feb 2002 14:37:11 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA21316;
	Sun, 10 Feb 2002 14:21:27 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA21285
	for <tsvwg@optimus.ietf.org>; Sun, 10 Feb 2002 14:21:25 -0500 (EST)
Received: from d12lmsgate-3.de.ibm.com (d12lmsgate-3.de.ibm.com [195.212.91.201])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA09608
	for <tsvwg@ietf.org>; Sun, 10 Feb 2002 14:21:21 -0500 (EST)
Received: from d12relay01.de.ibm.com (d12relay01.de.ibm.com [9.165.215.22])
	by d12lmsgate-3.de.ibm.com (1.0.0) with ESMTP id UAA51242;
	Sun, 10 Feb 2002 20:20:52 +0100
Received: from d10hubm1.telaviv.ibm.com (d10ml001.telaviv.ibm.com [9.148.216.55])
	by d12relay01.de.ibm.com (8.11.1m3/NCO v5.01) with ESMTP id g1AJMLm93002;
	Sun, 10 Feb 2002 20:22:22 +0100
To: "Leonid Grossman" <leonidg@attbi.com>
Cc: rdma@yahoogroups.com, tsvwg@ietf.org
MIME-Version: 1.0
X-Mailer: Lotus Notes Release 5.0.9  November 16, 2001
From: "Julian Satran" <Julian_Satran@il.ibm.com>
Message-ID: <OFE707DB63.E1CAA37B-ONC2256B5C.00686473@telaviv.ibm.com>
Date: Sun, 10 Feb 2002 21:22:22 +0200
X-MIMETrack: Serialize by Router on D10ML001/10/M/IBM(Release 5.0.8 |June 18, 2001) at
 10/02/2002 21:22:25,
	Serialize complete at 10/02/2002 21:22:25
Content-Type: multipart/alternative; boundary="=_alternative 00694576C2256B5C_="
Subject: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

This is a multipart message in MIME format.
--=_alternative 00694576C2256B5C_=
Content-Type: text/plain; charset="us-ascii"

Leonid,

The probability is not 0 - that is true. But for the form used in the ULP 
framing draft it is in theory 1 and that is unacceptable.
The new proposal is aimed at macing it a very low (acceptable number) - 
equivalent with the probability of an undetected error by a CRC (in fact 
substantially less).

As for directing DDP traffic to a specific port that is an interesting 
idea (although not related to the issue at hand).
Even in a separate ring you are still faced with the same probability as 
long as the stream is infinite. The gain that you can obtain is the 
coexistence of DDP streams with "non-steered" streams and you pay by 
maintaining a more complex stack.


Julo




"Leonid Grossman" <leonidg@attbi.com>
09-02-02 19:22

 
        To:     <satran@haifa.vnet.ibm.com>
        cc: 
        Subject:        Re: [rdma] a proposal for a different No-Data-Touch framer for TCP

 

Julian - either proposal has a non-zero probability to mistaken TCP
payload in non-DDP frame for a valid framing (and potentially cause 
a data corruption).
I'm deciding whether to put some framing and DDP on TCP support in our
upcoming 10GigE NIC ASIC, and to me it looks like a simpler and 
arguably more reliable way to recognize DDP frames would be to just 
use a dedicated port number for DDP.
For outgoing frames, DDP shim can replace application port number 
with DDP port number, and store the original port in iWARP/DDPP 
header somewhere.
On receive, NIC hardware would use DDP port number to recognize DDP
traffic and steer it into a separate receive ring (with protocol and
DDPP headers separated into pre-allocated buffers); 
then the hardware would verify iWAPR/DDPP header, do map tag into
physical address of the locked DDP buffers and DMA DDP payload.
SDP shim on receive would steer the traffic based upon the original 
port number.
Anyway, identifying DDP traffic in TCP header rather than in TCP payload 
looks like a better approach.

Any comments?

Thanks in advance,
Leonid Grossman
Software Engineering
S2IO Technologies
Cupertino, California




> The mathematics behind the proposal are simple. You have two random 
> numbers "against" a finite-length PDU.
> In you original proposal you have single random number (the marker) 
> against a quasi infinite stream. The chance to miss is 1.
> With any decent digest (e.g. a CRC) you have a small (2**-64) chance to 
> miss per PDU word (for every new PDU you have a new pattern). All the 
> tricks like alignment etc do not decrease the cardinality of stream - 
> i.e., it is still infinite and then a constant framer is no good.
> The cost of the new scheme is also tolerable.




--=_alternative 00694576C2256B5C_=
Content-Type: text/html; charset="us-ascii"


<br><font size=2 face="sans-serif">Leonid,</font>
<br>
<br><font size=2 face="sans-serif">The probability is not 0 - that is true. But for the form used in the ULP framing draft it is in theory 1 and that is unacceptable.</font>
<br><font size=2 face="sans-serif">The new proposal is aimed at macing it a very low (acceptable number) - equivalent with the probability of an undetected error by a CRC (in fact substantially less).</font>
<br>
<br><font size=2 face="sans-serif">As for directing DDP traffic to a specific port that is an interesting idea (although not related to the issue at hand).</font>
<br><font size=2 face="sans-serif">Even in a separate ring you are still faced with the same probability as long as the stream is infinite. The gain that you can obtain is the coexistence of DDP streams with &quot;non-steered&quot; streams and you pay by maintaining a more complex stack.</font>
<br>
<br>
<br><font size=2 face="sans-serif">Julo</font>
<br>
<br>
<br>
<table width=100%>
<tr valign=top>
<td>
<td><font size=1 face="sans-serif"><b>&quot;Leonid Grossman&quot; &lt;leonidg@attbi.com&gt;</b></font>
<p><font size=1 face="sans-serif">09-02-02 19:22</font>
<br>
<td><font size=1 face="Arial">&nbsp; &nbsp; &nbsp; &nbsp; </font>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; To: &nbsp; &nbsp; &nbsp; &nbsp;&lt;satran@haifa.vnet.ibm.com&gt;</font>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; cc: &nbsp; &nbsp; &nbsp; &nbsp;</font>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; Subject: &nbsp; &nbsp; &nbsp; &nbsp;Re: [rdma] a proposal for a different No-Data-Touch framer for TCP</font>
<br>
<br><font size=1 face="Arial">&nbsp; &nbsp; &nbsp; &nbsp;</font></table>
<br>
<br><font size=2 face="Courier New">Julian - either proposal has a non-zero probability to mistaken TCP<br>
payload in non-DDP frame for a valid framing (and potentially cause <br>
a data corruption).<br>
I'm deciding whether to put some framing and DDP on TCP support in our<br>
upcoming 10GigE NIC ASIC, and to me it looks like a simpler and <br>
arguably more reliable way to recognize DDP frames would be to just <br>
use a dedicated port number for DDP.<br>
For outgoing frames, DDP shim can replace application port number <br>
with DDP port number, and store the original port in iWARP/DDPP <br>
header somewhere.<br>
On receive, NIC hardware would use DDP port number to recognize DDP<br>
traffic and steer it into a separate receive ring (with protocol and<br>
DDPP headers separated into pre-allocated buffers); <br>
then the hardware would verify iWAPR/DDPP header, do map tag into<br>
physical address of the locked DDP buffers and DMA DDP payload.<br>
SDP shim on receive would steer the traffic based upon the original <br>
port number.<br>
Anyway, identifying DDP traffic in TCP header rather than in TCP payload <br>
looks like a better approach.<br>
<br>
Any comments?<br>
<br>
Thanks in advance,<br>
Leonid Grossman<br>
Software Engineering<br>
S2IO Technologies<br>
Cupertino, California<br>
<br>
<br>
<br>
<br>
&gt; The mathematics behind the proposal are simple. You have two random <br>
&gt; numbers &quot;against&quot; a finite-length PDU.<br>
&gt; In you original proposal you have single random number (the marker) <br>
&gt; against a quasi infinite stream. The chance to miss is 1.<br>
&gt; With any decent digest (e.g. a CRC) you have a small (2**-64) chance to <br>
&gt; miss per PDU word (for every new PDU you have a new pattern). All the <br>
&gt; tricks like alignment etc do not decrease the cardinality of stream - <br>
&gt; i.e., it is still infinite and then a constant framer is no good.<br>
&gt; The cost of the new scheme is also tolerable.<br>
<br>
</font>
<br>
<br>
--=_alternative 00694576C2256B5C_=--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sun Feb 10 20:44:42 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA13275
	for <tsvwg-archive@odin.ietf.org>; Sun, 10 Feb 2002 20:44:41 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA02385
	for tsvwg-archive@odin.ietf.org; Sun, 10 Feb 2002 20:44:44 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA01605;
	Sun, 10 Feb 2002 20:23:39 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA01576
	for <tsvwg@optimus.ietf.org>; Sun, 10 Feb 2002 20:23:35 -0500 (EST)
Received: from e31.co.us.ibm.com (e31.co.us.ibm.com [32.97.110.129])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA12940
	for <tsvwg@ietf.org>; Sun, 10 Feb 2002 20:23:32 -0500 (EST)
Received: from westrelay02.boulder.ibm.com (westrelay02.boulder.ibm.com [9.99.140.23])
	by e31.co.us.ibm.com (8.9.3/8.9.3) with ESMTP id UAA42612;
	Sun, 10 Feb 2002 20:20:20 -0500
Received: from d03nm014.boulder.ibm.com (avpilot.boulder.ibm.com [9.17.188.135])
	by westrelay02.boulder.ibm.com (8.11.1m3/NCO/VER6.00) with ESMTP id g1B1NTu129588;
	Sun, 10 Feb 2002 18:23:29 -0700
Importance: Normal
To: Stephen Bailey <steph@cs.uchicago.edu>,
        "Julian Satran" <Julian_Satran@il.ibm.com>
Cc: tsvwg@ietf.org
X-Mailer: Lotus Notes Release 5.0.3 (Intl) 21 March 2000
Message-ID: <OF5E554C4E.D550AFAA-ON88256B5D.0004CD53@boulder.ibm.com>
From: "John Hufferd" <hufferd@us.ibm.com>
Date: Sun, 10 Feb 2002 17:22:35 -0800
X-MIMETrack: Serialize by Router on D03NM014/03/M/IBM(Release 5.0.9 |November 16, 2001) at
 02/10/2002 06:23:29 PM
MIME-Version: 1.0
Content-type: text/plain; charset=us-ascii
Subject: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Step, Julian,
While looking at the calculations in fresh light, I think I spotted an
error that gave Julian's ESC approach, to high a probability.  That is I
multiplied both ex and ey by 361.  I think I should have only multiplied
one by that value.

"ey" should have been left at 2.3e-10  and the resulting total probability
(ex * ey) should have been 1.934e-19

This would mean that the probability of a False Positive, given a
segmentation occurred, to be 1.934e-19 for ESC and the current TUF value
would still be 1.962e-17.

That would mean that if the CRC was part of the PDU, then the probability
would be even less, and yet only use 2 extra words would be used in the
framing header.

It seems clear however, if three words of header are used with TUF, vrs
normal ESC, that the value probability of a false positive with the 3 word
TUF at 4.557e-27 is  a lot better then the 1.934e-19 in the normal 3 word
header of ESC.


 .
.
 .
John L. Hufferd
Senior Technical Staff Member (STSM)
IBM/SSG San Jose Ca
Main Office (408) 256-0403, Tie: 276-0403,  eFax: (408) 904-4688
Home Office (408) 997-6136, Cell: (408) 499-9702
Internet address: hufferd@us.ibm.com
---------------------- Forwarded by John Hufferd/San Jose/IBM on 02/10/2002
04:52 PM ---------------------------

John Hufferd
02/09/2002 05:05 PM

To:    Stephen Bailey <steph@cs.uchicago.edu>, Julian
       Satran/Haifa/IBM@IBMIL
cc:    tsvwg@ietf.org
From:  John Hufferd/San Jose/IBM@IBMUS
Subject:    Re: [rdma] a proposal for a different No-Data-Touch framer for
       TCP  (Document link: John Hufferd)

Steph, Julian,
The following is an attempt to determine the probability of a false
positive, both with the TUF approach as defined in
draft-ietf-tsvwg-tcp-ulp-frame-01.txt, and the semi draft that Julian
proposed.

Since we are only interested in the case of False Positives given a
re-segmentation occurs, I think it is possible to determine which of the
two approaches is less risky.

Here are the two layouts:

Current TUF has a 64 bit header at the beginning of Segment:
1. Length of 16 bits (Must match up with the physical length of the
segment)
2. Key  48 bits (Key Randomly Chosen by ULP,  -- Known by sender and
receiver, and changed as some undefined interval, if at all)

Julian's Eye-catcher-Salt-CRC approach (which I will call ESC) is 96 bits
1. Eye-catcher - 32 bits (Chosen by the ULP,  could also be chosen like the
Key in TUF, known by both sides)
2. Random number - 32 bits (Generated by ULP, unique to this PDU only)
3. CRC  - 32 bits (Generated by ULP based on 1. & 2.)

Lets first examine the ESC.

Once the Eye-catcher is determined, the Random number is chosen and the CRC
is computed, but to the Framer this is just string of bits passed from the
ULP.  The Receiving Framer (deFramer?) does not know the actual magic
string of bits which it should look for within the incoming payload.
Therefore it needs to do a computation at the start of each segment to make
sure the approprate string is received.

Given any beginning Eye-catcher, randomly chosen or not, there are a number
of 64 bit strings which can be interpreted as a valid salt and CRC.  That
is, for every 32 bit Salt, there is a valid 32 bit CRC.  Therefore, there
is as many valid 64-bit strings, which can follow the Eye-catcher as there
are unique 32-bit Salt numbers.  This means, there are (2**32) valid 8 byte
strings that can follow the Eye-catcher.

So in the case of ESC we have:

 The probability "ex" that a 32-bit Random Eye-catcher (or Key) is
"incorrectly" (do to re-segmentation) located at the beginning of a segment
, and the probability "ey" that one of the 2**32 valid 8 byte strings will
follow that Eye-catcher.  I think that makes it the total probability equal
to:
ex times  ey.

We have "ex" equal to the probability of the Eye-catcher occurring in any
word times the number of chances that might happen.  This means the
probability of any 32 bit value occurring  is 1/(2**32) times 361  where
361 is the number of possible segment points that could have been made in a
1460 byte payload (with a total of 3 header words and one byte of data).
Therefore:

ex = (1/2**32) times 361
ex = 2.3e-10 * 361
or
ex = 8.41e-8

The value of "ey" is equal to the probability that one of the 2**32 valid 8
byte strings will follow the Eye-catcher.  This is 1/2**32 times the number
of chances.  In this case the number of chances equals the number of
possible segment points that could have been made in a 1460 byte payload
(with a total of 3 header words and one byte of data), which is 361.
Therefore:

ey = 2.3e-10 * 361
or
ey = 8.41e-8

Therefore with ESC  we have the total probability of false positive of:

ex*ey = 8.41e-8 * 8.41e-8
or
ex*ey = 7.06 e-15


Now we will examine  the case of the current TUF proposal:

We have a probability "tx" that a 48-bit Random Key (or Eye-catcher) is
"incorrectly" (do to re-segmentation) located at + 2 bytes from the
beginning of a segment , and the probably "ty" that a 16 bit field
preceding it, will be a valid length of the current Segment.

I think that probability is equal to  tx  times  ty.

Now "ty" has 1 chance of occurring in 2**16 -10 (there must be at lease 8
bytes of TUF header + one byte of data, and zero is not valid so subtract
10 from the number of valid lengths)
ty = 1/((2**16)-10)
ty = 1/65526
ty = 1.526e-5

"tx" is the probability of a specific 48 bit string occurring at +2 bytes
into each possible segmentation point.  This is the probability of the
string occurring times the number of possible chances it has to occur.
Since there can be at most 365 TCP payload words in a normal Ethernet Frame
((1500-20-20)/4 =1460/4 = 365 words) then that means there are 365-3 or 362
possible segment point, which could leave at least a 2 word header and 1
byte of data.
Therefore:
tx = (1/2**48) times 362
tx = ((3.55e-15) * 362
tx = 1.286e-12).

Therefore with Current TUF we have the total probability of false positive
of:

tx*ty = 1.286e-12 * 1.526e-5
or
tx =  1.962e-17

Therefore, based on the calculations above, I believe that the Current TUF
approach offers a lower probability of False positive then does the
straight version of ESC.

But remember, Current TUF has a 64 bit header and ESC has a 96 bit header.
If we were to extend the TUF proposal to 96 bits the calculation would look
like:

ty = 1.526e-5 (Same as before)

AND,

tx =l (1/2**80) * 361
tx = 8.27e-25 * 361
tx = 2.986e-22

and
tx * ty = 2.986e-22 * 1.526e-5
tx * ty = 4.557e-27

However, ECS can also use a PDU header as part of the total salt.  The PDU
is not a fully random set of bits, but it does also reduce the probability
of False Positive a great deal more.  (I will let someone else compute that
probability.) The current probability of a false positive with a TUF's two
word header is 1.962e-17, but I think it is safe to guess that the ECS when
factoring in a PDU, would lower its probability a great deal lower then
7.06 e-15 (probably lower then current TUF), and it would then only need a
two word header like TUF.

Summary:
If you compare Current TUF (which has 2 word headers) to ECS (which has 3
word headers) you have a lower probability of False Positives with TUF.  If
you compare Current TUF with ECS that uses the CRC with the Header of the
PDU (so ECS only has a 2 byte Header) there is reason to believe that ECS
would have the lower probability of false positive.

So the questions should be: is  1.962e-17 a low enough probability of false
positives that we can get on with the Framing per the current TUFs
proposal?  Or do we need to add a third word to TUFs, or use the ECS with
PDU CRC.


 .
.
 .
John L. Hufferd
Senior Technical Staff Member (STSM)
IBM/SSG San Jose Ca
Main Office (408) 256-0403, Tie: 276-0403,  eFax: (408) 904-4688
Home Office (408) 997-6136, Cell: (408) 499-9702
Internet address: hufferd@us.ibm.com


Stephen Bailey <steph@cs.uchicago.edu> on 02/09/2002 08:35:17 AM

To:    Julian Satran/Haifa/IBM@IBMIL
cc:    rdma@yahoogroups.com, tsvwg@ietf.org
Subject:    Re: [rdma] a proposal for a different No-Data-Touch framer for
       TCP



For a proper safety analysis, you can not assume that the the contents
of the the stream (or the PDU) are randomly distributed.  You have to
assume the stream has worst-case contents.  Defining `worst case' is
the hard part.

Steph

> The mathematics behind the proposal are simple. You have two random
> numbers "against" a finite-length PDU.
> In you original proposal you have single random number (the marker)
> against a quasi infinite stream. The chance to miss is 1.
> With any decent digest (e.g. a CRC) you have a small (2**-64) chance to
> miss per PDU word (for every new PDU you have a new pattern).  All the
> tricks like alignment etc do not decrease the cardinality of stream -
> i.e., it is still infinite and then a constant framer is no good.
> The cost of the new scheme is also tolerable.

------------------------ Yahoo! Groups Sponsor ---------------------~-->
Sponsored by VeriSign - The Value of Trust
Pinpoint the right security solution for your company - FREE
Guide from industry leader VeriSign gives you all the facts.
http://us.click.yahoo.com/pCuuSA/WdiDAA/yigFAA/W6uqlB/TM
---------------------------------------------------------------------~->

To unsubscribe from this group, send an email to:
rdma-unsubscribe@egroups.com



Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/








_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sun Feb 10 20:44:55 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA13304
	for <tsvwg-archive@odin.ietf.org>; Sun, 10 Feb 2002 20:44:55 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA02399
	for tsvwg-archive@odin.ietf.org; Sun, 10 Feb 2002 20:44:57 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA01683;
	Sun, 10 Feb 2002 20:27:08 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA01654
	for <tsvwg@optimus.ietf.org>; Sun, 10 Feb 2002 20:27:05 -0500 (EST)
Received: from e31.co.us.ibm.com (e31.co.us.ibm.com [32.97.110.129])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA12961
	for <tsvwg@ietf.org>; Sun, 10 Feb 2002 20:27:01 -0500 (EST)
Received: from westrelay03.boulder.ibm.com (westrelay03.boulder.ibm.com [9.99.140.24])
	by e31.co.us.ibm.com (8.9.3/8.9.3) with ESMTP id UAA111372;
	Sun, 10 Feb 2002 20:23:54 -0500
Received: from d03nm014.boulder.ibm.com (avpilot.boulder.ibm.com [9.17.188.135])
	by westrelay03.boulder.ibm.com (8.11.1m3/NCO/VER6.00) with ESMTP id g1B1R3B78242;
	Sun, 10 Feb 2002 18:27:03 -0700
X-Priority: 1 (High)
Importance: Normal
To: "Julian Satran" <Julian_Satran@il.ibm.com>
Cc: "Leonid Grossman" <leonidg@attbi.com>, rdma@yahoogroups.com,
        tsvwg@ietf.org
X-Mailer: Lotus Notes Release 5.0.3 (Intl) 21 March 2000
Message-ID: <OFA2AD92F7.551FB699-ON88256B5D.00079FC8@boulder.ibm.com>
From: "John Hufferd" <hufferd@us.ibm.com>
Date: Sun, 10 Feb 2002 17:26:09 -0800
X-MIMETrack: Serialize by Router on D03NM014/03/M/IBM(Release 5.0.9 |November 16, 2001) at
 02/10/2002 06:27:02 PM
MIME-Version: 1.0
Content-type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id UAA01655
Subject: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit


Julian,
I do not think any of that matters, since you are not trying to determine
the probability of a False Negative.  Only the probability of a False
Positive at the beginning of a frame, given a re-segmentation, is the
issue.

 .
.
 .
John L. Hufferd
Senior Technical Staff Member (STSM)
IBM/SSG San Jose Ca
Main Office (408) 256-0403, Tie: 276-0403,  eFax: (408) 904-4688
Home Office (408) 997-6136, Cell: (408) 499-9702
Internet address: hufferd@us.ibm.com


Julian Satran/Haifa/IBM@IBMIL on 02/10/2002 11:22:22 AM

To:    "Leonid Grossman" <leonidg@attbi.com>
cc:    rdma@yahoogroups.com, tsvwg@ietf.org
Subject:    Re: [rdma] a proposal for a different No-Data-Touch framer for
       TCP




Leonid,

The probability is not 0 - that is true. But for the form used in the ULP
framing draft it is in theory 1 and that is unacceptable.
The new proposal is aimed at macing it a very low (acceptable number) -
equivalent with the probability of an undetected error by a CRC (in fact
substantially less).

As for directing DDP traffic to a specific port that is an interesting idea
(although not related to the issue at hand).
Even in a separate ring you are still faced with the same probability as
long as the stream is infinite. The gain that you can obtain is the
coexistence of DDP streams with "non-steered" streams and you pay by
maintaining a more complex stack.


Julo


                                                                           
     "Leonid Grossman"                                                     
     <leonidg@attbi.com>                 To:                               
                                 <satran@haifa.vnet.ibm.com>               
     09-02-02 19:22                      cc:                               
                                         Subject:        Re: [rdma] a      
                                 proposal for a different No-Data-Touch    
                                 framer for TCP                            
                                                                           
                                                                           
                                                                           



Julian - either proposal has a non-zero probability to mistaken TCP
payload in non-DDP frame for a valid framing (and potentially cause
a data corruption).
I'm deciding whether to put some framing and DDP on TCP support in our
upcoming 10GigE NIC ASIC, and to me it looks like a simpler and
arguably more reliable way to recognize DDP frames would be to just
use a dedicated port number for DDP.
For outgoing frames, DDP shim can replace application port number
with DDP port number, and store the original port in iWARP/DDPP
header somewhere.
On receive, NIC hardware would use DDP port number to recognize DDP
traffic and steer it into a separate receive ring (with protocol and
DDPP headers separated into pre-allocated buffers);
then the hardware would verify iWAPR/DDPP header, do map tag into
physical address of the locked DDP buffers and DMA DDP payload.
SDP shim on receive would steer the traffic based upon the original
port number.
Anyway, identifying DDP traffic in TCP header rather than in TCP payload
looks like a better approach.

Any comments?

Thanks in advance,
Leonid Grossman
Software Engineering
S2IO Technologies
Cupertino, California




> The mathematics behind the proposal are simple. You have two random
> numbers "against" a finite-length PDU.
> In you original proposal you have single random number (the marker)
> against a quasi infinite stream. The chance to miss is 1.
> With any decent digest (e.g. a CRC) you have a small (2**-64) chance to
> miss per PDU word (for every new PDU you have a new pattern). All the
> tricks like alignment etc do not decrease the cardinality of stream -
> i.e., it is still infinite and then a constant framer is no good.
> The cost of the new scheme is also tolerable.




                                                                            
                               Yahoo! Groups Sponsor                        
                                                                            
                                                                            
                                                                            
                                       ADVERTISEMENT                        
                                          [IMAGE]                           
                                                                            
                                                                            
         [IMAGE]                                                            
                                                                            


 To unsubscribe from this group, send an email to:
rdma-unsubscribe@egroups.com



Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.





_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Feb 11 01:33:38 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA18008
	for <tsvwg-archive@odin.ietf.org>; Mon, 11 Feb 2002 01:33:38 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id BAA19943
	for tsvwg-archive@odin.ietf.org; Mon, 11 Feb 2002 01:33:38 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA17386;
	Mon, 11 Feb 2002 01:12:39 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA17344
	for <tsvwg@ns.ietf.org>; Mon, 11 Feb 2002 01:12:36 -0500 (EST)
Received: from blaze.hcltech.com ([203.199.199.225])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA17831
	for <tsvwg@ietf.org>; Mon, 11 Feb 2002 01:12:35 -0500 (EST)
Received: from netlab.hcltech.com (arun [192.168.201.67])
	by blaze.hcltech.com (8.9.3/8.9.3) with ESMTP id LAA14264;
	Mon, 11 Feb 2002 11:42:34 +0530
Message-ID: <3C675B99.E6E0FBE0@netlab.hcltech.com>
Date: Mon, 11 Feb 2002 11:20:17 +0530
From: Arun Prasad <arun@netlab.hcltech.com>
X-Mailer: Mozilla 4.72 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Coene Lode <Lode.Coene@siemens.atea.be>
CC: "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "'multi6@ops.ietf.org'" <multi6@ops.ietf.org>
Subject: Re: [Tsvwg] SCTP multihoming issues draft
References: <E76F715C0429D5118F2100508BB9EDEE0795AC@hrtades7.atea.be>
Content-Type: multipart/alternative;
 boundary="------------873F2CBC2D112028880E143C"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


--------------873F2CBC2D112028880E143C
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

hi Coene,
    Regarding the below section in the draft,
    2.2 SCTP multihoming and the size of routing tables
    what are we trying to achieve....   section 2.1 basically explained the restrictions
of SCTP's multihoming feature due to existing IP architecture....  The problems
are well sorted out, and one solution I get from that is,   changing the Source
Ip address for the packets send ( periodically, if the acknowledgemnet for it is
not reaching the endpoint,) from a multihomed endpoint....
    But regarding section 2.2,  SCTP could do nothing about it ( because its
basically IP's restriction or the restriction of the currect routing architecture....)
SCTP defines some rules or features, using them properly is the responsiblity of the
Application which uses it (it may the administrator of the system also...) ..  This
each of them using it can implement their own ways to get the maximum from use
of multihoming feature of SCTP......
      Will it not make SCTP more heavy, if we try to load it with more and more
sub features like this... and it will make it more dependent on the existing architecture
which will make the maintanance tough.....
    But definitely we should be spelling out the problems or restrictions of the
multihoming feature of SCTP, with which the users of SCTP users will do
appropriate action to use it best....

Is the observation valid??

Thanks
-arun

Coene Lode wrote:

> Previous week, the draft on SCTP multihoming issues was published.
>
> Apart from some editorial changes concerning author list and missing/wrong
> page numbers,
> the only major change is a sentence added at the end of paragraph 2.3 SCTP
> multihoming and Network Adress Translators(NAT):
> "It should be noted that the NAT box becomes a single
>     point-of-failure in this case, as ALL the paths of the SCTP
>     association have to go through that single NAT box."
> Theoretical it is possible to let all the different paths run through
> different NAT boxes, but then the NAT boxes would never know exactly when to
> teardown/terminate the association as the terminate or abort takes only one
> of the paths(and there are other difficulties, such as the heartbeats
> ...etc...) So in practical terms this sentence should be correct. If you
> feel otherwise, please let me know and I'll remove or add something to it to
> make everything clear.
>
> The draft can be found at the IETF draft directory:
> Multihoming issues in the Stream Control Transmission Protocol
> <draft-coene-sctp-multihome-02.txt>
>
> If by then end of next week, no further issues have been identified, then I
> would ask for this draft to be moved to informational.
>
> yours sincerly
> Lode Coene
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

--
****************************************************************
V.Arun Prasad
HCL Technologies Ltd.
51, Jawaharlal Nehru Road,
Ekkattuthangal,
Guindy Industrial Estate,
Chennai - 600097.

Contact # : 9144 - 2334174
            9144 - 2334181
            extn : 233
****************************************************************


--------------873F2CBC2D112028880E143C
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
hi Coene,
<br>&nbsp;&nbsp;&nbsp; Regarding the below section in the draft,
<br><i>&nbsp;&nbsp;&nbsp; 2.2 SCTP multihoming and the size of routing
tables</i>
<br>&nbsp;&nbsp;&nbsp; what are we trying to achieve....&nbsp;&nbsp; section
2.1 basically explained the restrictions
<br>of SCTP's multihoming feature due to existing IP architecture....&nbsp;
The problems
<br>are well sorted out, and one solution I get from that is,&nbsp;&nbsp;
changing the Source
<br>Ip address for the packets send ( periodically, if the acknowledgemnet
for it is
<br>not reaching the endpoint,) from a multihomed endpoint....
<br>&nbsp;&nbsp;&nbsp; But regarding section 2.2,&nbsp; SCTP could do nothing
about it ( because its
<br>basically IP's restriction or the restriction of the currect routing
architecture....)
<br>SCTP defines some rules or features, using them properly is the responsiblity
of the
<br>Application which uses it (it may the administrator of the system also...)
..&nbsp; This
<br>each of them using it can implement their own ways to get the maximum
from use
<br>of multihoming feature of SCTP......
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Will it not make SCTP more heavy, if
we try to load it with more and more
<br>sub features like this... and it will make it more dependent on the
existing architecture
<br>which will make the maintanance tough.....
<br>&nbsp;&nbsp;&nbsp; But definitely we should be spelling out the problems
or restrictions of the
<br>multihoming feature of SCTP, with which the users of SCTP users will
do
<br>appropriate action to use it best....
<p>Is the observation valid??
<p>Thanks
<br>-arun
<p>Coene Lode wrote:
<blockquote TYPE=CITE>Previous week, the draft on SCTP multihoming issues
was published.
<p>Apart from some editorial changes concerning author list and missing/wrong
<br>page numbers,
<br>the only major change is a sentence added at the end of paragraph 2.3
SCTP
<br>multihoming and Network Adress Translators(NAT):
<br>"It should be noted that the NAT box becomes a single
<br>&nbsp;&nbsp;&nbsp; point-of-failure in this case, as ALL the paths
of the SCTP
<br>&nbsp;&nbsp;&nbsp; association have to go through that single NAT box."
<br>Theoretical it is possible to let all the different paths run through
<br>different NAT boxes, but then the NAT boxes would never know exactly
when to
<br>teardown/terminate the association as the terminate or abort takes
only one
<br>of the paths(and there are other difficulties, such as the heartbeats
<br>...etc...) So in practical terms this sentence should be correct. If
you
<br>feel otherwise, please let me know and I'll remove or add something
to it to
<br>make everything clear.
<p>The draft can be found at the IETF draft directory:
<br>Multihoming issues in the Stream Control Transmission Protocol
<br>&lt;draft-coene-sctp-multihome-02.txt>
<p>If by then end of next week, no further issues have been identified,
then I
<br>would ask for this draft to be moved to informational.
<p>yours sincerly
<br>Lode Coene
<p>_______________________________________________
<br>tsvwg mailing list
<br>tsvwg@ietf.org
<br><a href="https://www1.ietf.org/mailman/listinfo/tsvwg">https://www1.ietf.org/mailman/listinfo/tsvwg</a></blockquote>
--
<br>****************************************************************
<br>V.Arun Prasad
<br>HCL Technologies Ltd.
<br>51, Jawaharlal Nehru Road,
<br>Ekkattuthangal,
<br>Guindy Industrial Estate,
<br>Chennai - 600097.
<p>Contact # : 9144 - 2334174
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
9144 - 2334181
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
extn : 233
<br>****************************************************************
<br>&nbsp;</html>

--------------873F2CBC2D112028880E143C--


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Feb 11 09:10:49 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA02728
	for <tsvwg-archive@odin.ietf.org>; Mon, 11 Feb 2002 09:10:49 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA09496
	for tsvwg-archive@odin.ietf.org; Mon, 11 Feb 2002 09:10:50 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA08549;
	Mon, 11 Feb 2002 08:54:20 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA08471
	for <tsvwg@ns.ietf.org>; Mon, 11 Feb 2002 08:54:17 -0500 (EST)
Received: from sj-msg-core-4.cisco.com (sj-msg-core-4.cisco.com [171.71.163.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA01919
	for <tsvwg@ietf.org>; Mon, 11 Feb 2002 08:54:15 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-4.cisco.com (8.11.3/8.9.1) with ESMTP id g1BDrkt18129;
	Mon, 11 Feb 2002 05:53:46 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABB19043;
	Mon, 11 Feb 2002 05:53:43 -0800 (PST)
Message-ID: <3C67CCE6.51AB848A@stewart.chicago.il.us>
Date: Mon, 11 Feb 2002 07:53:42 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Arun Prasad <arun@netlab.hcltech.com>
CC: Coene Lode <Lode.Coene@siemens.atea.be>,
        "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "'multi6@ops.ietf.org'" <multi6@ops.ietf.org>
Subject: Re: [Tsvwg] SCTP multihoming issues draft
References: <E76F715C0429D5118F2100508BB9EDEE0795AC@hrtades7.atea.be> <3C675B99.E6E0FBE0@netlab.hcltech.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Arun:

I don't think we need to have a host become
routing aware as section 2.2 seems to imply to
some degree. Some clever work with the routing
system in the host can alleviate the problem that
Lode is mentioning here I think..

For example in our BSD work we have now the ability
of the routing system to have multiple default routes
and a method to lookup a route via an alternate interface.
We probably need to expand this to actually look at the gateway
and not the interface.. but for now we use interface :/

This allows the sctp stack to call 

rtalloc_alternate(dst,old_route);

which then trys to give a route out a different interface if
necessary hunting up the routing tree.

With this change we then get:

A) The host only needs to know about the edge routes available
   via its multiple default routes.

B) Full use of the multi-homing provided without injection
   of static routes as the v4 side does now...

C) Host only knows about the default routes and possibly
   some special routes applied by the local administrator...

R

Arun Prasad wrote:
> 
> hi Coene,
>     Regarding the below section in the draft,
>     2.2 SCTP multihoming and the size of routing tables
>     what are we trying to achieve....   section 2.1 basically
> explained the restrictions
> of SCTP's multihoming feature due to existing IP architecture....  The
> problems
> are well sorted out, and one solution I get from that is,   changing
> the Source
> Ip address for the packets send ( periodically, if the acknowledgemnet
> for it is
> not reaching the endpoint,) from a multihomed endpoint....
>     But regarding section 2.2,  SCTP could do nothing about it (
> because its
> basically IP's restriction or the restriction of the currect routing
> architecture....)
> SCTP defines some rules or features, using them properly is the
> responsiblity of the
> Application which uses it (it may the administrator of the system
> also...) ..  This
> each of them using it can implement their own ways to get the maximum
> from use
> of multihoming feature of SCTP......
>       Will it not make SCTP more heavy, if we try to load it with more
> and more
> sub features like this... and it will make it more dependent on the
> existing architecture
> which will make the maintanance tough.....
>     But definitely we should be spelling out the problems or
> restrictions of the
> multihoming feature of SCTP, with which the users of SCTP users will
> do
> appropriate action to use it best....
> 
> Is the observation valid??
> 
> Thanks
> -arun
> 
> Coene Lode wrote:
> 
> > Previous week, the draft on SCTP multihoming issues was published.
> >
> > Apart from some editorial changes concerning author list and
> > missing/wrong
> > page numbers,
> > the only major change is a sentence added at the end of paragraph
> > 2.3 SCTP
> > multihoming and Network Adress Translators(NAT):
> > "It should be noted that the NAT box becomes a single
> >     point-of-failure in this case, as ALL the paths of the SCTP
> >     association have to go through that single NAT box."
> > Theoretical it is possible to let all the different paths run
> > through
> > different NAT boxes, but then the NAT boxes would never know exactly
> > when to
> > teardown/terminate the association as the terminate or abort takes
> > only one
> > of the paths(and there are other difficulties, such as the
> > heartbeats
> > ...etc...) So in practical terms this sentence should be correct. If
> > you
> > feel otherwise, please let me know and I'll remove or add something
> > to it to
> > make everything clear.
> >
> > The draft can be found at the IETF draft directory:
> > Multihoming issues in the Stream Control Transmission Protocol
> > <draft-coene-sctp-multihome-02.txt>
> >
> > If by then end of next week, no further issues have been identified,
> > then I
> > would ask for this draft to be moved to informational.
> >
> > yours sincerly
> > Lode Coene
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> --
> ****************************************************************
> V.Arun Prasad
> HCL Technologies Ltd.
> 51, Jawaharlal Nehru Road,
> Ekkattuthangal,
> Guindy Industrial Estate,
> Chennai - 600097.
> 
> Contact # : 9144 - 2334174
>             9144 - 2334181
>             extn : 233
> ****************************************************************
> 

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Feb 11 09:15:03 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA02892
	for <tsvwg-archive@odin.ietf.org>; Mon, 11 Feb 2002 09:15:02 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA09679
	for tsvwg-archive@odin.ietf.org; Mon, 11 Feb 2002 09:15:03 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA09184;
	Mon, 11 Feb 2002 09:01:09 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA09131
	for <tsvwg@ns.ietf.org>; Mon, 11 Feb 2002 09:01:07 -0500 (EST)
Received: from sj-msg-core-4.cisco.com (sj-msg-core-4.cisco.com [171.71.163.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA02194;
	Mon, 11 Feb 2002 09:01:05 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-4.cisco.com (8.11.3/8.9.1) with ESMTP id g1BE0at20821;
	Mon, 11 Feb 2002 06:00:36 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABB19094;
	Mon, 11 Feb 2002 06:00:34 -0800 (PST)
Message-ID: <3C67CE81.463F99B4@stewart.chicago.il.us>
Date: Mon, 11 Feb 2002 08:00:33 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
CC: Pekka Savola <pekkas@netcore.fi>, Coene Lode <Lode.Coene@siemens.atea.be>,
        "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)" <sigtran@ietf.org>
Subject: Re: [Tsvwg] Re: SCTP multihoming issues draft
References: <Pine.LNX.4.44.0202081211000.4989-100000@netcore.fi> <200202081214.g18CECj9028264@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Andreas Jungmaier wrote:
...snip...snip...
> > ==> consider the security of introducing new addresses to the association
> > and capturing traffic: granted, this is an SCTP concern but is highlighted
> 
> No it is not, because SCTP according to all RFC 2960 does not
> allow for adding addresses.
> 

Andreas:

The current RFC does not allow the addition of addresses but
the extension in the tsvwg does.. aka sctp-addip...

R
-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Feb 11 11:39:25 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA08863
	for <tsvwg-archive@odin.ietf.org>; Mon, 11 Feb 2002 11:39:25 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA18622
	for tsvwg-archive@odin.ietf.org; Mon, 11 Feb 2002 11:39:27 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA17292;
	Mon, 11 Feb 2002 11:21:33 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA17243
	for <tsvwg@ns.ietf.org>; Mon, 11 Feb 2002 11:21:31 -0500 (EST)
Received: from irmgard.exp-math.uni-essen.de (irmgard.exp-math.uni-essen.de [132.252.150.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA07938;
	Mon, 11 Feb 2002 11:21:28 -0500 (EST)
Received: from there (biene.exp-math.uni-essen.de [132.252.150.215])
	by irmgard.exp-math.uni-essen.de (8.12.2/8.12.2) with SMTP id g1BGLSj9030904;
	Mon, 11 Feb 2002 17:21:29 +0100
Message-Id: <200202111621.g1BGLSj9030904@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain;
  charset="iso-8859-1"
From: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
Organization: University Essen
To: Randall Stewart <randall@stewart.chicago.il.us>
Subject: Re: [Tsvwg] Re: SCTP multihoming issues draft
Date: Mon, 11 Feb 2002 18:21:05 +0100
X-Mailer: KMail [version 1.3.1]
Cc: "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)" <sigtran@ietf.org>
References: <Pine.LNX.4.44.0202081211000.4989-100000@netcore.fi> <200202081214.g18CECj9028264@irmgard.exp-math.uni-essen.de> <3C67CE81.463F99B4@stewart.chicago.il.us>
In-Reply-To: <3C67CE81.463F99B4@stewart.chicago.il.us>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Randy, 

I count on testing the extension on the Connectathon, so I was 
actually aware of that :-) 
However, the security section should only address the SCTP base 
protocol, I think, so there is no need to add that. It should 
be mentioned (and I think it already is)  in the security section 
of the add-ip draft.

Best regards,
Andreas

On Monday, 11. February 2002 15:00, Randall Stewart wrote:
> Andreas Jungmaier wrote:
> ...snip...snip...
>
> > > ==> consider the security of introducing new addresses to the
> > > association and capturing traffic: granted, this is an SCTP concern but
> > > is highlighted
> >
> > No it is not, because SCTP according to all RFC 2960 does not
> > allow for adding addresses.
>
> Andreas:
>
> The current RFC does not allow the addition of addresses but
> the extension in the tsvwg does.. aka sctp-addip...
>
> R

- -- 
- ------------------------------------------------------------------------
Dipl.-Ing. Andreas Jungmaier              |                         _
Computer Networking Technology Group      | ASCII ribbon campaign  ( )
University of Essen                       |  - against HTML email   X
http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8Z/2HXAsLBNOCSsARAgnyAJ9CpYEASNqzropx9zcmYzJE03F1MQCg8yyV
HLr4ND0Yi4T2I9hsO+s1cWk=
=ULys
-----END PGP SIGNATURE-----

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 11 12:56:46 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA12619
	for <tsvwg-archive@odin.ietf.org>; Mon, 11 Feb 2002 12:56:46 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id MAA23797
	for tsvwg-archive@odin.ietf.org; Mon, 11 Feb 2002 12:56:47 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA23206;
	Mon, 11 Feb 2002 12:39:01 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA23175
	for <tsvwg@optimus.ietf.org>; Mon, 11 Feb 2002 12:38:59 -0500 (EST)
Received: from d12lmsgate-3.de.ibm.com (d12lmsgate-3.de.ibm.com [195.212.91.201])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA12185
	for <tsvwg@ietf.org>; Mon, 11 Feb 2002 12:38:56 -0500 (EST)
Received: from d12relay01.de.ibm.com (d12relay01.de.ibm.com [9.165.215.22])
	by d12lmsgate-3.de.ibm.com (1.0.0) with ESMTP id SAA54374;
	Mon, 11 Feb 2002 18:38:27 +0100
Received: from d10hubm1.telaviv.ibm.com (d10ml001.telaviv.ibm.com [9.148.216.55])
	by d12relay01.de.ibm.com (8.11.1m3/NCO v5.01) with ESMTP id g1BHdvG83194;
	Mon, 11 Feb 2002 18:39:57 +0100
To: "John Hufferd" <hufferd@us.ibm.com>
Cc: "Leonid Grossman" <leonidg@attbi.com>, rdma@yahoogroups.com,
        tsvwg@ietf.org
MIME-Version: 1.0
X-Mailer: Lotus Notes Release 5.0.9  November 16, 2001
From: "Julian Satran" <Julian_Satran@il.ibm.com>
Message-ID: <OF2853CB76.419150F7-ONC2256B5D.0060D522@telaviv.ibm.com>
Date: Mon, 11 Feb 2002 19:39:58 +0200
X-MIMETrack: Serialize by Router on D10ML001/10/M/IBM(Release 5.0.8 |June 18, 2001) at
 11/02/2002 19:40:01,
	Serialize complete at 11/02/2002 19:40:01
Content-Type: multipart/alternative; boundary="=_alternative 0060E3E9C2256B5D_="
Subject: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

This is a multipart message in MIME format.
--=_alternative 0060E3E9C2256B5D_=
Content-Type: text/plain; charset="us-ascii"

John - it is all about false positive.

Julo




John Hufferd@IBMUS
11-02-02 03:26


        To:     Julian Satran/Haifa/IBM@IBMIL
        cc:     "Leonid Grossman" <leonidg@attbi.com>, rdma@yahoogroups.com, 
tsvwg@ietf.org
        From:   John Hufferd/San Jose/IBM@IBMUS
        Subject:        Re: [rdma] a proposal for a different No-Data-Touch framer for TCP
 





Julian,
I do not think any of that matters, since you are not trying to determine 
the probability of a False Negative.  Only the probability of a False 
Positive at the beginning of a frame, given a re-segmentation, is the 
issue.

 .
.
 .
John L. Hufferd
Senior Technical Staff Member (STSM)
IBM/SSG San Jose Ca
Main Office (408) 256-0403, Tie: 276-0403,  eFax: (408) 904-4688
Home Office (408) 997-6136, Cell: (408) 499-9702
Internet address: hufferd@us.ibm.com

To:     "Leonid Grossman" <leonidg@attbi.com>
cc:     rdma@yahoogroups.com, tsvwg@ietf.org 
Subject:        Re: [rdma] a proposal for a different No-Data-Touch framer for TCP




Leonid, 

The probability is not 0 - that is true. But for the form used in the ULP 
framing draft it is in theory 1 and that is unacceptable. 
The new proposal is aimed at macing it a very low (acceptable number) - 
equivalent with the probability of an undetected error by a CRC (in fact 
substantially less). 

As for directing DDP traffic to a specific port that is an interesting 
idea (although not related to the issue at hand). 
Even in a separate ring you are still faced with the same probability as 
long as the stream is infinite. The gain that you can obtain is the 
coexistence of DDP streams with "non-steered" streams and you pay by 
maintaining a more complex stack. 


Julo 




"Leonid Grossman" <leonidg@attbi.com> 

09-02-02 19:22 

        
        To:        <satran@haifa.vnet.ibm.com> 
        cc:         
        Subject:        Re: [rdma] a proposal for a different 
No-Data-Touch framer for TCP 

       

Julian - either proposal has a non-zero probability to mistaken TCP
payload in non-DDP frame for a valid framing (and potentially cause 
a data corruption).
I'm deciding whether to put some framing and DDP on TCP support in our
upcoming 10GigE NIC ASIC, and to me it looks like a simpler and 
arguably more reliable way to recognize DDP frames would be to just 
use a dedicated port number for DDP.
For outgoing frames, DDP shim can replace application port number 
with DDP port number, and store the original port in iWARP/DDPP 
header somewhere.
On receive, NIC hardware would use DDP port number to recognize DDP
traffic and steer it into a separate receive ring (with protocol and
DDPP headers separated into pre-allocated buffers); 
then the hardware would verify iWAPR/DDPP header, do map tag into
physical address of the locked DDP buffers and DMA DDP payload.
SDP shim on receive would steer the traffic based upon the original 
port number.
Anyway, identifying DDP traffic in TCP header rather than in TCP payload 
looks like a better approach.

Any comments?

Thanks in advance,
Leonid Grossman
Software Engineering
S2IO Technologies
Cupertino, California




> The mathematics behind the proposal are simple. You have two random 
> numbers "against" a finite-length PDU.
> In you original proposal you have single random number (the marker) 
> against a quasi infinite stream. The chance to miss is 1.
> With any decent digest (e.g. a CRC) you have a small (2**-64) chance to 
> miss per PDU word (for every new PDU you have a new pattern). All the 
> tricks like alignment etc do not decrease the cardinality of stream - 
> i.e., it is still infinite and then a constant framer is no good.
> The cost of the new scheme is also tolerable.

 



Yahoo! Groups Sponsor

ADVERTISEMENT
[IMAGE]

[IMAGE]
 To unsubscribe from this group, send an email to:
rdma-unsubscribe@egroups.com

 

Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service. 





--=_alternative 0060E3E9C2256B5D_=
Content-Type: text/html; charset="us-ascii"


<br><font size=2 face="sans-serif">John - it is all about false positive.</font>
<br>
<br><font size=2 face="sans-serif">Julo</font>
<br>
<br>
<br>
<table width=100%>
<tr valign=top>
<td>
<td><font size=1 face="sans-serif"><b>John Hufferd@IBMUS</b></font>
<p><font size=1 face="sans-serif">11-02-02 03:26</font>
<br>
<td>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; To: &nbsp; &nbsp; &nbsp; &nbsp;Julian Satran/Haifa/IBM@IBMIL</font>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; cc: &nbsp; &nbsp; &nbsp; &nbsp;&quot;Leonid Grossman&quot; &lt;leonidg@attbi.com&gt;, rdma@yahoogroups.com, tsvwg@ietf.org</font>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; From: &nbsp; &nbsp; &nbsp; &nbsp;John Hufferd/San Jose/IBM@IBMUS</font>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; Subject: &nbsp; &nbsp; &nbsp; &nbsp;Re: [rdma] a proposal for a different No-Data-Touch framer for TCP</font><a href=Notes:///C225670D0041573F/38D46BF5E8F08834852564B500129B2C/082887EC4039776B87256B5C006A5775>Link</a>
<br><font size=1 face="Arial">&nbsp; </font>
<br>
<br>
<br>
<br></table>
<br>
<br><font size=2 face="sans-serif">Julian,</font>
<br><font size=2 face="sans-serif">I do not think any of that matters, since you are not trying to determine the probability of a False Negative. &nbsp;Only the probability of a False Positive at the beginning of a frame, given a re-segmentation, is the issue.<br>
<br>
.<br>
.<br>
.<br>
John L. Hufferd<br>
Senior Technical Staff Member (STSM)<br>
IBM/SSG San Jose Ca<br>
Main Office (408) 256-0403, Tie: 276-0403, &nbsp;eFax: (408) 904-4688<br>
Home Office (408) 997-6136, Cell: (408) 499-9702<br>
Internet address: hufferd@us.ibm.com<br>
</font>
<p><font size=1 color=#800080 face="sans-serif">To: &nbsp; &nbsp; &nbsp; &nbsp;</font><font size=1 face="sans-serif">&quot;Leonid Grossman&quot; &lt;leonidg@attbi.com&gt;</font>
<br><font size=1 color=#800080 face="sans-serif">cc: &nbsp; &nbsp; &nbsp; &nbsp;</font><font size=1 face="sans-serif">rdma@yahoogroups.com, tsvwg@ietf.org</font><font size=1 color=#800080 face="sans-serif"> </font>
<br><font size=1 color=#800080 face="sans-serif">Subject: &nbsp; &nbsp; &nbsp; &nbsp;</font><font size=1 face="sans-serif">Re: [rdma] a proposal for a different No-Data-Touch framer for TCP</font>
<br>
<br>
<br>
<br>
<br><font size=2>Leonid,</font><font size=3> </font>
<br>
<br><font size=2>The probability is not 0 - that is true. But for the form used in the ULP framing draft it is in theory 1 and that is unacceptable.</font><font size=3> </font>
<br><font size=2>The new proposal is aimed at macing it a very low (acceptable number) - equivalent with the probability of an undetected error by a CRC (in fact substantially less).</font><font size=3> </font>
<br>
<br><font size=2>As for directing DDP traffic to a specific port that is an interesting idea (although not related to the issue at hand).</font><font size=3> </font>
<br><font size=2>Even in a separate ring you are still faced with the same probability as long as the stream is infinite. The gain that you can obtain is the coexistence of DDP streams with &quot;non-steered&quot; streams and you pay by maintaining a more complex stack.</font><font size=3> </font>
<br>
<br>
<br><font size=2>Julo</font><font size=3> </font>
<br>
<br>
<br>
<table width=100%>
<tr valign=top>
<td width=0%>
<td width=35%><font size=1><b>&quot;Leonid Grossman&quot; &lt;leonidg@attbi.com&gt;</b></font><font size=3> </font>
<br>
<br><font size=1>09-02-02 19:22</font><font size=3> </font>
<br>
<td width=63%><font size=1>&nbsp; &nbsp; &nbsp; &nbsp; </font>
<br><font size=1>&nbsp; &nbsp; &nbsp; &nbsp; To: &nbsp; &nbsp; &nbsp; &nbsp;&lt;satran@haifa.vnet.ibm.com&gt;</font><font size=3> </font>
<br><font size=1>&nbsp; &nbsp; &nbsp; &nbsp; cc: &nbsp; &nbsp; &nbsp; &nbsp;</font><font size=3> </font>
<br><font size=1>&nbsp; &nbsp; &nbsp; &nbsp; Subject: &nbsp; &nbsp; &nbsp; &nbsp;Re: [rdma] a proposal for a different No-Data-Touch framer for TCP</font><font size=3> </font>
<br>
<br><font size=1>&nbsp; &nbsp; &nbsp; &nbsp;</font></table>
<br>
<br><font size=2>Julian - either proposal has a non-zero probability to mistaken TCP</font>
<br><font size=2>payload in non-DDP frame for a valid framing (and potentially cause </font>
<br><font size=2>a data corruption).</font>
<br><font size=2>I'm deciding whether to put some framing and DDP on TCP support in our</font>
<br><font size=2>upcoming 10GigE NIC ASIC, and to me it looks like a simpler and </font>
<br><font size=2>arguably more reliable way to recognize DDP frames would be to just </font>
<br><font size=2>use a dedicated port number for DDP.</font>
<br><font size=2>For outgoing frames, DDP shim can replace application port number </font>
<br><font size=2>with DDP port number, and store the original port in iWARP/DDPP </font>
<br><font size=2>header somewhere.</font>
<br><font size=2>On receive, NIC hardware would use DDP port number to recognize DDP</font>
<br><font size=2>traffic and steer it into a separate receive ring (with protocol and</font>
<br><font size=2>DDPP headers separated into pre-allocated buffers); </font>
<br><font size=2>then the hardware would verify iWAPR/DDPP header, do map tag into</font>
<br><font size=2>physical address of the locked DDP buffers and DMA DDP payload.</font>
<br><font size=2>SDP shim on receive would steer the traffic based upon the original </font>
<br><font size=2>port number.</font>
<br><font size=2>Anyway, identifying DDP traffic in TCP header rather than in TCP payload </font>
<br><font size=2>looks like a better approach.</font>
<br>
<br><font size=2>Any comments?</font>
<br>
<br><font size=2>Thanks in advance,</font>
<br><font size=2>Leonid Grossman</font>
<br><font size=2>Software Engineering</font>
<br><font size=2>S2IO Technologies</font>
<br><font size=2>Cupertino, California</font>
<br>
<br>
<br>
<br>
<br><font size=2>&gt; The mathematics behind the proposal are simple. You have two random </font>
<br><font size=2>&gt; numbers &quot;against&quot; a finite-length PDU.</font>
<br><font size=2>&gt; In you original proposal you have single random number (the marker) </font>
<br><font size=2>&gt; against a quasi infinite stream. The chance to miss is 1.</font>
<br><font size=2>&gt; With any decent digest (e.g. a CRC) you have a small (2**-64) chance to </font>
<br><font size=2>&gt; miss per PDU word (for every new PDU you have a new pattern). All the </font>
<br><font size=2>&gt; tricks like alignment etc do not decrease the cardinality of stream - </font>
<br><font size=2>&gt; i.e., it is still infinite and then a constant framer is no good.</font>
<br><font size=2>&gt; The cost of the new scheme is also tolerable.</font>
<br>
<br><font size=3>&nbsp;</font>
<br>
<br>
<br>
<table width=100%>
<tr valign=top>
<td width=100%><font size=2><b>Yahoo! Groups Sponsor</b></font>
<tr valign=top>
<td>
<table width=100%>
<tr valign=top>
<td width=100%><font size=1>ADVERTISEMENT</font>
<br><a href="http://rd.yahoo.com/M=221000.1882886.3382503.1261774/D=egroupweb/S=1705701014:HM/A=965714/R=0/O=1/I=brandr-promo-flowersale-alerts-lrecm/*http://shopping.yahoo.com/promotions/flowers/index.html"><font size=1 color=red face="sans-serif">[IMAGE]</font></a></table>
<br>
<tr valign=top>
<td><font size=1 color=red face="sans-serif">[IMAGE]</font></table>
<br><font size=3><tt>&nbsp;To unsubscribe from this group, send an email to:</tt></font>
<br><font size=3><tt>rdma-unsubscribe@egroups.com</tt></font>
<br>
<br><font size=3>&nbsp;</font>
<br>
<br><font size=3><tt>Your use of Yahoo! Groups is subject to the </tt></font><a href=http://docs.yahoo.com/info/terms/><font size=3 color=blue><tt><u>Yahoo! Terms of Service</u></tt></font></a><font size=3><tt>.</tt></font><font size=3> </font>
<br>
<br>
<br>
<br>
<br>
--=_alternative 0060E3E9C2256B5D_=--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 11 18:25:55 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA20255
	for <tsvwg-archive@odin.ietf.org>; Mon, 11 Feb 2002 18:25:55 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id SAA08666
	for tsvwg-archive@odin.ietf.org; Mon, 11 Feb 2002 18:25:57 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA07115;
	Mon, 11 Feb 2002 17:42:05 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA07086
	for <tsvwg@optimus.ietf.org>; Mon, 11 Feb 2002 17:42:03 -0500 (EST)
Received: from sandmail.sandburst.com (sandmail.sandburst.com [216.57.132.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA19795
	for <tsvwg@ietf.org>; Mon, 11 Feb 2002 17:42:01 -0500 (EST)
To: "John Hufferd" <hufferd@us.ibm.com>
Cc: tsvwg@ietf.org, rdma@yahoogroups.com
In-Reply-To: Message from "John Hufferd" <hufferd@us.ibm.com> 
   of "Sat, 09 Feb 2002 17:05:37 PST." <OF31443FAC.EDF1393F-ON88256B5B.006A9AA4@boulder.ibm.com> 
References: <OF31443FAC.EDF1393F-ON88256B5B.006A9AA4@boulder.ibm.com> 
Date: Mon, 11 Feb 2002 17:41:26 -0500
From: Stephen Bailey <steph@cs.uchicago.edu>
Message-Id: <20020211224132.63F364E8F@sandmail.sandburst.com>
Subject: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

John,

Thanks for the exhausting analysis.

As Paul points out, you do not want to scan the received segment to
recover synchronization with either of these techniques, since it's
too risky.  The TUF draft specifically makes this point.  Basically,
you get one event per segment.  However, if you want to be really
pessemistic you could say the stream IS segmented into tiny frames
that almost result in scanning.

The TUF draft also says that the use of a TUF-derived PDU containment
property should be discontinued after repeated failure of PDU
containment.  Given Julian's repeated statements that TUF should be
analyzed against an infinite stream, I can conclude that the draft
does not make this point strongly or clearly enough.

The TUF assumption is that there are two distinct, quasi-stable
scenarios: segmentation is preserved or it is not.  Obviously if
segmentation is not preserved, there's no point in trying to use the
PDU containment property because it will never hold.  If a receiver
detect that the PDU containment property does not hold for longer than
would represent a transient condition like PMTU change, it gives up
and stops using it.

Julian's argument that TUF needs to be analyzed against an infinite
stream is a straw man.  You can pick the threshold at which a TUF
receiver gives up to be any number that makes you comfortable (10e6
packets, 10e4, whatever).

I'm not clear on how your analysis incorporates worst case data
pattern distribution assumptions.  There is a ULP running inside the
framing protocol, and if we assume the ULP is efficient, MOST of the
data IT carries is actually controlled by the user of the ULP.  This
user data can just as well be a `search and destroy' sequence that
tries every `eye catcher' in a well-formed context.  This technique
boils down to a blind search for a 32-bit integer, and there's the
additional probability of hitting any particular element of the blind
search pattern (basically like your ex and ey).  Julian's extra 64
bits (salt & digest) are not doing anything for you in this case.
It's as simple as the probability of the resegmented stream hitting
the correct 32-bit number.

Thanks again for your work on this.

Steph

> Steph, Julian,
> The following is an attempt to determine the probability of a false
> positive, both with the TUF approach as defined in
> draft-ietf-tsvwg-tcp-ulp-frame-01.txt, and the semi draft that Julian
> proposed.
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 11 19:54:27 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA21197
	for <tsvwg-archive@odin.ietf.org>; Mon, 11 Feb 2002 19:54:26 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id TAA12071
	for tsvwg-archive@odin.ietf.org; Mon, 11 Feb 2002 19:54:28 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA10833;
	Mon, 11 Feb 2002 19:13:38 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA10802
	for <tsvwg@optimus.ietf.org>; Mon, 11 Feb 2002 19:13:36 -0500 (EST)
Received: from c003.snv.cp.net (c003-h000.c003.snv.cp.net [209.228.32.214])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id TAA20754
	for <tsvwg@ietf.org>; Mon, 11 Feb 2002 19:13:34 -0500 (EST)
Received: (cpmta 26363 invoked from network); 11 Feb 2002 16:13:04 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.214) with SMTP; 11 Feb 2002 16:13:04 -0800
X-Sent: 12 Feb 2002 00:13:04 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <tsvwg@ietf.org>
Cc: <rdma@yahoogroups.com>
Subject: RE: [Tsvwg] a proposal for a different No-Data-Touch framer for TCP
Date: Mon, 11 Feb 2002 16:15:12 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJAEKDCPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Importance: Normal
In-Reply-To: <20020211224132.63F364E8F@sandmail.sandburst.com>
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Steph,

Schemes exclusively supporting layers prior to TCP, used in conjunction with
application designators for placement, are advocating a TCP/application
transport amalgam depreciating the TCP interface paradigms.  There is an
effort at adding a shim to SCTP as a generalization of placement
capabilities.  This TCP amalgam concept introduces a new transport and
interface that should attempt application consolidation if such changes are
allowed.  Would you see a benefit in combining generalized structures found
within the SCTP shim within these various introductions of a new layer for
framing and "out of sequence" synchronization for direct placement within
TCP?

Doug

> John,
>
> Thanks for the exhausting analysis.
>
> As Paul points out, you do not want to scan the received segment to
> recover synchronization with either of these techniques, since it's
> too risky.  The TUF draft specifically makes this point.  Basically,
> you get one event per segment.  However, if you want to be really
> pessemistic you could say the stream IS segmented into tiny frames
> that almost result in scanning.
>
> The TUF draft also says that the use of a TUF-derived PDU containment
> property should be discontinued after repeated failure of PDU
> containment.  Given Julian's repeated statements that TUF should be
> analyzed against an infinite stream, I can conclude that the draft
> does not make this point strongly or clearly enough.
>
> The TUF assumption is that there are two distinct, quasi-stable
> scenarios: segmentation is preserved or it is not.  Obviously if
> segmentation is not preserved, there's no point in trying to use the
> PDU containment property because it will never hold.  If a receiver
> detect that the PDU containment property does not hold for longer than
> would represent a transient condition like PMTU change, it gives up
> and stops using it.
>
> Julian's argument that TUF needs to be analyzed against an infinite
> stream is a straw man.  You can pick the threshold at which a TUF
> receiver gives up to be any number that makes you comfortable (10e6
> packets, 10e4, whatever).
>
> I'm not clear on how your analysis incorporates worst case data
> pattern distribution assumptions.  There is a ULP running inside the
> framing protocol, and if we assume the ULP is efficient, MOST of the
> data IT carries is actually controlled by the user of the ULP.  This
> user data can just as well be a `search and destroy' sequence that
> tries every `eye catcher' in a well-formed context.  This technique
> boils down to a blind search for a 32-bit integer, and there's the
> additional probability of hitting any particular element of the blind
> search pattern (basically like your ex and ey).  Julian's extra 64
> bits (salt & digest) are not doing anything for you in this case.
> It's as simple as the probability of the resegmented stream hitting
> the correct 32-bit number.
>
> Thanks again for your work on this.
>
> Steph
>
> > Steph, Julian,
> > The following is an attempt to determine the probability of a false
> > positive, both with the TUF approach as defined in
> > draft-ietf-tsvwg-tcp-ulp-frame-01.txt, and the semi draft that Julian
> > proposed.
> >
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
>


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 11 20:21:38 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA21436
	for <tsvwg-archive@odin.ietf.org>; Mon, 11 Feb 2002 20:21:38 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA13277
	for tsvwg-archive@odin.ietf.org; Mon, 11 Feb 2002 20:21:36 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA12722;
	Mon, 11 Feb 2002 20:04:27 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA12691
	for <tsvwg@optimus.ietf.org>; Mon, 11 Feb 2002 20:04:25 -0500 (EST)
Received: from e31.co.us.ibm.com (e31.co.us.ibm.com [32.97.110.129])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA21255
	for <tsvwg@ietf.org>; Mon, 11 Feb 2002 20:04:23 -0500 (EST)
Received: from westrelay03.boulder.ibm.com (westrelay03.boulder.ibm.com [9.99.140.24])
	by e31.co.us.ibm.com (8.9.3/8.9.3) with ESMTP id UAA25912;
	Mon, 11 Feb 2002 20:01:12 -0500
Received: from d03nm014.boulder.ibm.com (avpilot.boulder.ibm.com [9.17.188.135])
	by westrelay03.boulder.ibm.com (8.11.1m3/NCO/VER6.00) with ESMTP id g1C14M1191258;
	Mon, 11 Feb 2002 18:04:22 -0700
X-Priority: 1 (High)
Importance: Normal
To: Stephen Bailey <steph@cs.uchicago.edu>
Cc: tsvwg@ietf.org, rdma@yahoogroups.com
X-Mailer: Lotus Notes Release 5.0.3 (Intl) 21 March 2000
Message-ID: <OF77785F44.528B8C96-ON88256B5E.000241D4@boulder.ibm.com>
From: "John Hufferd" <hufferd@us.ibm.com>
Date: Mon, 11 Feb 2002 17:03:27 -0800
X-MIMETrack: Serialize by Router on D03NM014/03/M/IBM(Release 5.0.9 |November 16, 2001) at
 02/11/2002 06:04:21 PM
MIME-Version: 1.0
Content-type: text/plain; charset=us-ascii
Subject: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


Steph,
How small do you think the segmentation will be, and will they generally be
on a word boundary?  The value I used for 361 or 362 (lets call it the
segmentation-point-number) was based on the possibility that the
segmentation could occur on any word boundary in a normal 1460 byte
payload.  If this segment will never get smaller then say 500 bytes, then
clearly the number will change and the probability will be even lower,  on
the other hand if the segmentation can be on a byte level, the
segmentation-point-number will be larger and the probability of false
positive would be higher.

So some guidance on the techniques of fragmentation and resultant sizes
would be useful.

I had assumed, based on what I thought I read in your document, that as
soon as you detected the Re-Segmentation, you would stop TUFing, since I
thought that the re-segmenting was likely to last for a long time.  Then,
when you thought things had reach a steady state again, you would re-probe
for the PMTU, and use that for your new  Frame Size and begin again.  If I
had that right, then, the shut down would be immediate, and only one False
Positive Event would need to be detected.  In that single event case the
probabilities I calculated would be (approximately) correct.

Now, given a single segmentation, and the other assumptions I used, we were
able to get a probability of a single occurrence.  However, if this
re-segmentation and the resulting needed detection, was to continue for
some period of time, then that would clearly change (raise) the probability
very quickly.



 .
.
 .
John L. Hufferd
Senior Technical Staff Member (STSM)
IBM/SSG San Jose Ca
Main Office (408) 256-0403, Tie: 276-0403,  eFax: (408) 904-4688
Home Office (408) 997-6136, Cell: (408) 499-9702
Internet address: hufferd@us.ibm.com


Stephen Bailey <steph@cs.uchicago.edu> on 02/11/2002 02:41:26 PM

To:    John Hufferd/San Jose/IBM@IBMUS
cc:    tsvwg@ietf.org, rdma@yahoogroups.com
Subject:    Re: [rdma] a proposal for a different No-Data-Touch framer for
       TCP



John,

Thanks for the exhausting analysis.

As Paul points out, you do not want to scan the received segment to
recover synchronization with either of these techniques, since it's
too risky.  The TUF draft specifically makes this point.  Basically,
you get one event per segment.  However, if you want to be really
pessemistic you could say the stream IS segmented into tiny frames
that almost result in scanning.

The TUF draft also says that the use of a TUF-derived PDU containment
property should be discontinued after repeated failure of PDU
containment.  Given Julian's repeated statements that TUF should be
analyzed against an infinite stream, I can conclude that the draft
does not make this point strongly or clearly enough.

The TUF assumption is that there are two distinct, quasi-stable
scenarios: segmentation is preserved or it is not.  Obviously if
segmentation is not preserved, there's no point in trying to use the
PDU containment property because it will never hold.  If a receiver
detect that the PDU containment property does not hold for longer than
would represent a transient condition like PMTU change, it gives up
and stops using it.

Julian's argument that TUF needs to be analyzed against an infinite
stream is a straw man.  You can pick the threshold at which a TUF
receiver gives up to be any number that makes you comfortable (10e6
packets, 10e4, whatever).

I'm not clear on how your analysis incorporates worst case data
pattern distribution assumptions.  There is a ULP running inside the
framing protocol, and if we assume the ULP is efficient, MOST of the
data IT carries is actually controlled by the user of the ULP.  This
user data can just as well be a `search and destroy' sequence that
tries every `eye catcher' in a well-formed context.  This technique
boils down to a blind search for a 32-bit integer, and there's the
additional probability of hitting any particular element of the blind
search pattern (basically like your ex and ey).  Julian's extra 64
bits (salt & digest) are not doing anything for you in this case.
It's as simple as the probability of the resegmented stream hitting
the correct 32-bit number.

Thanks again for your work on this.

Steph

> Steph, Julian,
> The following is an attempt to determine the probability of a false
> positive, both with the TUF approach as defined in
> draft-ietf-tsvwg-tcp-ulp-frame-01.txt, and the semi draft that Julian
> proposed.
>




_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 11 23:14:57 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA25068
	for <tsvwg-archive@odin.ietf.org>; Mon, 11 Feb 2002 23:14:56 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id XAA19549
	for tsvwg-archive@odin.ietf.org; Mon, 11 Feb 2002 23:14:58 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id XAA19360;
	Mon, 11 Feb 2002 23:02:14 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id XAA19331
	for <tsvwg@optimus.ietf.org>; Mon, 11 Feb 2002 23:02:12 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-200.dallas.net [209.44.42.200])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA24981
	for <tsvwg@ietf.org>; Mon, 11 Feb 2002 23:02:09 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id WAA08517;
	Mon, 11 Feb 2002 22:01:11 -0600
Date: Mon, 11 Feb 2002 22:01:11 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <randall@stewart.chicago.il.us>
Cc: Arun Prasad <arun@netlab.hcltech.com>,
        Coene Lode <Lode.Coene@siemens.atea.be>,
        "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "'multi6@ops.ietf.org'" <multi6@ops.ietf.org>
Subject: Re: [Tsvwg] SCTP multihoming issues draft
Message-ID: <20020211220111.B6341@openss7.org>
Reply-To: bidulock@openss7.org
References: <E76F715C0429D5118F2100508BB9EDEE0795AC@hrtades7.atea.be> <3C675B99.E6E0FBE0@netlab.hcltech.com> <3C67CCE6.51AB848A@stewart.chicago.il.us>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3C67CCE6.51AB848A@stewart.chicago.il.us>; from randall@stewart.chicago.il.us on Mon, Feb 11, 2002 at 07:53:42AM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Randall,

Gee, I think the SCTP only needs to know the IP address.
IP does the rest.

--brian

On Mon, 11 Feb 2002, Randall Stewart wrote:

> Arun:
> 
> I don't think we need to have a host become
> routing aware as section 2.2 seems to imply to
> some degree. Some clever work with the routing
> system in the host can alleviate the problem that
> Lode is mentioning here I think..
> 
> For example in our BSD work we have now the ability
> of the routing system to have multiple default routes
> and a method to lookup a route via an alternate interface.
> We probably need to expand this to actually look at the gateway
> and not the interface.. but for now we use interface :/
> 
> This allows the sctp stack to call 
> 
> rtalloc_alternate(dst,old_route);
> 
> which then trys to give a route out a different interface if
> necessary hunting up the routing tree.
> 
> With this change we then get:
> 
> A) The host only needs to know about the edge routes available
>    via its multiple default routes.
> 
> B) Full use of the multi-homing provided without injection
>    of static routes as the v4 side does now...
> 
> C) Host only knows about the default routes and possibly
>    some special routes applied by the local administrator...
> 
> R
> 
> Arun Prasad wrote:
> > 
> > hi Coene,
> >     Regarding the below section in the draft,
> >     2.2 SCTP multihoming and the size of routing tables
> >     what are we trying to achieve....   section 2.1 basically
> > explained the restrictions
> > of SCTP's multihoming feature due to existing IP architecture....  The
> > problems
> > are well sorted out, and one solution I get from that is,   changing
> > the Source
> > Ip address for the packets send ( periodically, if the acknowledgemnet
> > for it is
> > not reaching the endpoint,) from a multihomed endpoint....
> >     But regarding section 2.2,  SCTP could do nothing about it (
> > because its
> > basically IP's restriction or the restriction of the currect routing
> > architecture....)
> > SCTP defines some rules or features, using them properly is the
> > responsiblity of the
> > Application which uses it (it may the administrator of the system
> > also...) ..  This
> > each of them using it can implement their own ways to get the maximum
> > from use
> > of multihoming feature of SCTP......
> >       Will it not make SCTP more heavy, if we try to load it with more
> > and more
> > sub features like this... and it will make it more dependent on the
> > existing architecture
> > which will make the maintanance tough.....
> >     But definitely we should be spelling out the problems or
> > restrictions of the
> > multihoming feature of SCTP, with which the users of SCTP users will
> > do
> > appropriate action to use it best....
> > 
> > Is the observation valid??
> > 
> > Thanks
> > -arun
> > 
> > Coene Lode wrote:
> > 
> > > Previous week, the draft on SCTP multihoming issues was published.
> > >
> > > Apart from some editorial changes concerning author list and
> > > missing/wrong
> > > page numbers,
> > > the only major change is a sentence added at the end of paragraph
> > > 2.3 SCTP
> > > multihoming and Network Adress Translators(NAT):
> > > "It should be noted that the NAT box becomes a single
> > >     point-of-failure in this case, as ALL the paths of the SCTP
> > >     association have to go through that single NAT box."
> > > Theoretical it is possible to let all the different paths run
> > > through
> > > different NAT boxes, but then the NAT boxes would never know exactly
> > > when to
> > > teardown/terminate the association as the terminate or abort takes
> > > only one
> > > of the paths(and there are other difficulties, such as the
> > > heartbeats
> > > ...etc...) So in practical terms this sentence should be correct. If
> > > you
> > > feel otherwise, please let me know and I'll remove or add something
> > > to it to
> > > make everything clear.
> > >
> > > The draft can be found at the IETF draft directory:
> > > Multihoming issues in the Stream Control Transmission Protocol
> > > <draft-coene-sctp-multihome-02.txt>
> > >
> > > If by then end of next week, no further issues have been identified,
> > > then I
> > > would ask for this draft to be moved to informational.
> > >
> > > yours sincerly
> > > Lode Coene
> > >
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > 
> > --
> > ****************************************************************
> > V.Arun Prasad
> > HCL Technologies Ltd.
> > 51, Jawaharlal Nehru Road,
> > Ekkattuthangal,
> > Guindy Industrial Estate,
> > Chennai - 600097.
> > 
> > Contact # : 9144 - 2334174
> >             9144 - 2334181
> >             extn : 233
> > ****************************************************************
> > 
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 02:11:45 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA05307
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 02:11:45 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id CAA05118
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 02:11:45 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA04905;
	Tue, 12 Feb 2002 02:00:52 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id CAA04820
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 02:00:48 -0500 (EST)
Received: from d12lmsgate.de.ibm.com (d12lmsgate.de.ibm.com [195.212.91.199])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA28522
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 02:00:47 -0500 (EST)
Received: from d12relay02.de.ibm.com (d12relay02.de.ibm.com [9.165.215.23])
	by d12lmsgate.de.ibm.com (1.0.0) with ESMTP id IAA85972;
	Tue, 12 Feb 2002 08:00:12 +0100
Received: from d10hubm1.telaviv.ibm.com (d10ml001.telaviv.ibm.com [9.148.216.55])
	by d12relay02.de.ibm.com (8.11.1m3/NCO v5.01) with ESMTP id g1C71jD63976;
	Tue, 12 Feb 2002 08:01:46 +0100
To: Stephen Bailey <steph@cs.uchicago.edu>
Cc: "John Hufferd" <hufferd@us.ibm.com>, rdma@yahoogroups.com, tsvwg@ietf.org
MIME-Version: 1.0
X-Mailer: Lotus Notes Release 5.0.9  November 16, 2001
From: "Julian Satran" <Julian_Satran@il.ibm.com>
Message-ID: <OFA76FE5DC.23742C0E-ONC2256B5E.00255476@telaviv.ibm.com>
Date: Tue, 12 Feb 2002 09:01:42 +0200
X-MIMETrack: Serialize by Router on D10ML001/10/M/IBM(Release 5.0.8 |June 18, 2001) at
 12/02/2002 09:01:46,
	Serialize complete at 12/02/2002 09:01:46
Content-Type: multipart/alternative; boundary="=_alternative 0025FC68C2256B5E_="
Subject: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

This is a multipart message in MIME format.
--=_alternative 0025FC68C2256B5E_=
Content-Type: text/plain; charset="us-ascii"

Steph,

You fail to see that a good synchronization scheme should work with or 
without a TCP alignment scheme.
If the instream alignment can be infered by the ULP the so it can by a 
generic DDP. Alignment can somewhat improve performance
but it is not mandatory neither for a generic nor for a protocol specific 
placement scheme.  In other words - steering every TCP packet is the 
ultimate effectivenss but not necessarily the only steering scheme.  An IP 
steering scheme would work unchanged on every transport.

Regards,
Julo





Stephen Bailey <steph@cs.uchicago.edu>
12-02-02 00:41

 
        To:     "John Hufferd" <hufferd@us.ibm.com>
        cc:     tsvwg@ietf.org, rdma@yahoogroups.com
        Subject:        Re: [rdma] a proposal for a different No-Data-Touch framer for TCP

 

John,

Thanks for the exhausting analysis.

As Paul points out, you do not want to scan the received segment to
recover synchronization with either of these techniques, since it's
too risky.  The TUF draft specifically makes this point.  Basically,
you get one event per segment.  However, if you want to be really
pessemistic you could say the stream IS segmented into tiny frames
that almost result in scanning.

The TUF draft also says that the use of a TUF-derived PDU containment
property should be discontinued after repeated failure of PDU
containment.  Given Julian's repeated statements that TUF should be
analyzed against an infinite stream, I can conclude that the draft
does not make this point strongly or clearly enough.

The TUF assumption is that there are two distinct, quasi-stable
scenarios: segmentation is preserved or it is not.  Obviously if
segmentation is not preserved, there's no point in trying to use the
PDU containment property because it will never hold.  If a receiver
detect that the PDU containment property does not hold for longer than
would represent a transient condition like PMTU change, it gives up
and stops using it.

Julian's argument that TUF needs to be analyzed against an infinite
stream is a straw man.  You can pick the threshold at which a TUF
receiver gives up to be any number that makes you comfortable (10e6
packets, 10e4, whatever).

I'm not clear on how your analysis incorporates worst case data
pattern distribution assumptions.  There is a ULP running inside the
framing protocol, and if we assume the ULP is efficient, MOST of the
data IT carries is actually controlled by the user of the ULP.  This
user data can just as well be a `search and destroy' sequence that
tries every `eye catcher' in a well-formed context.  This technique
boils down to a blind search for a 32-bit integer, and there's the
additional probability of hitting any particular element of the blind
search pattern (basically like your ex and ey).  Julian's extra 64
bits (salt & digest) are not doing anything for you in this case.
It's as simple as the probability of the resegmented stream hitting
the correct 32-bit number.

Thanks again for your work on this.

Steph

> Steph, Julian,
> The following is an attempt to determine the probability of a false
> positive, both with the TUF approach as defined in
> draft-ietf-tsvwg-tcp-ulp-frame-01.txt, and the semi draft that Julian
> proposed.
> 

------------------------ Yahoo! Groups Sponsor ---------------------~-->
Sponsored by VeriSign - The Value of Trust
Do you need to encrypt all your online transactions? Find
the perfect solution in this FREE Guide from VeriSign.
http://us.click.yahoo.com/vCuuSA/UdiDAA/yigFAA/W6uqlB/TM
---------------------------------------------------------------------~->

To unsubscribe from this group, send an email to:
rdma-unsubscribe@egroups.com

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ 





--=_alternative 0025FC68C2256B5E_=
Content-Type: text/html; charset="us-ascii"


<br><font size=2 face="sans-serif">Steph,</font>
<br>
<br><font size=2 face="sans-serif">You fail to see that a good synchronization scheme should work with or without a TCP alignment scheme.</font>
<br><font size=2 face="sans-serif">If the instream alignment can be infered by the ULP the so it can by a generic DDP. Alignment can somewhat improve performance</font>
<br><font size=2 face="sans-serif">but it is not mandatory neither for a generic nor for a protocol specific placement scheme. &nbsp;In other words - steering every TCP packet is the ultimate effectivenss but not necessarily the only steering scheme. &nbsp;An IP steering scheme would work unchanged on every transport.</font>
<br>
<br><font size=2 face="sans-serif">Regards,</font>
<br><font size=2 face="sans-serif">Julo</font>
<br>
<br>
<br>
<br>
<table width=100%>
<tr valign=top>
<td>
<td><font size=1 face="sans-serif"><b>Stephen Bailey &lt;steph@cs.uchicago.edu&gt;</b></font>
<p><font size=1 face="sans-serif">12-02-02 00:41</font>
<br>
<td><font size=1 face="Arial">&nbsp; &nbsp; &nbsp; &nbsp; </font>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; To: &nbsp; &nbsp; &nbsp; &nbsp;&quot;John Hufferd&quot; &lt;hufferd@us.ibm.com&gt;</font>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; cc: &nbsp; &nbsp; &nbsp; &nbsp;tsvwg@ietf.org, rdma@yahoogroups.com</font>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; Subject: &nbsp; &nbsp; &nbsp; &nbsp;Re: [rdma] a proposal for a different No-Data-Touch framer for TCP</font>
<br>
<br><font size=1 face="Arial">&nbsp; &nbsp; &nbsp; &nbsp;</font></table>
<br>
<br><font size=2 face="Courier New">John,<br>
<br>
Thanks for the exhausting analysis.<br>
<br>
As Paul points out, you do not want to scan the received segment to<br>
recover synchronization with either of these techniques, since it's<br>
too risky. &nbsp;The TUF draft specifically makes this point. &nbsp;Basically,<br>
you get one event per segment. &nbsp;However, if you want to be really<br>
pessemistic you could say the stream IS segmented into tiny frames<br>
that almost result in scanning.<br>
<br>
The TUF draft also says that the use of a TUF-derived PDU containment<br>
property should be discontinued after repeated failure of PDU<br>
containment. &nbsp;Given Julian's repeated statements that TUF should be<br>
analyzed against an infinite stream, I can conclude that the draft<br>
does not make this point strongly or clearly enough.<br>
<br>
The TUF assumption is that there are two distinct, quasi-stable<br>
scenarios: segmentation is preserved or it is not. &nbsp;Obviously if<br>
segmentation is not preserved, there's no point in trying to use the<br>
PDU containment property because it will never hold. &nbsp;If a receiver<br>
detect that the PDU containment property does not hold for longer than<br>
would represent a transient condition like PMTU change, it gives up<br>
and stops using it.<br>
<br>
Julian's argument that TUF needs to be analyzed against an infinite<br>
stream is a straw man. &nbsp;You can pick the threshold at which a TUF<br>
receiver gives up to be any number that makes you comfortable (10e6<br>
packets, 10e4, whatever).<br>
<br>
I'm not clear on how your analysis incorporates worst case data<br>
pattern distribution assumptions. &nbsp;There is a ULP running inside the<br>
framing protocol, and if we assume the ULP is efficient, MOST of the<br>
data IT carries is actually controlled by the user of the ULP. &nbsp;This<br>
user data can just as well be a `search and destroy' sequence that<br>
tries every `eye catcher' in a well-formed context. &nbsp;This technique<br>
boils down to a blind search for a 32-bit integer, and there's the<br>
additional probability of hitting any particular element of the blind<br>
search pattern (basically like your ex and ey). &nbsp;Julian's extra 64<br>
bits (salt &amp; digest) are not doing anything for you in this case.<br>
It's as simple as the probability of the resegmented stream hitting<br>
the correct 32-bit number.<br>
<br>
Thanks again for your work on this.<br>
<br>
Steph<br>
<br>
&gt; Steph, Julian,<br>
&gt; The following is an attempt to determine the probability of a false<br>
&gt; positive, both with the TUF approach as defined in<br>
&gt; draft-ietf-tsvwg-tcp-ulp-frame-01.txt, and the semi draft that Julian<br>
&gt; proposed.<br>
&gt; <br>
<br>
------------------------ Yahoo! Groups Sponsor ---------------------~--&gt;<br>
Sponsored by VeriSign - The Value of Trust<br>
Do you need to encrypt all your online transactions? Find<br>
the perfect solution in this FREE Guide from VeriSign.<br>
http://us.click.yahoo.com/vCuuSA/UdiDAA/yigFAA/W6uqlB/TM<br>
---------------------------------------------------------------------~-&gt;<br>
<br>
To unsubscribe from this group, send an email to:<br>
rdma-unsubscribe@egroups.com<br>
<br>
 <br>
<br>
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ <br>
</font>
<br><font size=2 face="Courier New"><br>
</font>
<br>
<br>
--=_alternative 0025FC68C2256B5E_=--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 05:27:35 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA07305
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 05:27:34 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA12936
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 05:27:36 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA12603;
	Tue, 12 Feb 2002 05:15:45 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA12572
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 05:15:39 -0500 (EST)
Received: from irmgard.exp-math.uni-essen.de (irmgard.exp-math.uni-essen.de [132.252.150.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA07207
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 05:15:36 -0500 (EST)
Received: from there (biene.exp-math.uni-essen.de [132.252.150.215])
	by irmgard.exp-math.uni-essen.de (8.12.2/8.12.2) with SMTP id g1CAFKj9030134;
	Tue, 12 Feb 2002 11:15:21 +0100
Message-Id: <200202121015.g1CAFKj9030134@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain;
  charset="iso-8859-1"
From: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
Organization: University Essen
To: bidulock@openss7.org, Randall Stewart <randall@stewart.chicago.il.us>
Subject: Re: [Tsvwg] SCTP multihoming issues draft
Date: Tue, 12 Feb 2002 12:15:03 +0100
X-Mailer: KMail [version 1.3.1]
Cc: Arun Prasad <arun@netlab.hcltech.com>,
        Coene Lode <Lode.Coene@siemens.atea.be>,
        "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "'multi6@ops.ietf.org'" <multi6@ops.ietf.org>
References: <E76F715C0429D5118F2100508BB9EDEE0795AC@hrtades7.atea.be> <3C67CCE6.51AB848A@stewart.chicago.il.us> <20020211220111.B6341@openss7.org>
In-Reply-To: <20020211220111.B6341@openss7.org>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Brian, 

I think with the modifications suggested by Randy one would
obtain better performance/better failover behaviour in a typical 
network failure scenario than were otherwise the case.
(otherwise == with normal routing, one default route)

Best regards,
Andreas


On Tuesday, 12. February 2002 05:01, Brian F. G. Bidulock wrote:
> Randall,
>
> Gee, I think the SCTP only needs to know the IP address.
> IP does the rest.
>
>

> >
> > For example in our BSD work we have now the ability
> > of the routing system to have multiple default routes
> > and a method to lookup a route via an alternate interface.
> > We probably need to expand this to actually look at the gateway
> > and not the interface.. but for now we use interface :/
> >
> > This allows the sctp stack to call
> >
> > rtalloc_alternate(dst,old_route);
> >
> > which then trys to give a route out a different interface if
> > necessary hunting up the routing tree.
> >
> > With this change we then get:
> >
> > A) The host only needs to know about the edge routes available
> >    via its multiple default routes.
> >
> > B) Full use of the multi-homing provided without injection
> >    of static routes as the v4 side does now...
> >
> > C) Host only knows about the default routes and possibly
> >    some special routes applied by the local administrator...
> >
> > R
> >
> > Arun Prasad wrote:
> > > hi Coene,
> > >     Regarding the below section in the draft,
> > >     2.2 SCTP multihoming and the size of routing tables
> > >     what are we trying to achieve....   section 2.1 basically
> > > explained the restrictions
> > > of SCTP's multihoming feature due to existing IP architecture....  The
> > > problems
> > > are well sorted out, and one solution I get from that is,   changing
> > > the Source
> > > Ip address for the packets send ( periodically, if the acknowledgemnet
> > > for it is
> > > not reaching the endpoint,) from a multihomed endpoint....
> > >     But regarding section 2.2,  SCTP could do nothing about it (
> > > because its
> > > basically IP's restriction or the restriction of the currect routing
> > > architecture....)
> > > SCTP defines some rules or features, using them properly is the
> > > responsiblity of the
> > > Application which uses it (it may the administrator of the system
> > > also...) ..  This
> > > each of them using it can implement their own ways to get the maximum
> > > from use
> > > of multihoming feature of SCTP......
> > >       Will it not make SCTP more heavy, if we try to load it with more
> > > and more
> > > sub features like this... and it will make it more dependent on the
> > > existing architecture
> > > which will make the maintanance tough.....
> > >     But definitely we should be spelling out the problems or
> > > restrictions of the
> > > multihoming feature of SCTP, with which the users of SCTP users will
> > > do
> > > appropriate action to use it best....
> > >
> > > Is the observation valid??
> > >
> > > Thanks
> > > -arun
> > >
> > > Coene Lode wrote:
> > > > Previous week, the draft on SCTP multihoming issues was published.
> > > >
> > > > Apart from some editorial changes concerning author list and
> > > > missing/wrong
> > > > page numbers,
> > > > the only major change is a sentence added at the end of paragraph
> > > > 2.3 SCTP
> > > > multihoming and Network Adress Translators(NAT):
> > > > "It should be noted that the NAT box becomes a single
> > > >     point-of-failure in this case, as ALL the paths of the SCTP
> > > >     association have to go through that single NAT box."
> > > > Theoretical it is possible to let all the different paths run
> > > > through
> > > > different NAT boxes, but then the NAT boxes would never know exactly
> > > > when to
> > > > teardown/terminate the association as the terminate or abort takes
> > > > only one
> > > > of the paths(and there are other difficulties, such as the
> > > > heartbeats
> > > > ...etc...) So in practical terms this sentence should be correct. If
> > > > you
> > > > feel otherwise, please let me know and I'll remove or add something
> > > > to it to
> > > > make everything clear.
> > > >
> > > > The draft can be found at the IETF draft directory:
> > > > Multihoming issues in the Stream Control Transmission Protocol
> > > > <draft-coene-sctp-multihome-02.txt>
> > > >
> > > > If by then end of next week, no further issues have been identified,
> > > > then I
> > > > would ask for this draft to be moved to informational.
> > > >
> > > > yours sincerly
> > > > Lode Coene
> > > >
> > > > _______________________________________________
> > > > tsvwg mailing list
> > > > tsvwg@ietf.org
> > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > >
> > > --
> > > ****************************************************************
> > > V.Arun Prasad
> > > HCL Technologies Ltd.
> > > 51, Jawaharlal Nehru Road,
> > > Ekkattuthangal,
> > > Guindy Industrial Estate,
> > > Chennai - 600097.
> > >
> > > Contact # : 9144 - 2334174
> > >             9144 - 2334181
> > >             extn : 233
> > > ****************************************************************
> >
> > --
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg

- -- 
- ------------------------------------------------------------------------
Dipl.-Ing. Andreas Jungmaier              |                         _
Computer Networking Technology Group      | ASCII ribbon campaign  ( )
University of Essen                       |  - against HTML email   X
http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8aPk3XAsLBNOCSsARAileAKC6uaW7C3CzMTEktuFrtYatzSh61ACeL52k
fwb1bJG49MBAZCwj5vSU6+A=
=TNXt
-----END PGP SIGNATURE-----

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 06:27:34 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA07921
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 06:27:34 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA15688
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 06:27:34 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA14936;
	Tue, 12 Feb 2002 06:14:49 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA14908
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 06:14:47 -0500 (EST)
Received: from sj-msg-core-4.cisco.com (sj-msg-core-4.cisco.com [171.71.163.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA07735
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 06:14:44 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-4.cisco.com (8.11.3/8.9.1) with ESMTP id g1CBAht03045;
	Tue, 12 Feb 2002 03:10:43 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABB48688;
	Tue, 12 Feb 2002 03:10:41 -0800 (PST)
Message-ID: <3C68F830.2FFE9522@stewart.chicago.il.us>
Date: Tue, 12 Feb 2002 05:10:40 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
CC: bidulock@openss7.org, Arun Prasad <arun@netlab.hcltech.com>,
        Coene Lode <Lode.Coene@siemens.atea.be>,
        "Transport Area wg (E-mail)" <tsvwg@ietf.org>
Subject: Re: [Tsvwg] SCTP multihoming issues draft
References: <E76F715C0429D5118F2100508BB9EDEE0795AC@hrtades7.atea.be> <3C67CCE6.51AB848A@stewart.chicago.il.us> <20020211220111.B6341@openss7.org> <200202121015.g1CAFKj9030134@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Andreas:

What is nice about the change I will be submitting
to KAME is it provides for ANY transport to be able
to take better account of them being multi-homed.

It is not strictly a transport type change.. but a modification
to the routing system. I think the base problem is that when
BSD was built multi-homing was not considered an important
issue...

R

Andreas Jungmaier wrote:
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Brian,
> 
> I think with the modifications suggested by Randy one would
> obtain better performance/better failover behaviour in a typical
> network failure scenario than were otherwise the case.
> (otherwise == with normal routing, one default route)
> 
> Best regards,
> Andreas
> 
> On Tuesday, 12. February 2002 05:01, Brian F. G. Bidulock wrote:
> > Randall,
> >
> > Gee, I think the SCTP only needs to know the IP address.
> > IP does the rest.
> >
> >
> 
> > >
> > > For example in our BSD work we have now the ability
> > > of the routing system to have multiple default routes
> > > and a method to lookup a route via an alternate interface.
> > > We probably need to expand this to actually look at the gateway
> > > and not the interface.. but for now we use interface :/
> > >
> > > This allows the sctp stack to call
> > >
> > > rtalloc_alternate(dst,old_route);
> > >
> > > which then trys to give a route out a different interface if
> > > necessary hunting up the routing tree.
> > >
> > > With this change we then get:
> > >
> > > A) The host only needs to know about the edge routes available
> > >    via its multiple default routes.
> > >
> > > B) Full use of the multi-homing provided without injection
> > >    of static routes as the v4 side does now...
> > >
> > > C) Host only knows about the default routes and possibly
> > >    some special routes applied by the local administrator...
> > >
> > > R
> > >
> > > Arun Prasad wrote:
> > > > hi Coene,
> > > >     Regarding the below section in the draft,
> > > >     2.2 SCTP multihoming and the size of routing tables
> > > >     what are we trying to achieve....   section 2.1 basically
> > > > explained the restrictions
> > > > of SCTP's multihoming feature due to existing IP architecture....  The
> > > > problems
> > > > are well sorted out, and one solution I get from that is,   changing
> > > > the Source
> > > > Ip address for the packets send ( periodically, if the acknowledgemnet
> > > > for it is
> > > > not reaching the endpoint,) from a multihomed endpoint....
> > > >     But regarding section 2.2,  SCTP could do nothing about it (
> > > > because its
> > > > basically IP's restriction or the restriction of the currect routing
> > > > architecture....)
> > > > SCTP defines some rules or features, using them properly is the
> > > > responsiblity of the
> > > > Application which uses it (it may the administrator of the system
> > > > also...) ..  This
> > > > each of them using it can implement their own ways to get the maximum
> > > > from use
> > > > of multihoming feature of SCTP......
> > > >       Will it not make SCTP more heavy, if we try to load it with more
> > > > and more
> > > > sub features like this... and it will make it more dependent on the
> > > > existing architecture
> > > > which will make the maintanance tough.....
> > > >     But definitely we should be spelling out the problems or
> > > > restrictions of the
> > > > multihoming feature of SCTP, with which the users of SCTP users will
> > > > do
> > > > appropriate action to use it best....
> > > >
> > > > Is the observation valid??
> > > >
> > > > Thanks
> > > > -arun
> > > >
> > > > Coene Lode wrote:
> > > > > Previous week, the draft on SCTP multihoming issues was published.
> > > > >
> > > > > Apart from some editorial changes concerning author list and
> > > > > missing/wrong
> > > > > page numbers,
> > > > > the only major change is a sentence added at the end of paragraph
> > > > > 2.3 SCTP
> > > > > multihoming and Network Adress Translators(NAT):
> > > > > "It should be noted that the NAT box becomes a single
> > > > >     point-of-failure in this case, as ALL the paths of the SCTP
> > > > >     association have to go through that single NAT box."
> > > > > Theoretical it is possible to let all the different paths run
> > > > > through
> > > > > different NAT boxes, but then the NAT boxes would never know exactly
> > > > > when to
> > > > > teardown/terminate the association as the terminate or abort takes
> > > > > only one
> > > > > of the paths(and there are other difficulties, such as the
> > > > > heartbeats
> > > > > ...etc...) So in practical terms this sentence should be correct. If
> > > > > you
> > > > > feel otherwise, please let me know and I'll remove or add something
> > > > > to it to
> > > > > make everything clear.
> > > > >
> > > > > The draft can be found at the IETF draft directory:
> > > > > Multihoming issues in the Stream Control Transmission Protocol
> > > > > <draft-coene-sctp-multihome-02.txt>
> > > > >
> > > > > If by then end of next week, no further issues have been identified,
> > > > > then I
> > > > > would ask for this draft to be moved to informational.
> > > > >
> > > > > yours sincerly
> > > > > Lode Coene
> > > > >
> > > > > _______________________________________________
> > > > > tsvwg mailing list
> > > > > tsvwg@ietf.org
> > > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > >
> > > > --
> > > > ****************************************************************
> > > > V.Arun Prasad
> > > > HCL Technologies Ltd.
> > > > 51, Jawaharlal Nehru Road,
> > > > Ekkattuthangal,
> > > > Guindy Industrial Estate,
> > > > Chennai - 600097.
> > > >
> > > > Contact # : 9144 - 2334174
> > > >             9144 - 2334181
> > > >             extn : 233
> > > > ****************************************************************
> > >
> > > --
> > > Randall R. Stewart
> > > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> > >
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> - --
> - ------------------------------------------------------------------------
> Dipl.-Ing. Andreas Jungmaier              |                         _
> Computer Networking Technology Group      | ASCII ribbon campaign  ( )
> University of Essen                       |  - against HTML email   X
> http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.0.6 (GNU/Linux)
> Comment: For info see http://www.gnupg.org
> 
> iD8DBQE8aPk3XAsLBNOCSsARAileAKC6uaW7C3CzMTEktuFrtYatzSh61ACeL52k
> fwb1bJG49MBAZCwj5vSU6+A=
> =TNXt
> -----END PGP SIGNATURE-----
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 06:32:32 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA07993
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 06:32:32 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA16211
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 06:32:32 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA15301;
	Tue, 12 Feb 2002 06:21:12 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA15206
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 06:21:06 -0500 (EST)
Received: from hrtades9.atea.be (viap111.atea.be [194.78.143.111])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA07816
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 06:21:04 -0500 (EST)
Received: from hrtades10.atea.be (siemens.atea.be [139.10.143.141]) by hrtades9.atea.be with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13)
	id DF4YSNW7; Tue, 12 Feb 2002 12:20:28 +0100
Received: by siemens.atea.be with Internet Mail Service (5.5.2653.19)
	id <D6SNHRL2>; Tue, 12 Feb 2002 12:20:51 +0100
Message-ID: <E76F715C0429D5118F2100508BB9EDEE0795B7@hrtades7.atea.be>
From: Coene Lode <Lode.Coene@siemens.atea.be>
To: "'Arun Prasad'" <arun@netlab.hcltech.com>,
        Coene Lode
	 <Lode.Coene@siemens.atea.be>
Cc: "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "'multi6@ops.ietf.org'"
	 <multi6@ops.ietf.org>
Subject: RE: [Tsvwg] SCTP multihoming issues draft
Date: Tue, 12 Feb 2002 12:18:09 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hello Arun,

>-----Original Message-----
>From: Arun Prasad [mailto:arun@netlab.hcltech.com]
>Sent: maandag 11 februari 2002 6:50
>To: Coene Lode
>Cc: Transport Area wg (E-mail); 'multi6@ops.ietf.org'
>Subject: Re: [Tsvwg] SCTP multihoming issues draft
>
>
>hi Coene, 
>    Regarding the below section in the draft, 
>    2.2 SCTP multihoming and the size of routing tables 
>    what are we trying to achieve....   section 2.1 basically explained the
restrictions 
>of SCTP's multihoming feature due to existing IP architecture....  The
problems 
>are well sorted out, and one solution I get from that is,   changing the
Source 
>Ip address for the packets send ( periodically, if the acknowledgemnet for
it is 
>not reaching the endpoint,) from a multihomed endpoint.... 
>    But regarding section 2.2,  SCTP could do nothing about it ( because
its 
>basically IP's restriction or the restriction of the currect routing
architecture....) 
>SCTP defines some rules or features, using them properly is the
responsiblity of the 
>Application which uses it (it may the administrator of the system also...)
..  This 
>each of them using it can implement their own ways to get the maximum from
use 
>of multihoming feature of SCTP...... 
>      Will it not make SCTP more heavy, if we try to load it with more and
more 
>sub features like this... and it will make it more dependent on the
existing architecture 
>which will make the maintanance tough..... 

That is true. But some of the features would be far more usable if some
things(like NAT's) didn't exist in the first place. And according to the
end-to-end pricniple, it is better to put some complexity in the the
endnode, than to put it in the network: I concede that this is more a
filosofical discussion, but as the NAT section shows, such boxes are not
that helpfull.

>    But definitely we should be spelling out the problems or restrictions
of the 
>multihoming feature of SCTP, with which the users of SCTP users will do 
>appropriate action to use it best.... 
>Is the observation valid?? 

That is indeed a correct observation. 
This draft tries to spell out problems or restrictions of the multihoming
feature of SCTP.
Some of those problems can be evaded in SCTP but others need to be solved in
IP (or somewhere else). That is not my call to make. 


yours sincerly,
Lode
 
>Thanks 
>-arun 
>Coene Lode wrote: 
>The draft can be found at the IETF draft directory: 
>Multihoming issues in the Stream Control Transmission Protocol 
><draft-coene-sctp-multihome-02.txt> 
  

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 06:33:55 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA08010
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 06:33:54 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA16231
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 06:33:55 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA15238;
	Tue, 12 Feb 2002 06:21:08 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA15191
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 06:21:05 -0500 (EST)
Received: from hrtades9.atea.be (viap111.atea.be [194.78.143.111])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA07813;
	Tue, 12 Feb 2002 06:21:03 -0500 (EST)
Received: from hrtades10.atea.be (siemens.atea.be [139.10.143.141]) by hrtades9.atea.be with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13)
	id DF4YSNW5; Tue, 12 Feb 2002 12:20:23 +0100
Received: by siemens.atea.be with Internet Mail Service (5.5.2653.19)
	id <D6SNHRLF>; Tue, 12 Feb 2002 12:20:46 +0100
Message-ID: <E76F715C0429D5118F2100508BB9EDEE0795B6@hrtades7.atea.be>
From: Coene Lode <Lode.Coene@siemens.atea.be>
To: "'Pekka Savola'" <pekkas@netcore.fi>,
        Coene Lode
	 <Lode.Coene@siemens.atea.be>
Cc: "Andreas Jungmayer (E-mail)" <ajung@exp-math.uni-essen.de>,
        "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)"
	 <sigtran@ietf.org>
Date: Tue, 12 Feb 2002 12:18:04 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: [Tsvwg] RE: [Sigtran] RE: SCTP multihoming issues draft
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hello Pekka,

Sorry for the delay(some minor internal system failures:-) 

> -----Original Message-----
> From: Pekka Savola [mailto:pekkas@netcore.fi]
> Sent: vrijdag 8 februari 2002 16:01
> To: Coene Lode
> Cc: Andreas Jungmayer (E-mail); Transport Area wg (E-mail); SIGTRAN
> (E-mail); 'multi6@ops.ietf.org'
> Subject: [Sigtran] RE: SCTP multihoming issues draft
> 
> > > ==> on a generic note: it seems to me there is a need for routing
> > > tables on hosts only if one wants to optimize the paths: 
> one is used
> > > by default, and if it fails, just (try to) switch to the next one,
> > > right?
> > > 
> > 
> > See previous remark on routing tables.
> 
> See my reply to Andreas. (note to others: multi6 was stripped 
> from Cc:).
> 
> In short, routing table is naturally required, but the draft 
> talks about 
> the need to "download" a fuller table from edge routers on 
> hosts.  IMO 
> this is unnecessary.

the text is indeed too broad(my fault). It should be saying something
about"distributing some prefixes from the edge routers to the host". That is
far nearer to the truth.
But my favourite solution to this should be something that does NOT need "a
download" or "receiving some prefixes"(in general any solution needing to
bugger the adjacend routers) on the host. As such Randy's BSD implementation
is clearly good step in that direction.(as I could infer from his
explanation, correct me if I am wrong)

yours sincerly,
Lode

> 
> -- 
> Pekka Savola                 "Tell me of difficulties surmounted,
> Netcore Oy                   not those you stumble over and fall"
> Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords
> 
> 
> 
> 
> _______________________________________________
> Sigtran mailing list
> Sigtran@ietf.org
> https://www1.ietf.org/mailman/listinfo/sigtran
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 06:35:15 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA08034
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 06:35:15 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA16255
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 06:35:16 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA15443;
	Tue, 12 Feb 2002 06:21:23 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA15310
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 06:21:14 -0500 (EST)
Received: from hrtades9.atea.be (viap111.atea.be [194.78.143.111])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA07824
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 06:21:12 -0500 (EST)
Received: from hrtades10.atea.be (siemens.atea.be [139.10.143.141]) by hrtades9.atea.be with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13)
	id DF4YSNXA; Tue, 12 Feb 2002 12:20:36 +0100
Received: by siemens.atea.be with Internet Mail Service (5.5.2653.19)
	id <D6SNHRLM>; Tue, 12 Feb 2002 12:20:59 +0100
Message-ID: <E76F715C0429D5118F2100508BB9EDEE0795B8@hrtades7.atea.be>
From: Coene Lode <Lode.Coene@siemens.atea.be>
To: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        Arun Prasad
	 <arun@netlab.hcltech.com>
Cc: Coene Lode <Lode.Coene@siemens.atea.be>,
        "Transport Area wg (E-mail)"
	 <tsvwg@ietf.org>,
        "'multi6@ops.ietf.org'" <multi6@ops.ietf.org>
Subject: RE: [Tsvwg] SCTP multihoming issues draft
Date: Tue, 12 Feb 2002 12:18:16 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hello Randy,

> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: maandag 11 februari 2002 14:54
> To: Arun Prasad
> Cc: Coene Lode; Transport Area wg (E-mail); 'multi6@ops.ietf.org'
> Subject: Re: [Tsvwg] SCTP multihoming issues draft
> 
> 
> Arun:
> 
> I don't think we need to have a host become
> routing aware as section 2.2 seems to imply to
> some degree. Some clever work with the routing
> system in the host can alleviate the problem that
> Lode is mentioning here I think..

Maybe the section is not clear enough. 
And a bit biased too. At least, I need to rephrase "downloading the routing
tables" sentences.  Stay tuned for the next iteration.

> 
> For example in our BSD work we have now the ability
> of the routing system to have multiple default routes
> and a method to lookup a route via an alternate interface.
> We probably need to expand this to actually look at the gateway
> and not the interface.. but for now we use interface :/
> 
> This allows the sctp stack to call 
> 
> rtalloc_alternate(dst,old_route);
> 
> which then trys to give a route out a different interface if
> necessary hunting up the routing tree.
> 
> With this change we then get:
> 
> A) The host only needs to know about the edge routes available
>    via its multiple default routes.
> 
> B) Full use of the multi-homing provided without injection
>    of static routes as the v4 side does now...
> 
> C) Host only knows about the default routes and possibly
>    some special routes applied by the local administrator...
> 
> R


yours sincerly,
Lode

> 
> Arun Prasad wrote:
> > 
> > hi Coene,
> >     Regarding the below section in the draft,
> >     2.2 SCTP multihoming and the size of routing tables
> >     what are we trying to achieve....   section 2.1 basically
> > explained the restrictions
> > of SCTP's multihoming feature due to existing IP 
> architecture....  The
> > problems
> > are well sorted out, and one solution I get from that is,   changing
> > the Source
> > Ip address for the packets send ( periodically, if the 
> acknowledgemnet
> > for it is
> > not reaching the endpoint,) from a multihomed endpoint....
> >     But regarding section 2.2,  SCTP could do nothing about it (
> > because its
> > basically IP's restriction or the restriction of the currect routing
> > architecture....)
> > SCTP defines some rules or features, using them properly is the
> > responsiblity of the
> > Application which uses it (it may the administrator of the system
> > also...) ..  This
> > each of them using it can implement their own ways to get 
> the maximum
> > from use
> > of multihoming feature of SCTP......
> >       Will it not make SCTP more heavy, if we try to load 
> it with more
> > and more
> > sub features like this... and it will make it more dependent on the
> > existing architecture
> > which will make the maintanance tough.....
> >     But definitely we should be spelling out the problems or
> > restrictions of the
> > multihoming feature of SCTP, with which the users of SCTP users will
> > do
> > appropriate action to use it best....
> > 
> > Is the observation valid??
> > 
> > Thanks
> > -arun
> > 
> > Coene Lode wrote:
> > 
> > > Previous week, the draft on SCTP multihoming issues was published.
> > >
> > > Apart from some editorial changes concerning author list and
> > > missing/wrong
> > > page numbers,
> > > the only major change is a sentence added at the end of paragraph
> > > 2.3 SCTP
> > > multihoming and Network Adress Translators(NAT):
> > > "It should be noted that the NAT box becomes a single
> > >     point-of-failure in this case, as ALL the paths of the SCTP
> > >     association have to go through that single NAT box."
> > > Theoretical it is possible to let all the different paths run
> > > through
> > > different NAT boxes, but then the NAT boxes would never 
> know exactly
> > > when to
> > > teardown/terminate the association as the terminate or abort takes
> > > only one
> > > of the paths(and there are other difficulties, such as the
> > > heartbeats
> > > ...etc...) So in practical terms this sentence should be 
> correct. If
> > > you
> > > feel otherwise, please let me know and I'll remove or add 
> something
> > > to it to
> > > make everything clear.
> > >
> > > The draft can be found at the IETF draft directory:
> > > Multihoming issues in the Stream Control Transmission Protocol
> > > <draft-coene-sctp-multihome-02.txt>
> > >
> > > If by then end of next week, no further issues have been 
> identified,
> > > then I
> > > would ask for this draft to be moved to informational.
> > >
> > > yours sincerly
> > > Lode Coene
> > >
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > 
> > --
> > ****************************************************************
> > V.Arun Prasad
> > HCL Technologies Ltd.
> > 51, Jawaharlal Nehru Road,
> > Ekkattuthangal,
> > Guindy Industrial Estate,
> > Chennai - 600097.
> > 
> > Contact # : 9144 - 2334174
> >             9144 - 2334181
> >             extn : 233
> > ****************************************************************
> > 
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 08:05:33 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA10102
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 08:05:28 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA20149
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 08:05:29 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA19433;
	Tue, 12 Feb 2002 07:48:52 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA19402
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 07:48:44 -0500 (EST)
Received: from c003.snv.cp.net (c003-h004.c003.snv.cp.net [209.228.32.218])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id HAA09794
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 07:48:22 -0500 (EST)
Received: (cpmta 8952 invoked from network); 12 Feb 2002 04:46:23 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.218) with SMTP; 12 Feb 2002 04:46:23 -0800
X-Sent: 12 Feb 2002 12:46:23 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <tsvwg@ietf.org>
Cc: <rdma@yahoogroups.com>
Subject: RE: [Tsvwg]  a proposal for a different No-Data-Touch framer for TCP
Date: Tue, 12 Feb 2002 04:48:33 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJGEKHCPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
In-Reply-To: <OFA76FE5DC.23742C0E-ONC2256B5E.00255476@telaviv.ibm.com>
Importance: Normal
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Julian,

An application using DDP must continuously communicate with a steering layer
added to the transport.  Placement is an intimate aspect of transport
function and the transport serves as the interface to the application.  It
is
inappropriate to designate this as an IP steering scheme, however a scheme
incorporated into TCP can offer common conventions and structures.

Doug

> Steph,
>
> You fail to see that a good synchronization scheme should work with or
without a
> TCP alignment scheme.  If the instream alignment can be infered by the ULP
the so
> it can by a generic DDP. Alignment can somewhat improve performance but it
is not
> mandatory neither for a generic nor for a protocol specific placement
scheme.  In
> other words - steering every TCP packet is the ultimate effectivenss but
not
> necessarily the only steering scheme.  An IP steering scheme would work
unchanged
> on every transport.
>
> Regards,
> Julo


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 11:58:09 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA18382
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 11:58:09 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA02013
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 11:58:11 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA00384;
	Tue, 12 Feb 2002 11:30:21 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA00278
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 11:30:15 -0500 (EST)
Received: from eins.siemens.at (eins.siemens.at [193.81.246.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA17252
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 11:30:12 -0500 (EST)
Received: from scesie13.sie.siemens.at (forix [10.1.140.2])
	by eins.siemens.at  with ESMTP id g1CGThp22494
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 17:29:43 +0100
Received: from vies194a.sie.siemens.at (atws15tc.sie.siemens.at [158.226.135.41])
	by scesie13.sie.siemens.at (8.12.1/8.12.1) with ESMTP id g1CGTgkv015943
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 17:29:43 +0100 (MET)
Received: by vies194a.sie.siemens.at with Internet Mail Service (5.5.2653.19)
	id <1XS6QKNS>; Tue, 12 Feb 2002 17:29:40 +0100
Message-ID: <D9F2B9CD7BD5D21196BC0800060D9ED603E91F43@vies186a.sie.siemens.at>
From: Benedict Rainer <rainer.benedict@siemens.at>
To: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Date: Tue, 12 Feb 2002 17:29:39 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C1B3E2.778D8A30"
Subject: [Tsvwg] fast retransmit in SCTP Implementors Guide
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01C1B3E2.778D8A30
Content-Type: text/plain;
	charset="iso-8859-1"

Hello SCTP experts,

when checking our current SCTP implementation for compatibility with the
implementors guide some questions concerning fast retransmit or SACK
processing arose:

What happens, when a SACK contains no 'HighestTSNinSack', i.e. it doesnt ack
new data, but the receiver reneges some chunks previously acked by not
including them in the SACK?

According to M2 I would not qualify the reneged chunks as 'Stray DATA
chunks',
but I guess the actions 6.2.1 D) iii) (mark as missing, start T3 if not
running, see below) must be taken.
Otherwise a deadlock situation could occur, because for the sender the
reneged chunks are already acked and are not subject for retransmission,
neither after timeout nor for fast retransmit.

In many situation this would probably clear if RWND grows by such an amount,
that the receiver can send 4 SACK that ack new chunks and so the transmitter
can  use all the information from the sack, but I'm not sure if this is
always true.

What was the reason for introducing M1-M5, especially 'HighestTSNinSack'?

Any comments welcome

best regards
Rainer

SCTP Implementors Guide,
   M2) Each time a SACK arrives reporting 'Stray DATA chunk(s)' record
       the highest TSN reported as newly acknowledged, call this
       value 'HighestTSNinSack'. A newly acknowledged DATA chunk is one
       not previously acknowledged in a SACK.

       When the SCTP sender of data receives a SACK chunk that
       acknowledges, for the first time, the receipt of a DATA chunk,
       all the still unacknowledged DATA chunks whose TSN is older than
       that newly acknowledged DATA chunk, are qualified as
       'Stray DATA chunks'.

RFC2960,
6.2.1 D) Any time a SACK arrives, the endpoint performs the following
         ...
         iii) If the SACK is missing a TSN that was previously
         acknowledged via a Gap Ack Block (e.g., the data receiver
         reneged on the data), then mark the corresponding DATA chunk as
         available for retransmit:  Mark it as missing for fast
         retransmit as described in Section 7.2.4 and if no retransmit
         timer is running for the destination address to which the DATA
         chunk was originally transmitted, then T3-rtx is started for
         that destination address.


------_=_NextPart_001_01C1B3E2.778D8A30
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: base64
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------_=_NextPart_001_01C1B3E2.778D8A30--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 13:00:04 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA20284
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 13:00:04 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA06006
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 13:00:04 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA05579;
	Tue, 12 Feb 2002 12:45:54 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA05526
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 12:45:48 -0500 (EST)
Received: from calliope1.fm.intel.com (fmfdns01.fm.intel.com [132.233.247.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA19968
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 12:45:46 -0500 (EST)
Received: from fmsmsxvs042.fm.intel.com (fmsmsxv042-1.fm.intel.com [132.233.48.110])
	by calliope1.fm.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.50 2002/02/08 23:45:02 root Exp $) with SMTP id RAA23914
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 17:45:46 GMT
Received: from fmsmsx28.fm.intel.com ([132.233.42.28])
 by fmsmsxvs042.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002021209482605312
 for <tsvwg@ietf.org>; Tue, 12 Feb 2002 09:48:26 -0800
Received: by fmsmsx28.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <1YPAPB7S>; Tue, 12 Feb 2002 09:45:46 -0800
Message-ID: <F1CE15E08172D4119247009027AE9D500B5ECF45@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Date: Tue, 12 Feb 2002 09:45:43 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Subject: [Tsvwg] SCTP Implementer's guide question...
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi all, 
  I have a question regarding SCTP implementer's guide. I remember this
issue was discussed during 2nd Bakeoff. 

In http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt
section 5.2.2. We are describing the case of association RESTART when INIT
is received in any state other than CLOSED, COOKIE-ECHOED, COOKIE-WAIT and
SHUTDOWN-ACK-SENT. The text says 

 " Before responding the endpoint MUST check to see if the
   unexpected INIT adds new addresses to the association. If new
   addresses are added to the association, the endpoint MUST respond
   with an ABORT copying the 'Initiation Tag' of the unexpected INIT
   into the 'Verification Tag' of the outbound packet carrying the ABORT. " 

I think that we should check the additional IP addresses when we receive
COOKIE-ECHO from the peer not when we receive INIT from peer, because if we
send ABORT when we receive INIT then an *INTRUDER* can break an ESTABLISHED
association by sending INIT on an existing association with addition of any
IP addresses.


Correct me here please . 

Thanks
sandeep 


 



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 13:17:05 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA20911
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 13:17:04 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA06723
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 13:17:05 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA06464;
	Tue, 12 Feb 2002 13:03:31 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA06433
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 13:03:30 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-248.dallas.net [209.44.42.248])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA20343
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 13:03:27 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id MAA20630;
	Tue, 12 Feb 2002 12:03:19 -0600
Date: Tue, 12 Feb 2002 12:03:19 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: "Balani, Sandeep" <s_balani@trillium.com>
Cc: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Subject: Re: [Tsvwg] SCTP Implementer's guide question...
Message-ID: <20020212120319.P6341@openss7.org>
Reply-To: bidulock@openss7.org
References: <F1CE15E08172D4119247009027AE9D500B5ECF45@FMSMSX37>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <F1CE15E08172D4119247009027AE9D500B5ECF45@FMSMSX37>; from s_balani@trillium.com on Tue, Feb 12, 2002 at 09:45:43AM -0800
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Sandeep,

For an intruder to abort the association, the intruder would also
have to place the verification tag of the existing association in
the INIT.  That is rather difficult.

--brian

On Tue, 12 Feb 2002, Balani, Sandeep wrote:

> Hi all, 
>   I have a question regarding SCTP implementer's guide. I remember this
> issue was discussed during 2nd Bakeoff. 
> 
> In http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt
> section 5.2.2. We are describing the case of association RESTART when INIT
> is received in any state other than CLOSED, COOKIE-ECHOED, COOKIE-WAIT and
> SHUTDOWN-ACK-SENT. The text says 
> 
>  " Before responding the endpoint MUST check to see if the
>    unexpected INIT adds new addresses to the association. If new
>    addresses are added to the association, the endpoint MUST respond
>    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
>    into the 'Verification Tag' of the outbound packet carrying the ABORT. " 
> 
> I think that we should check the additional IP addresses when we receive
> COOKIE-ECHO from the peer not when we receive INIT from peer, because if we
> send ABORT when we receive INIT then an *INTRUDER* can break an ESTABLISHED
> association by sending INIT on an existing association with addition of any
> IP addresses.
> 
> 
> Correct me here please . 
> 
> Thanks
> sandeep 
> 

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 13:52:50 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA21883
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 13:52:49 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA08122
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 13:52:51 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA07745;
	Tue, 12 Feb 2002 13:34:48 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA07720
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 13:34:46 -0500 (EST)
Received: from thalia.fm.intel.com (fmfdns02.fm.intel.com [132.233.247.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA21502
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 13:34:44 -0500 (EST)
Received: from fmsmsxvs042.fm.intel.com (fmsmsxv042-1.fm.intel.com [132.233.48.110])
	by thalia.fm.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.50 2002/02/08 23:45:02 root Exp $) with SMTP id SAA10965
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 18:34:43 GMT
Received: from fmsmsx28.fm.intel.com ([132.233.42.28])
 by fmsmsxvs042.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002021210372318704
 ; Tue, 12 Feb 2002 10:37:23 -0800
Received: by fmsmsx28.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <1YPAPD5W>; Tue, 12 Feb 2002 10:34:43 -0800
Message-ID: <F1CE15E08172D4119247009027AE9D500B5ECF46@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'bidulock@openss7.org'" <bidulock@openss7.org>
Cc: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Subject: RE: [Tsvwg] SCTP Implementer's guide question...
Date: Tue, 12 Feb 2002 10:34:36 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id NAA07721
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi Brian, 

Should'nt INIT will have ( vTag == 0 )  ?  

So INTRUDER, only needs correct set of addresses & port numbers to mess up
an association. 
 

-sandeep


-----Original Message-----
From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
Sent: Tuesday, February 12, 2002 10:03 AM
To: Balani, Sandeep
Cc: 'tsvwg@ietf.org'
Subject: Re: [Tsvwg] SCTP Implementer's guide question...


Sandeep,

For an intruder to abort the association, the intruder would also
have to place the verification tag of the existing association in
the INIT.  That is rather difficult.

--brian

On Tue, 12 Feb 2002, Balani, Sandeep wrote:

> Hi all, 
>   I have a question regarding SCTP implementer's guide. I remember this
> issue was discussed during 2nd Bakeoff. 
> 
> In
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt
> section 5.2.2. We are describing the case of association RESTART when INIT
> is received in any state other than CLOSED, COOKIE-ECHOED, COOKIE-WAIT and
> SHUTDOWN-ACK-SENT. The text says 
> 
>  " Before responding the endpoint MUST check to see if the
>    unexpected INIT adds new addresses to the association. If new
>    addresses are added to the association, the endpoint MUST respond
>    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
>    into the 'Verification Tag' of the outbound packet carrying the ABORT.
" 
> 
> I think that we should check the additional IP addresses when we receive
> COOKIE-ECHO from the peer not when we receive INIT from peer, because if
we
> send ABORT when we receive INIT then an *INTRUDER* can break an
ESTABLISHED
> association by sending INIT on an existing association with addition of
any
> IP addresses.
> 
> 
> Correct me here please . 
> 
> Thanks
> sandeep 
> 

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 14:37:09 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA23274
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 14:37:09 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA10555
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 14:37:11 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA09581;
	Tue, 12 Feb 2002 14:15:50 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA09554
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 14:15:48 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-248.dallas.net [209.44.42.248])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA22630
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 14:15:40 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id NAA21936;
	Tue, 12 Feb 2002 13:15:34 -0600
Date: Tue, 12 Feb 2002 13:15:34 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: "Balani, Sandeep" <s_balani@trillium.com>
Cc: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Subject: Re: [Tsvwg] SCTP Implementer's guide question...
Message-ID: <20020212131534.A21856@openss7.org>
Reply-To: bidulock@openss7.org
References: <F1CE15E08172D4119247009027AE9D500B5ECF46@FMSMSX37>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <F1CE15E08172D4119247009027AE9D500B5ECF46@FMSMSX37>; from s_balani@trillium.com on Tue, Feb 12, 2002 at 10:34:36AM -0800
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Sandeep,

> >  " Before responding the endpoint MUST check to see if the
> >    unexpected INIT adds new addresses to the association. If new
> >    addresses are added to the association, the endpoint MUST respond
> >    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> >    into the 'Verification Tag' of the outbound packet carrying the ABORT.
       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

The sender of the INIT tag must guess the verification tag of the
existing association and place it in the Initiation Tag field of
the INIT.  It is of no moment that the vTag is zero in the INIT.
The ABORT will arrive with the T-bit clear and a vTag that corresponds
to the Initiate Tag from the INIT.  If the vTag in the ABORT does
not match the local tag, the receiver ignores the ABORT.

So your intruder must guess the verification tag associated with the
existing association to perform this attack.  If the verification tag
is known or discernable, there are easier ways of aborting an existing
association (just send ABORT).

--brian


On Tue, 12 Feb 2002, Balani, Sandeep wrote:

> Hi Brian, 
> 
> Should'nt INIT will have ( vTag == 0 )  ?  
> 
> So INTRUDER, only needs correct set of addresses & port numbers to mess up
> an association. 
>  
> 
> -sandeep
> 
> 
> -----Original Message-----
> From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> Sent: Tuesday, February 12, 2002 10:03 AM
> To: Balani, Sandeep
> Cc: 'tsvwg@ietf.org'
> Subject: Re: [Tsvwg] SCTP Implementer's guide question...
> 
> 
> Sandeep,
> 
> For an intruder to abort the association, the intruder would also
> have to place the verification tag of the existing association in
> the INIT.  That is rather difficult.
> 
> --brian
> 
> On Tue, 12 Feb 2002, Balani, Sandeep wrote:
> 
> > Hi all, 
> >   I have a question regarding SCTP implementer's guide. I remember this
> > issue was discussed during 2nd Bakeoff. 
> > 
> > In
> http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt
> > section 5.2.2. We are describing the case of association RESTART when INIT
> > is received in any state other than CLOSED, COOKIE-ECHOED, COOKIE-WAIT and
> > SHUTDOWN-ACK-SENT. The text says 
> > 
> >  " Before responding the endpoint MUST check to see if the
> >    unexpected INIT adds new addresses to the association. If new
> >    addresses are added to the association, the endpoint MUST respond
> >    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
> >    into the 'Verification Tag' of the outbound packet carrying the ABORT.
> " 
> > 
> > I think that we should check the additional IP addresses when we receive
> > COOKIE-ECHO from the peer not when we receive INIT from peer, because if
> we
> > send ABORT when we receive INIT then an *INTRUDER* can break an
> ESTABLISHED
> > association by sending INIT on an existing association with addition of
> any
> > IP addresses.
> > 
> > 
> > Correct me here please . 
> > 
> > Thanks
> > sandeep 
> > 
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 15:38:39 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA25786
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 15:38:39 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA12977
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 15:38:41 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA12077;
	Tue, 12 Feb 2002 15:19:54 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA12048
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 15:19:52 -0500 (EST)
Received: from sandmail.sandburst.com (sandmail.sandburst.com [216.57.132.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA24876
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 15:19:49 -0500 (EST)
To: "Julian Satran" <Julian_Satran@il.ibm.com>
Cc: rdma@yahoogroups.com, tsvwg@ietf.org
In-Reply-To: Message from "Julian Satran" <Julian_Satran@il.ibm.com> 
   of "Tue, 12 Feb 2002 09:01:42 +0200." <OFA76FE5DC.23742C0E-ONC2256B5E.00255476@telaviv.ibm.com> 
References: <OFA76FE5DC.23742C0E-ONC2256B5E.00255476@telaviv.ibm.com> 
Date: Tue, 12 Feb 2002 15:19:13 -0500
From: Stephen Bailey <steph@cs.uchicago.edu>
Message-Id: <20020212201920.657244E7A@sandmail.sandburst.com>
Subject: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

> You fail to see that a good synchronization scheme should work with or 
> without a TCP alignment scheme.

I see it clearly.  However, this seems to be a case where we can't
have everything.  If we want to touch the data we can have a
synchronization technique which can also be used to determine PDU
alignment (COWS).  If we don't want to touch the data, we can have PDU
alignment (TUF or NDT) OR synchronization (FIM), but not both.

The TUF draft specifically says an implementation must not use the TUF
in-stream information to recover synchronization in an arbitrary
stream.  It can only be used to verify PDU containment.  That means
there is only one toss of the dice per PDU rather than PDU length
worth.

If you are proposing that your No-Data-Touch framer CAN be used this
way, I strongly disagree.  You drastically reduce the safety of the
scheme if you do this.

Steph



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 15:54:13 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA26657
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 15:54:12 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA13234
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 15:54:15 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA12881;
	Tue, 12 Feb 2002 15:36:16 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA12852
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 15:36:14 -0500 (EST)
Received: from sandmail.sandburst.com (sandmail.sandburst.com [216.57.132.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA25671
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 15:36:11 -0500 (EST)
To: "John Hufferd" <hufferd@us.ibm.com>
Cc: tsvwg@ietf.org
In-Reply-To: Message from "John Hufferd" <hufferd@us.ibm.com> 
   of "Mon, 11 Feb 2002 17:03:27 PST." <OF77785F44.528B8C96-ON88256B5E.000241D4@boulder.ibm.com> 
References: <OF77785F44.528B8C96-ON88256B5E.000241D4@boulder.ibm.com> 
Date: Tue, 12 Feb 2002 15:35:33 -0500
From: Stephen Bailey <steph@cs.uchicago.edu>
Message-Id: <20020212203542.15ACD4E7A@sandmail.sandburst.com>
Subject: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch framer for TCP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

> I had assumed, based on what I thought I read in your document, that as
> soon as you detected the Re-Segmentation, you would stop TUFing, since I
> thought that the re-segmenting was likely to last for a long time.

Oh.  I guess I missed this in your analysis.  Sorry.

So, you probably got it roughly right then.  The key point is that you
get one toss per frame.  You probably would not immediately stop
TUFing after a single burp, but after a relatively small number, which
would multiplicatively increase the chances of a false positive, but
that's why you have to pick a SMALL number.

10-17 sounds fine to me.

The other point to note is that you only play the dice at all if
resegmentation DOES result from an intermediary, since the source will
only send segments which are known not to fail false-positively.  This
technique applies equally to Julian's proposal.

Thanks again.

Steph


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 16:35:19 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA28283
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 16:35:19 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA15193
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 16:35:22 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA14117;
	Tue, 12 Feb 2002 16:11:23 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA14087
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 16:11:18 -0500 (EST)
Received: from mail.eecis.udel.edu (louie.udel.edu [128.4.40.12])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id QAA27469
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 16:11:15 -0500 (EST)
Received: from ren.eecis.udel.edu by mail.eecis.udel.edu id aa00754;
          12 Feb 2002 16:10 EST
Date: Tue, 12 Feb 2002 16:10:26 -0500 (EST)
From: Janardhan Iyengar <iyengar@mail.eecis.udel.edu>
To: Benedict Rainer <rainer.benedict@siemens.at>
cc: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Subject: Re: [Tsvwg] fast retransmit in SCTP Implementors Guide
In-Reply-To: <D9F2B9CD7BD5D21196BC0800060D9ED603E91F43@vies186a.sie.siemens.at>
Message-ID: <Pine.GSO.4.33.0202121538460.29285-100000@ren.eecis.udel.edu>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi Rainer,

> What happens, when a SACK contains no 'HighestTSNinSack', i.e. it
> doesnt ack new data, but the receiver reneges some chunks previously
> acked by not including them in the SACK?

As per the RFC/Imp Guide, the TSN should get marked for FR after 4
missing reports. Consider the following case:

1 ------>
2 ------>
3 ---X
4 ------>
5 ------>

1 <-----
2 <-----
2 (4-4) <-----
2 (4-5) <-----

At this point, 3 has been struck twice. Now the receiver renegs on TSN4

2 (5-5) <----

Now, TSN4 gets struck once. 3 more SACKs striking TSN4 will cause FR.

>
> According to M2 I would not qualify the reneged chunks as 'Stray DATA
> chunks', but I guess the actions 6.2.1 D) iii) (mark as missing, start
> T3 if not running, see below) must be taken. Otherwise a deadlock
> situation could occur, because for the sender the reneged chunks are
> already acked and are not subject for retransmission, neither after
> timeout nor for fast retransmit.
>
> In many situation this would probably clear if RWND grows by such an
> amount, that the receiver can send 4 SACK that ack new chunks and so
> the transmitter can use all the information from the sack, but I'm not
> sure if this is always true.
>
> What was the reason for introducing M1-M5, especially
> 'HighestTSNinSack'?

AFAIK, the FR algo in the Implementor's guide protects from a few out of
order TSNs from incorrectly causing FR. Consider:
1 --->
2 --->
...
8 --->

Suppose that somehow 8 got ahead of all the TSNs, then according to the FR
algo, TSN 8 would cause a SACK of 0(8-8), TSN 1 through 7 would cause
SACKs of the form 1(8-8), 2(8-8), 3(8-8)... "HighestTSNinSACK" for these
SACKs will be 1, 2, 3 respectively.

In such a case, using the RFC's FR algo, 3(8-8) would cause an FR of TSNs
4-7, but using the Implementor guide's FR algo, no FR would occur.
HighestTSNinSACK helps implement this modification

I don't know if I answered your question completely.. hope that clarifies
things though.

regards,
jana

-----------------------------------------------------------------------
Janardhan R. Iyengar				   iyengar@cis.udel.edu
University of Delaware		       http://www.cis.udel.edu/~iyengar
-----------------------------------------------------------------------



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 20:04:34 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA02150
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 20:04:34 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA24194
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 20:04:35 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA23352;
	Tue, 12 Feb 2002 19:52:40 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA23323
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 19:52:38 -0500 (EST)
Received: from motgate.mot.com (motgate.mot.com [129.188.136.100])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA01920
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 19:52:36 -0500 (EST)
Received: [from pobox.mot.com (pobox.mot.com [129.188.137.100]) by motgate.mot.com (motgate 2.1) with ESMTP id RAA10218; Tue, 12 Feb 2002 17:52:37 -0700 (MST)]
Received: [from relay2.cig.mot.com (relay2.cig.mot.com [136.182.15.24]) by pobox.mot.com (MOT-pobox 2.0) with ESMTP id RAA06376; Tue, 12 Feb 2002 17:52:37 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay2.cig.mot.com (8.11.4/8.11.4) with ESMTP id g1D0qUr09435;
	Tue, 12 Feb 2002 18:52:30 -0600 (CST)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id SAA05389;
	Tue, 12 Feb 2002 18:57:24 -0600 (CST)
Message-ID: <3C69B956.BA0F5687@email.mot.com>
Date: Tue, 12 Feb 2002 18:54:46 -0600
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: bidulock@openss7.org
CC: "Balani, Sandeep" <s_balani@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP Implementer's guide question...
References: <F1CE15E08172D4119247009027AE9D500B5ECF46@FMSMSX37> <20020212131534.A21856@openss7.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Brian/Sandeep:

I believe Sandeep is right. The impl-guide should be corrected. See
below for more...

"Brian F. G. Bidulock" wrote:
> 
> Sandeep,
> 
> > >  " Before responding the endpoint MUST check to see if the
> > >    unexpected INIT adds new addresses to the association. If new
> > >    addresses are added to the association, the endpoint MUST respond
> > >    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
>        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > >    into the 'Verification Tag' of the outbound packet carrying the ABORT.
>        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> 
> The sender of the INIT tag must guess the verification tag of the
> existing association and place it in the Initiation Tag field of
> the INIT.  It is of no moment that the vTag is zero in the INIT.

No. vTag is always zero in the INIT, no matter it's for a fresh
association or a re-started one. The endpoint that restarts would most
likely lose any memory of its previous life.

.....
> > On Tue, 12 Feb 2002, Balani, Sandeep wrote:
> >
> > > Hi all,
> > >   I have a question regarding SCTP implementer's guide. I remember this
> > > issue was discussed during 2nd Bakeoff.
> > >
> > > In
> > http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt
> > > section 5.2.2. We are describing the case of association RESTART when INIT
> > > is received in any state other than CLOSED, COOKIE-ECHOED, COOKIE-WAIT and
> > > SHUTDOWN-ACK-SENT. The text says
> > >
> > >  " Before responding the endpoint MUST check to see if the
> > >    unexpected INIT adds new addresses to the association. If new
> > >    addresses are added to the association, the endpoint MUST respond
> > >    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
> > >    into the 'Verification Tag' of the outbound packet carrying the ABORT.
> > "
> > >
> > > I think that we should check the additional IP addresses when we receive
> > > COOKIE-ECHO from the peer not when we receive INIT from peer, 

Yes, this check on whether "new addresses are added to the association"
is only performed *after* the endpoint received the COOKIE-ECHO *and*
determined that this is a re-start case. 

regards,
-Qiaobing

> > > because if we
> > > send ABORT when we receive INIT then an *INTRUDER* can break an ESTABLISHED
> > > association by sending INIT on an existing association with addition of any
> > > IP addresses.

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 20:44:54 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA03112
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 20:44:53 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA25529
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 20:44:55 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA25263;
	Tue, 12 Feb 2002 20:34:11 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA25218
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 20:34:02 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-248.dallas.net [209.44.42.248])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA02905
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 20:33:59 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id TAA26751;
	Tue, 12 Feb 2002 19:31:23 -0600
Date: Tue, 12 Feb 2002 19:31:23 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Qiaobing Xie <qxie1@email.mot.com>
Cc: "Balani, Sandeep" <s_balani@trillium.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP Implementer's guide question...
Message-ID: <20020212193123.B25887@openss7.org>
Reply-To: bidulock@openss7.org
References: <F1CE15E08172D4119247009027AE9D500B5ECF46@FMSMSX37> <20020212131534.A21856@openss7.org> <3C69B956.BA0F5687@email.mot.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3C69B956.BA0F5687@email.mot.com>; from qxie1@email.mot.com on Tue, Feb 12, 2002 at 06:54:46PM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Qiaobing,

On Tue, 12 Feb 2002, Qiaobing Xie wrote:

> Brian/Sandeep:
> 
> I believe Sandeep is right. The impl-guide should be corrected. See
> below for more...
> 
> "Brian F. G. Bidulock" wrote:
> > 
> > Sandeep,
> > 
> > > >  " Before responding the endpoint MUST check to see if the
> > > >    unexpected INIT adds new addresses to the association. If new
> > > >    addresses are added to the association, the endpoint MUST respond
> > > >    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
> >        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > >    into the 'Verification Tag' of the outbound packet carrying the ABORT.
> >        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > 
> > The sender of the INIT tag must guess the verification tag of the
> > existing association and place it in the Initiation Tag field of
> > the INIT.  It is of no moment that the vTag is zero in the INIT.
> 
> No. vTag is always zero in the INIT, no matter it's for a fresh
> association or a re-started one. The endpoint that restarts would most
> likely lose any memory of its previous life.

The vTag is not zero in the ABORT.  As the vTag in the received
ABORT must match the existing association's tag (otherwise the
ABORT is treated OOTB) the intruder must know the tag of the
existinging association to abort it.  Not only that, but the
intruder must also know a source address of the attacked host's
peer to fake the source address and get it sent there.

It requires the intruder to know both the existing association
tag an a valid peer address of the association.  If the intruder
knows these things, it might as well send an ABORT to the attacked
host faked from the known source address with the known tag value.

Sorry, but there is no 'sploit here that is different from the
normal cases for RFC 2960.

> 
> .....
> > > On Tue, 12 Feb 2002, Balani, Sandeep wrote:
> > >
> > > > Hi all,
> > > >   I have a question regarding SCTP implementer's guide. I remember this
> > > > issue was discussed during 2nd Bakeoff.
> > > >
> > > > In
> > > http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt
> > > > section 5.2.2. We are describing the case of association RESTART when INIT
> > > > is received in any state other than CLOSED, COOKIE-ECHOED, COOKIE-WAIT and
> > > > SHUTDOWN-ACK-SENT. The text says
> > > >
> > > >  " Before responding the endpoint MUST check to see if the
> > > >    unexpected INIT adds new addresses to the association. If new
> > > >    addresses are added to the association, the endpoint MUST respond
> > > >    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
> > > >    into the 'Verification Tag' of the outbound packet carrying the ABORT.
> > > "
> > > >
> > > > I think that we should check the additional IP addresses when we receive
> > > > COOKIE-ECHO from the peer not when we receive INIT from peer, 
> 
> Yes, this check on whether "new addresses are added to the association"
> is only performed *after* the endpoint received the COOKIE-ECHO *and*
> determined that this is a re-start case. 

There is no need to wait for COOKIE-ECHO when then ABORT can be sent in
response to INIT.

--brian

> 
> regards,
> -Qiaobing
> 
> > > > because if we
> > > > send ABORT when we receive INIT then an *INTRUDER* can break an ESTABLISHED
> > > > association by sending INIT on an existing association with addition of any
> > > > IP addresses.

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 21:05:49 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA03411
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 21:05:48 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id VAA26428
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 21:05:49 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA25705;
	Tue, 12 Feb 2002 20:54:41 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA25675
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 20:54:40 -0500 (EST)
Received: from thalia.fm.intel.com (fmfdns02.fm.intel.com [132.233.247.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA03244
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 20:54:37 -0500 (EST)
Received: from fmsmsxvs043.fm.intel.com (fmsmsxv043-1.fm.intel.com [132.233.48.128])
	by thalia.fm.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.50 2002/02/08 23:45:02 root Exp $) with SMTP id BAA20931
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 01:54:38 GMT
Received: from FMSMSX018.fm.intel.com ([132.233.42.197])
 by fmsmsxvs043.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002021217534511960
 ; Tue, 12 Feb 2002 17:53:45 -0800
Received: by fmsmsx018.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <1Y36H33P>; Tue, 12 Feb 2002 17:54:36 -0800
Message-ID: <F1CE15E08172D4119247009027AE9D500B5ECF4E@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        Qiaobing Xie <qxie1@email.mot.com>
Cc: "Balani, Sandeep" <s_balani@trillium.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] SCTP Implementer's guide question...
Date: Tue, 12 Feb 2002 17:54:35 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi Brian,

  Somehow i still think that we should do this check when we receive
COOKIE-ECHO. 
Because, even though It send ABORT with vTag=iTag of INIT chunk, It still
means that Association is terminated.

When the receiver of INIT respond with ABORT it also send TERMINATION-IND to
upper user to close the socket. which means that although u may be right
that peer won't accept ABORT because of mismatching vTag but this end still
assume the association to be TERMINATED. 

On the other hand, If we wait till COOKIE-ECHO,  the intruder can never be
able to pass thru the association.
 
In my understanding getting srcIPaddress and port number (which are
standard) is easy.  U r correct that vTag are difficult to guess.

Thanks
Sandeep   



-----Original Message-----
From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
Sent: Tuesday, February 12, 2002 5:31 PM
To: Qiaobing Xie
Cc: Balani, Sandeep; tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP Implementer's guide question...


Qiaobing,

On Tue, 12 Feb 2002, Qiaobing Xie wrote:

> Brian/Sandeep:
> 
> I believe Sandeep is right. The impl-guide should be corrected. See
> below for more...
> 
> "Brian F. G. Bidulock" wrote:
> > 
> > Sandeep,
> > 
> > > >  " Before responding the endpoint MUST check to see if the
> > > >    unexpected INIT adds new addresses to the association. If new
> > > >    addresses are added to the association, the endpoint MUST respond
> > > >    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
> >        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > >    into the 'Verification Tag' of the outbound packet carrying the
ABORT.
> >        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > 
> > The sender of the INIT tag must guess the verification tag of the
> > existing association and place it in the Initiation Tag field of
> > the INIT.  It is of no moment that the vTag is zero in the INIT.
> 
> No. vTag is always zero in the INIT, no matter it's for a fresh
> association or a re-started one. The endpoint that restarts would most
> likely lose any memory of its previous life.

The vTag is not zero in the ABORT.  As the vTag in the received
ABORT must match the existing association's tag (otherwise the
ABORT is treated OOTB) the intruder must know the tag of the
existinging association to abort it.  Not only that, but the
intruder must also know a source address of the attacked host's
peer to fake the source address and get it sent there.

It requires the intruder to know both the existing association
tag an a valid peer address of the association.  If the intruder
knows these things, it might as well send an ABORT to the attacked
host faked from the known source address with the known tag value.

Sorry, but there is no 'sploit here that is different from the
normal cases for RFC 2960.

> 
> .....
> > > On Tue, 12 Feb 2002, Balani, Sandeep wrote:
> > >
> > > > Hi all,
> > > >   I have a question regarding SCTP implementer's guide. I remember
this
> > > > issue was discussed during 2nd Bakeoff.
> > > >
> > > > In
> > >
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt
> > > > section 5.2.2. We are describing the case of association RESTART
when INIT
> > > > is received in any state other than CLOSED, COOKIE-ECHOED,
COOKIE-WAIT and
> > > > SHUTDOWN-ACK-SENT. The text says
> > > >
> > > >  " Before responding the endpoint MUST check to see if the
> > > >    unexpected INIT adds new addresses to the association. If new
> > > >    addresses are added to the association, the endpoint MUST respond
> > > >    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
> > > >    into the 'Verification Tag' of the outbound packet carrying the
ABORT.
> > > "
> > > >
> > > > I think that we should check the additional IP addresses when we
receive
> > > > COOKIE-ECHO from the peer not when we receive INIT from peer, 
> 
> Yes, this check on whether "new addresses are added to the association"
> is only performed *after* the endpoint received the COOKIE-ECHO *and*
> determined that this is a re-start case. 

There is no need to wait for COOKIE-ECHO when then ABORT can be sent in
response to INIT.

--brian

> 
> regards,
> -Qiaobing
> 
> > > > because if we
> > > > send ABORT when we receive INIT then an *INTRUDER* can break an
ESTABLISHED
> > > > association by sending INIT on an existing association with addition
of any
> > > > IP addresses.

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 12 21:36:17 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA04986
	for <tsvwg-archive@odin.ietf.org>; Tue, 12 Feb 2002 21:36:17 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id VAA27783
	for tsvwg-archive@odin.ietf.org; Tue, 12 Feb 2002 21:36:19 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA27131;
	Tue, 12 Feb 2002 21:25:13 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id VAA27053
	for <tsvwg@optimus.ietf.org>; Tue, 12 Feb 2002 21:25:09 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-248.dallas.net [209.44.42.248])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA03829
	for <tsvwg@ietf.org>; Tue, 12 Feb 2002 21:24:19 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id UAA27316;
	Tue, 12 Feb 2002 20:22:29 -0600
Date: Tue, 12 Feb 2002 20:22:29 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: "Balani, Sandeep" <s_balani@trillium.com>
Cc: Qiaobing Xie <qxie1@email.mot.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP Implementer's guide question...
Message-ID: <20020212202229.C25887@openss7.org>
Reply-To: bidulock@openss7.org
References: <F1CE15E08172D4119247009027AE9D500B5ECF4E@FMSMSX37>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <F1CE15E08172D4119247009027AE9D500B5ECF4E@FMSMSX37>; from s_balani@trillium.com on Tue, Feb 12, 2002 at 05:54:35PM -0800
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Balani,,

On Tue, 12 Feb 2002, Balani, Sandeep wrote:

> Hi Brian,
> 
>   Somehow i still think that we should do this check when we receive
> COOKIE-ECHO. 
> Because, even though It send ABORT with vTag=iTag of INIT chunk, It still
> means that Association is terminated.
> 
> When the receiver of INIT respond with ABORT it also send TERMINATION-IND to
> upper user to close the socket. which means that although u may be right
> that peer won't accept ABORT because of mismatching vTag but this end still
> assume the association to be TERMINATED. 

Where is it required to send termination to the user?  That could be
foolish behavior and might open up all manner of DoS attacks.  Perhaps
the receiver of the INIT should respond with ERROR rather than ABORT,
but I don't see where RFC 2960 requires the imp to send termination
to the user.

> 
> On the other hand, If we wait till COOKIE-ECHO,  the intruder can never be
> able to pass thru the association.

What does waiting for COOKIE-ECHO accomplish?  I don't believe the
receiver has any more information upon the receipt of COOKIE-ECHO
than it does upon receipt of INIT in this case.

>  
> In my understanding getting srcIPaddress and port number (which are
> standard) is easy.  U r correct that vTag are difficult to guess.

--brian

> 
> Thanks
> Sandeep   
> 
> 
> 
> -----Original Message-----
> From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> Sent: Tuesday, February 12, 2002 5:31 PM
> To: Qiaobing Xie
> Cc: Balani, Sandeep; tsvwg@ietf.org
> Subject: Re: [Tsvwg] SCTP Implementer's guide question...
> 
> 
> Qiaobing,
> 
> On Tue, 12 Feb 2002, Qiaobing Xie wrote:
> 
> > Brian/Sandeep:
> > 
> > I believe Sandeep is right. The impl-guide should be corrected. See
> > below for more...
> > 
> > "Brian F. G. Bidulock" wrote:
> > > 
> > > Sandeep,
> > > 
> > > > >  " Before responding the endpoint MUST check to see if the
> > > > >    unexpected INIT adds new addresses to the association. If new
> > > > >    addresses are added to the association, the endpoint MUST respond
> > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
> > >        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > > >    into the 'Verification Tag' of the outbound packet carrying the
> ABORT.
> > >        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > 
> > > The sender of the INIT tag must guess the verification tag of the
> > > existing association and place it in the Initiation Tag field of
> > > the INIT.  It is of no moment that the vTag is zero in the INIT.
> > 
> > No. vTag is always zero in the INIT, no matter it's for a fresh
> > association or a re-started one. The endpoint that restarts would most
> > likely lose any memory of its previous life.
> 
> The vTag is not zero in the ABORT.  As the vTag in the received
> ABORT must match the existing association's tag (otherwise the
> ABORT is treated OOTB) the intruder must know the tag of the
> existinging association to abort it.  Not only that, but the
> intruder must also know a source address of the attacked host's
> peer to fake the source address and get it sent there.
> 
> It requires the intruder to know both the existing association
> tag an a valid peer address of the association.  If the intruder
> knows these things, it might as well send an ABORT to the attacked
> host faked from the known source address with the known tag value.
> 
> Sorry, but there is no 'sploit here that is different from the
> normal cases for RFC 2960.
> 
> > 
> > .....
> > > > On Tue, 12 Feb 2002, Balani, Sandeep wrote:
> > > >
> > > > > Hi all,
> > > > >   I have a question regarding SCTP implementer's guide. I remember
> this
> > > > > issue was discussed during 2nd Bakeoff.
> > > > >
> > > > > In
> > > >
> http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt
> > > > > section 5.2.2. We are describing the case of association RESTART
> when INIT
> > > > > is received in any state other than CLOSED, COOKIE-ECHOED,
> COOKIE-WAIT and
> > > > > SHUTDOWN-ACK-SENT. The text says
> > > > >
> > > > >  " Before responding the endpoint MUST check to see if the
> > > > >    unexpected INIT adds new addresses to the association. If new
> > > > >    addresses are added to the association, the endpoint MUST respond
> > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
> > > > >    into the 'Verification Tag' of the outbound packet carrying the
> ABORT.
> > > > "
> > > > >
> > > > > I think that we should check the additional IP addresses when we
> receive
> > > > > COOKIE-ECHO from the peer not when we receive INIT from peer, 
> > 
> > Yes, this check on whether "new addresses are added to the association"
> > is only performed *after* the endpoint received the COOKIE-ECHO *and*
> > determined that this is a re-start case. 
> 
> There is no need to wait for COOKIE-ECHO when then ABORT can be sent in
> response to INIT.
> 
> --brian
> 
> > 
> > regards,
> > -Qiaobing
> > 
> > > > > because if we
> > > > > send ABORT when we receive INIT then an *INTRUDER* can break an
> ESTABLISHED
> > > > > association by sending INIT on an existing association with addition
> of any
> > > > > IP addresses.
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 01:10:29 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA11583
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 01:10:29 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id BAA12155
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 01:10:31 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA05517;
	Wed, 13 Feb 2002 00:55:38 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA05491
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 00:55:36 -0500 (EST)
Received: from motgate.mot.com (motgate.mot.com [129.188.136.100])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA11068
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 00:55:33 -0500 (EST)
Received: [from pobox3.mot.com (pobox3.mot.com [10.64.251.242]) by motgate.mot.com (motgate 2.1) with ESMTP id WAA08118 for <tsvwg@ietf.org>; Tue, 12 Feb 2002 22:55:34 -0700 (MST)]
Received: [from il27exb01.cig.mot.com (il27exb01.cig.mot.com [136.182.15.100]) by pobox3.mot.com (MOT-pobox3 2.0) with ESMTP id WAA03057 for <tsvwg@ietf.org>; Tue, 12 Feb 2002 22:44:13 -0700 (MST)]
Received: from email.mot.com (az33-t05-async35.corp.mot.com [10.64.18.77]) by il27exb01.cig.mot.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2654.52)
	id 1LK15Y3Q; Tue, 12 Feb 2002 23:55:32 -0600
Message-ID: <3C6A0067.80E9D68@email.mot.com>
Date: Tue, 12 Feb 2002 23:57:59 -0600
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: bidulock@openss7.org
CC: "Balani, Sandeep" <s_balani@trillium.com>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP Implementer's guide question...
References: <F1CE15E08172D4119247009027AE9D500B5ECF4E@FMSMSX37> <20020212202229.C25887@openss7.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Brian,

Probably we are talking about two diff. scenarios. Please see below.

"Brian F. G. Bidulock" wrote:
....
> Where is it required to send termination to the user?  That could be
> foolish behavior and might open up all manner of DoS attacks.  Perhaps
> the receiver of the INIT should respond with ERROR rather than ABORT,
> but I don't see where RFC 2960 requires the imp to send termination
> to the user.

It is defined clearly in RFC 2960 that the receiver of an INIT responds
with an INIT-ACK, not an ABORT nor an ERROR, period.

> >
> > On the other hand, If we wait till COOKIE-ECHO,  the intruder can never be
> > able to pass thru the association.
> 
> What does waiting for COOKIE-ECHO accomplish?  I don't believe the
> receiver has any more information upon the receipt of COOKIE-ECHO
> than it does upon receipt of INIT in this case.

By design, the receiver has zero information until receiving the
COOKIE-ECHO. Why bother to worry about sending an ABORT or ERROR before
the COOKIE-ECHO arrives.

-Qiaobing

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 02:06:48 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA20847
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 02:06:48 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id CAA17377
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 02:06:49 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA16788;
	Wed, 13 Feb 2002 01:56:07 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA16757
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 01:56:05 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-217.dallas.net [209.44.42.217])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA12863
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 01:56:02 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id AAA31537;
	Wed, 13 Feb 2002 00:55:50 -0600
Date: Wed, 13 Feb 2002 00:55:50 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Qiaobing Xie <qxie1@email.mot.com>
Cc: "Balani, Sandeep" <s_balani@trillium.com>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP Implementer's guide question...
Message-ID: <20020213005550.G27856@openss7.org>
Reply-To: bidulock@openss7.org
References: <F1CE15E08172D4119247009027AE9D500B5ECF4E@FMSMSX37> <20020212202229.C25887@openss7.org> <3C6A0067.80E9D68@email.mot.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3C6A0067.80E9D68@email.mot.com>; from qxie1@email.mot.com on Tue, Feb 12, 2002 at 11:57:59PM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Qiaobing,

On Tue, 12 Feb 2002, Qiaobing Xie wrote:

> Brian,
> 
> Probably we are talking about two diff. scenarios. Please see below.
> 
> "Brian F. G. Bidulock" wrote:
> ....
> > Where is it required to send termination to the user?  That could be
> > foolish behavior and might open up all manner of DoS attacks.  Perhaps
> > the receiver of the INIT should respond with ERROR rather than ABORT,
> > but I don't see where RFC 2960 requires the imp to send termination
> > to the user.
> 
> It is defined clearly in RFC 2960 that the receiver of an INIT responds
> with an INIT-ACK, not an ABORT nor an ERROR, period.

We are talking about the implementation guide.  I asked where in RFC 2960
it requires an imp to send termination to the user when sending an ABORT:
the point being that it doesn't say any such thing.

Besides, your statement above is false:  RFC 2960 Section 5.1 says:

   If an endpoint receives an INIT, INIT ACK, or COOKIE ECHO chunk but
   decides not to establish the new association due to missing mandatory
   parameters in the received INIT or INIT ACK, invalid parameter
   values, or lack of local resources, it MUST respond with an ABORT
   chunk.  It SHOULD also specify the cause of abort, such as the type
   of the missing mandatory parameters, etc., by including the error
   cause parameters with the ABORT chunk.  The Verification Tag field in
   the common header of the outbound SCTP packet containing the ABORT
   chunk MUST be set to the Initiate Tag value of the peer.

It looks to me like there are circumstances in which in implementation
"MUST" send an ABORT in response to an INIT according to RFC 2960.

> 
> > >
> > > On the other hand, If we wait till COOKIE-ECHO,  the intruder can never be
> > > able to pass thru the association.
> > 
> > What does waiting for COOKIE-ECHO accomplish?  I don't believe the
> > receiver has any more information upon the receipt of COOKIE-ECHO
> > than it does upon receipt of INIT in this case.
> 
> By design, the receiver has zero information until receiving the
> COOKIE-ECHO. Why bother to worry about sending an ABORT or ERROR before
> the COOKIE-ECHO arrives.

To avoid the resource consumption of generating the cookie.

--brian

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 08:21:01 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA00652
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 08:21:01 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA02825
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 08:21:04 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA02318;
	Wed, 13 Feb 2002 08:09:17 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA02290
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 08:09:15 -0500 (EST)
Received: from beamer.mchh.siemens.de (beamer.mchh.siemens.de [194.138.158.163])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA00337
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 08:09:11 -0500 (EST)
Received: from moody.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.226])
	by beamer.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id OAA06500;
	Wed, 13 Feb 2002 14:09:07 +0100 (MET)
Received: from mchh246e.demchh201e.icn.siemens.de ([139.21.200.56])
	by moody.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id OAA15535;
	Wed, 13 Feb 2002 14:09:06 +0100 (MET)
Received: by MCHH246E with Internet Mail Service (5.5.2653.19)
	id <1ZZHL4AX>; Wed, 13 Feb 2002 14:09:10 +0100
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF756@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'Balani, Sandeep'" <s_balani@trillium.com>,
        "'bidulock@openss7.org'"
	 <bidulock@openss7.org>,
        Qiaobing Xie <qxie1@email.mot.com>
Cc: tsvwg@ietf.org
Subject: RE: [Tsvwg] SCTP Implementer's guide question...
Date: Wed, 13 Feb 2002 14:09:08 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Dear all,
see my comments below.
Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Balani, Sandeep [SMTP:s_balani@trillium.com]
> Sent:	Wednesday, February 13, 2002 2:55 AM
> To:	'bidulock@openss7.org'; Qiaobing Xie
> Cc:	Balani, Sandeep; tsvwg@ietf.org
> Subject:	RE: [Tsvwg] SCTP Implementer's guide question...
> 
> Hi Brian,
> 
>   Somehow i still think that we should do this check when we receive
> COOKIE-ECHO. 
> Because, even though It send ABORT with vTag=iTag of INIT chunk, It still
> means that Association is terminated.
> 
> When the receiver of INIT respond with ABORT it also send TERMINATION-IND to
> upper user to close the socket. which means that although u may be right
> that peer won't accept ABORT because of mismatching vTag but this end still
> assume the association to be TERMINATED. 
	[Tuexen Michael]  My understanding of the RFC 2960 and
	the Implementors guide is the following:
	The intruder sends an INIT with an additional address. The
	verification tag in the SCTP common header is 0, the initiate
	tag in the init chunk is some value which is not the correct
	tag. The receiver of this INIT will send an Abort chunk with
	the initiate tag as a verification tag. It will not inform the
	user! The receiver of the Abort will discard it because it has
	not the correct tag. The user of that endpoint will also not
	be informed.
	There is no security problem and not user will be informed. This
	behaviour should be OK.

> On the other hand, If we wait till COOKIE-ECHO,  the intruder can never be
> able to pass thru the association.
>  
> In my understanding getting srcIPaddress and port number (which are
> standard) is easy.  U r correct that vTag are difficult to guess.
> 
> Thanks
> Sandeep   
> 
> 
> 
> -----Original Message-----
> From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> Sent: Tuesday, February 12, 2002 5:31 PM
> To: Qiaobing Xie
> Cc: Balani, Sandeep; tsvwg@ietf.org
> Subject: Re: [Tsvwg] SCTP Implementer's guide question...
> 
> 
> Qiaobing,
> 
> On Tue, 12 Feb 2002, Qiaobing Xie wrote:
> 
> > Brian/Sandeep:
> > 
> > I believe Sandeep is right. The impl-guide should be corrected. See
> > below for more...
> > 
> > "Brian F. G. Bidulock" wrote:
> > > 
> > > Sandeep,
> > > 
> > > > >  " Before responding the endpoint MUST check to see if the
> > > > >    unexpected INIT adds new addresses to the association. If new
> > > > >    addresses are added to the association, the endpoint MUST respond
> > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
> > >        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > > >    into the 'Verification Tag' of the outbound packet carrying the
> ABORT.
> > >        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > 
> > > The sender of the INIT tag must guess the verification tag of the
> > > existing association and place it in the Initiation Tag field of
> > > the INIT.  It is of no moment that the vTag is zero in the INIT.
> > 
> > No. vTag is always zero in the INIT, no matter it's for a fresh
> > association or a re-started one. The endpoint that restarts would most
> > likely lose any memory of its previous life.
> 
> The vTag is not zero in the ABORT.  As the vTag in the received
> ABORT must match the existing association's tag (otherwise the
> ABORT is treated OOTB) the intruder must know the tag of the
> existinging association to abort it.  Not only that, but the
> intruder must also know a source address of the attacked host's
> peer to fake the source address and get it sent there.
> 
> It requires the intruder to know both the existing association
> tag an a valid peer address of the association.  If the intruder> 
> knows these things, it might as well send an ABORT to the attacked
> host faked from the known source address with the known tag value.
> 
> Sorry, but there is no 'sploit here that is different from the
> normal cases for RFC 2960.
> 
> > 
> > .....
> > > > On Tue, 12 Feb 2002, Balani, Sandeep wrote:
> > > >
> > > > > Hi all,
> > > > >   I have a question regarding SCTP implementer's guide. I remember
> this
> > > > > issue was discussed during 2nd Bakeoff.
> > > > >
> > > > > In
> > > >
> http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt
> > > > > section 5.2.2. We are describing the case of association RESTART
> when INIT
> > > > > is received in any state other than CLOSED, COOKIE-ECHOED,
> COOKIE-WAIT and
> > > > > SHUTDOWN-ACK-SENT. The text says
> > > > >
> > > > >  " Before responding the endpoint MUST check to see if the
> > > > >    unexpected INIT adds new addresses to the association. If new
> > > > >    addresses are added to the association, the endpoint MUST respond
> > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
> > > > >    into the 'Verification Tag' of the outbound packet carrying the
> ABORT.
> > > > "
> > > > >
> > > > > I think that we should check the additional IP addresses when we
> receive
> > > > > COOKIE-ECHO from the peer not when we receive INIT from peer, 
> > 
> > Yes, this check on whether "new addresses are added to the association"
> > is only performed *after* the endpoint received the COOKIE-ECHO *and*
> > determined that this is a re-start case. 
> 
> There is no need to wait for COOKIE-ECHO when then ABORT can be sent in
> response to INIT.
> 
> --brian
> 
> > 
> > regards,
> > -Qiaobing
> > 
> > > > > because if we
> > > > > send ABORT when we receive INIT then an *INTRUDER* can break an
> ESTABLISHED
> > > > > association by sending INIT on an existing association with addition
> of any
> > > > > IP addresses.
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 09:21:42 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA03140
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 09:21:41 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA06025
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 09:21:43 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA05156;
	Wed, 13 Feb 2002 09:05:07 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA05131
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 09:05:05 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA02626
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 09:05:02 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir02nok.ntc.nokia.com (esvir02nokt.ntc.nokia.com [172.21.143.34])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1DE5AZ23626
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 16:05:11 +0200 (EET)
Received: from esebh001.NOE.Nokia.com (unverified) by esvir02nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T590ca617deac158f22076@esvir02nok.ntc.nokia.com>;
 Wed, 13 Feb 2002 16:05:01 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh001.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Wed, 13 Feb 2002 16:05:00 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] SCTP Implementer's guide question...
Date: Wed, 13 Feb 2002 16:04:59 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF26308EE97@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] SCTP Implementer's guide question...
Thread-Index: AcG0W5CUoo1mf2UbSp2bBJ6rwKjefAANOtzQ
To: <bidulock@openss7.org>, <qxie1@email.mot.com>
Cc: <s_balani@trillium.com>, <Qiaobing.Xie@motorola.com>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 13 Feb 2002 14:05:00.0251 (UTC) FILETIME=[6CDC1AB0:01C1B497]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id JAA05132
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi all! Some comments regarding the received mails so far...

	First of all, I completely agree with Brian about not sending anything to the upper user. The INIT receiver must not send any termination message to its upper user at all.

	As it has been told, to be able to make that attack, you should know the VT, one of the addresses and the port, so it is easier to simply send an ABORT...

	And about why not waiting for the COOKIE-ECHO... I was the one that proposed it :-0. Initially you should wait until you received the COOKIE-ECHO, but...

	The point is that the State Cookie parameter inside the INIT ACK chunk is authenticated, but not encrypted... So, the normal procedure in these cases, when receiving a duplicate INIT is setting the Tie Tags inside the State Cookie parameter to both the values of our VT and our peer's one.

	Ok, the attacker should not know about the internal structure, but the Tie Tags are 32-bit numbers that should always be there set to the same value and if the attacker sends several INITs and compares them, taking into account that some values as the IP addresses and stuff can easily be identified, it would be relatively simple to "guess" the VTs from those State Cookies... Even easier, if it sends INITs to both peers, then checking 32-bit values that are set to the same value inside State Cookies generated by different peers would make things simpler...

	Also, if you send the INIT ACK while in the 5.2.1 section case, it is even worse, as you are sending the "real" VTs inside the INIT ACK itself (ok, there is a really tiny window of time to do it, but still... :-0 it was easy to avoid)... Thus, for the reason stated above, to make things the same way (5.2.1 and 5.2.2), and to save time while calculating the INIT ACK (the MAC might take some time to calculate), I think it is better not to generate the INIT ACK chunk at all...

	The receiver of the ABORT should check which are the new addresses and try to restart not using them, then terminate the association and then start it again if it really need to use those other addresses... It could also do it using the AddIP extension... :-/

	I hope this answers your questions...

	BR Iván Arias Rodríguez

	

> > 
> > > >
> > > > On the other hand, If we wait till COOKIE-ECHO,  the 
> intruder can never be
> > > > able to pass thru the association.
> > > 
> > > What does waiting for COOKIE-ECHO accomplish?  I don't believe the
> > > receiver has any more information upon the receipt of COOKIE-ECHO
> > > than it does upon receipt of INIT in this case.
> > 
> > By design, the receiver has zero information until receiving the
> > COOKIE-ECHO. Why bother to worry about sending an ABORT or 
> ERROR before
> > the COOKIE-ECHO arrives.
> 
> To avoid the resource consumption of generating the cookie.
> 
> --brian
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 09:57:28 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA05808
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 09:57:28 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA08117
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 09:57:30 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA07020;
	Wed, 13 Feb 2002 09:40:58 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA06989
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 09:40:57 -0500 (EST)
Received: from sj-msg-core-3.cisco.com (sj-msg-core-3.cisco.com [171.70.157.152])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA04539
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 09:40:42 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-3.cisco.com (8.11.3/8.9.1) with ESMTP id g1DEdhZ15965;
	Wed, 13 Feb 2002 06:39:58 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABB81677;
	Wed, 13 Feb 2002 06:39:49 -0800 (PST)
Message-ID: <3C6A7AB4.E828083F@stewart.chicago.il.us>
Date: Wed, 13 Feb 2002 08:39:48 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: "Balani, Sandeep" <s_balani@trillium.com>
CC: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        Qiaobing Xie <qxie1@email.mot.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP Implementer's guide question...
References: <F1CE15E08172D4119247009027AE9D500B5ECF4E@FMSMSX37>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Balani:

It is quite unusal for me but I am in complete agreement
with Brian :-)

"Balani, Sandeep" wrote:
> 
> Hi Brian,
> 
>   Somehow i still think that we should do this check when we receive
> COOKIE-ECHO.
> Because, even though It send ABORT with vTag=iTag of INIT chunk, It still
> means that Association is terminated.
> 
> When the receiver of INIT respond with ABORT it also send TERMINATION-IND to
> upper user to close the socket. which means that although u may be right
> that peer won't accept ABORT because of mismatching vTag but this end still
> assume the association to be TERMINATED.

No it SHOULD NOT send anything up to the upper layer. The ABORT it sends
with the iTag would NOT be for THIS association. As far as the local
endpoint that receives the abort is concerned the association has
NOT restarted, the vtag is incorrect. 

> 
> On the other hand, If we wait till COOKIE-ECHO,  the intruder can never be
> able to pass thru the association.
> 
> In my understanding getting srcIPaddress and port number (which are
> standard) is easy.  U r correct that vTag are difficult to guess.
> 

I considered this for a LONG time before deciding the right way to
do it is in response to the INIT itself here is why:

We have two basic cases:
1) The peer has restarted
and
2) An attacker or STRAY INIT is arriving.

Now for case 1) I think we all will agree it does not
matter if I wait to send the abort at the cookie-echo or INIT. In
both case the "restart" is NOT going to happen until one of two
things occur:
a) A HB on the now dead association causes an abort (HB from
   Endpoint-Z to Endpoint-A)
<or>
b) The peer drops the extra address and resends its INIT. It could then
   later use add-ip to add the new address (if it supports the
   extension) or shutdown the association properly and restart again
   with the new address.

So just considering case 1) sending the abort from the INIT makes
sense (remember it DOES take a bit of cpu to do the hash .. depending
on the hash algo the implementation is using it can be QUITE a bit
of cpu). And it quickens the recovery time.

Now lets consider case 2) the one in contention..

We have an association:

Endpoint-A                                 Endpoint-Z
IP-1                                        IP-3
    +port5000 <--------------------------->+port6000
IP-2                                        IP-4

Now our evil attacker Endpoint-E sends in an INIT 
after guessing the association is up.

Endpoint-E-----srcip:INIT(Vtag=0,
                          iTag=N, 
                          IPaddress IP-1,
                          IP-2,IP-E)--->Endpoint-Z

Now Endpoint-Z sees a NEW address in the list. It can do one of 
two things:
a) Respond with a ----Abort(Vtag=N)----->src-ip
or
b) send a  -----Cookie()--->src-ip
   
If it does a)
--1) if src-ip is IP-1/IP-2 what happens? Endpoint-A gets the
     abort() checks the V-tag and dumps it on the floor since
     it is a old stale Abort .. not this association.
--2) if src-ip is IP-E then the attacker has failed.

   In either case Endpoint-Z would NOT tell the application
   that anything happened. It would possibly LOG the event to
   a security log. Which would then help point a finger at
   IP-E but it would not tell the app.

If it does b)
--1) It would send a cookie back that it KNOWS that it
     will reject. Thus wasting some CPU cycles to sign
     a bad cookie.
--2) If src-ip is IP-1/IP-2 the peer would get a cookie
     it did not send and discard it.
--3) If src-ip is IP-E the evil endpoint would then send
     the cookie-echo... which would of course be rejected
     with an abort in the same fashion as if we caught
     it early.

So the bottom line is that if you do the detection early 
it just saves CPU cycles on the victim AND CPU cycles on
the peer of the vicitim in the attackers case.

The other neat point is you gain a audit trail of the
events. When sending a Cookie instead of an abort, if the 
attacker uses IP-1/IP-2 as src-ip then we are responding 
normally and create no audit trail for our security people
to track down... if it uses IP-E then we DO have an audit 
trail but have done extra CPU computation to get something 
we could have had with less cost..

In the end the result is the same though..

R




> -----Original Message-----
> From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> Sent: Tuesday, February 12, 2002 5:31 PM
> To: Qiaobing Xie
> Cc: Balani, Sandeep; tsvwg@ietf.org
> Subject: Re: [Tsvwg] SCTP Implementer's guide question...
> 
> Qiaobing,
> 
> On Tue, 12 Feb 2002, Qiaobing Xie wrote:
> 
> > Brian/Sandeep:
> >
> > I believe Sandeep is right. The impl-guide should be corrected. See
> > below for more...
> >
> > "Brian F. G. Bidulock" wrote:
> > >
> > > Sandeep,
> > >
> > > > >  " Before responding the endpoint MUST check to see if the
> > > > >    unexpected INIT adds new addresses to the association. If new
> > > > >    addresses are added to the association, the endpoint MUST respond
> > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
> > >        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > > >    into the 'Verification Tag' of the outbound packet carrying the
> ABORT.
> > >        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > >
> > > The sender of the INIT tag must guess the verification tag of the
> > > existing association and place it in the Initiation Tag field of
> > > the INIT.  It is of no moment that the vTag is zero in the INIT.
> >
> > No. vTag is always zero in the INIT, no matter it's for a fresh
> > association or a re-started one. The endpoint that restarts would most
> > likely lose any memory of its previous life.
> 
> The vTag is not zero in the ABORT.  As the vTag in the received
> ABORT must match the existing association's tag (otherwise the
> ABORT is treated OOTB) the intruder must know the tag of the
> existinging association to abort it.  Not only that, but the
> intruder must also know a source address of the attacked host's
> peer to fake the source address and get it sent there.
> 
> It requires the intruder to know both the existing association
> tag an a valid peer address of the association.  If the intruder
> knows these things, it might as well send an ABORT to the attacked
> host faked from the known source address with the known tag value.
> 
> Sorry, but there is no 'sploit here that is different from the
> normal cases for RFC 2960.
> 
> >
> > .....
> > > > On Tue, 12 Feb 2002, Balani, Sandeep wrote:
> > > >
> > > > > Hi all,
> > > > >   I have a question regarding SCTP implementer's guide. I remember
> this
> > > > > issue was discussed during 2nd Bakeoff.
> > > > >
> > > > > In
> > > >
> http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt
> > > > > section 5.2.2. We are describing the case of association RESTART
> when INIT
> > > > > is received in any state other than CLOSED, COOKIE-ECHOED,
> COOKIE-WAIT and
> > > > > SHUTDOWN-ACK-SENT. The text says
> > > > >
> > > > >  " Before responding the endpoint MUST check to see if the
> > > > >    unexpected INIT adds new addresses to the association. If new
> > > > >    addresses are added to the association, the endpoint MUST respond
> > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected INIT
> > > > >    into the 'Verification Tag' of the outbound packet carrying the
> ABORT.
> > > > "
> > > > >
> > > > > I think that we should check the additional IP addresses when we
> receive
> > > > > COOKIE-ECHO from the peer not when we receive INIT from peer,
> >
> > Yes, this check on whether "new addresses are added to the association"
> > is only performed *after* the endpoint received the COOKIE-ECHO *and*
> > determined that this is a re-start case.
> 
> There is no need to wait for COOKIE-ECHO when then ABORT can be sent in
> response to INIT.
> 
> --brian
> 
> >
> > regards,
> > -Qiaobing
> >
> > > > > because if we
> > > > > send ABORT when we receive INIT then an *INTRUDER* can break an
> ESTABLISHED
> > > > > association by sending INIT on an existing association with addition
> of any
> > > > > IP addresses.
> 
> --
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 09:57:47 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA05826
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 09:57:46 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA08136
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 09:57:48 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA07163;
	Wed, 13 Feb 2002 09:42:55 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA07106
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 09:42:50 -0500 (EST)
Received: from sj-msg-core-3.cisco.com (sj-msg-core-3.cisco.com [171.70.157.152])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA04626
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 09:42:46 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-3.cisco.com (8.11.3/8.9.1) with ESMTP id g1DEfxZ17199;
	Wed, 13 Feb 2002 06:42:01 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABB81716;
	Wed, 13 Feb 2002 06:42:06 -0800 (PST)
Message-ID: <3C6A7B3D.76377D22@stewart.chicago.il.us>
Date: Wed, 13 Feb 2002 08:42:05 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Qiaobing Xie <qxie1@email.mot.com>
CC: bidulock@openss7.org, "Balani, Sandeep" <s_balani@trillium.com>,
        Qiaobing Xie <Qiaobing.Xie@motorola.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP Implementer's guide question...
References: <F1CE15E08172D4119247009027AE9D500B5ECF4E@FMSMSX37> <20020212202229.C25887@openss7.org> <3C6A0067.80E9D68@email.mot.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Qiaobing:

Brian is ABSOLUTLY correct... please see my previous mail 

R

Qiaobing Xie wrote:
> 
> Brian,
> 
> Probably we are talking about two diff. scenarios. Please see below.
> 
> "Brian F. G. Bidulock" wrote:
> ....
> > Where is it required to send termination to the user?  That could be
> > foolish behavior and might open up all manner of DoS attacks.  Perhaps
> > the receiver of the INIT should respond with ERROR rather than ABORT,
> > but I don't see where RFC 2960 requires the imp to send termination
> > to the user.
> 
> It is defined clearly in RFC 2960 that the receiver of an INIT responds
> with an INIT-ACK, not an ABORT nor an ERROR, period.
> 
> > >
> > > On the other hand, If we wait till COOKIE-ECHO,  the intruder can never be
> > > able to pass thru the association.
> >
> > What does waiting for COOKIE-ECHO accomplish?  I don't believe the
> > receiver has any more information upon the receipt of COOKIE-ECHO
> > than it does upon receipt of INIT in this case.
> 
> By design, the receiver has zero information until receiving the
> COOKIE-ECHO. Why bother to worry about sending an ABORT or ERROR before
> the COOKIE-ECHO arrives.
> 
> -Qiaobing
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 13:36:47 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA16643
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 13:36:47 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA22414
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 13:36:49 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA20845;
	Wed, 13 Feb 2002 13:10:11 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA20814
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 13:10:09 -0500 (EST)
Received: from thalia.fm.intel.com (fmfdns02.fm.intel.com [132.233.247.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA15794
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 13:10:06 -0500 (EST)
Received: from fmsmsxvs042.fm.intel.com (fmsmsxv042-1.fm.intel.com [132.233.48.110])
	by thalia.fm.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.50 2002/02/08 23:45:02 root Exp $) with SMTP id SAA08721
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 18:10:07 GMT
Received: from fmsmsx28.fm.intel.com ([132.233.42.28])
 by fmsmsxvs042.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002021310124631342
 ; Wed, 13 Feb 2002 10:12:46 -0800
Received: by fmsmsx28.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <1YPAQ14P>; Wed, 13 Feb 2002 10:10:05 -0800
Message-ID: <F1CE15E08172D4119247009027AE9D500B5ECF4F@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        "Balani, Sandeep" <s_balani@trillium.com>
Cc: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        Qiaobing Xie<qxie1@email.mot.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] SCTP Implementer's guide question...
Date: Wed, 13 Feb 2002 10:10:00 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi Randy,
 
Some Comments, 

-----Original Message-----
From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
Sent: Wednesday, February 13, 2002 6:40 AM
To: Balani, Sandeep
Cc: 'bidulock@openss7.org'; Qiaobing Xie; tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP Implementer's guide question...


Balani:

It is quite unusal for me but I am in complete agreement
with Brian :-)

"Balani, Sandeep" wrote:
> 
> Hi Brian,
> 
>   Somehow i still think that we should do this check when we receive
> COOKIE-ECHO.
> Because, even though It send ABORT with vTag=iTag of INIT chunk, It still
> means that Association is terminated.
> 
> When the receiver of INIT respond with ABORT it also send TERMINATION-IND
to
> upper user to close the socket. which means that although u may be right
> that peer won't accept ABORT because of mismatching vTag but this end
still
> assume the association to be TERMINATED.

No it SHOULD NOT send anything up to the upper layer. The ABORT it sends
with the iTag would NOT be for THIS association. As far as the local
endpoint that receives the abort is concerned the association has
NOT restarted, the vtag is incorrect. 

[ Balani Sandeep] Randy, Can u please explain Why would we not send
Termination Indication to local user when association is ABORTED. 
If there was an SCTP exception locally or some illegal event occured in
which case u send ABORT to peer & terminate the association. 

One more reason for Terminating the association is in case of RESTART
because the peer loses all its context it may not remember what addresses it
sent during first time of association. So in that case even though he
receives ABORT ( with additional addresses) he has to guess that what
addresses i sent previously. In that case it will be hit & trial for peer to
send IP addresses.
 
On the other hand if u teminate the assocition he can start with whatever IP
addresses it wants. Is'nt it ?  

Brian, I am sorry for my foolishness, but i still don't understand. 

Lastly the reason to wait for COOKIE-ECHO is because INIT-ACK would be send
to genuine user because it matched with its association, & that peer will
dump the INIT-ACK so we would never receive COOKIE-EACHO. that means No need
to send ABORT.
 
Although i agree that extra cpu cycles are utlized. 

- Sandeep 

> 
> On the other hand, If we wait till COOKIE-ECHO,  the intruder can never be
> able to pass thru the association.
> 
> In my understanding getting srcIPaddress and port number (which are
> standard) is easy.  U r correct that vTag are difficult to guess.
> 

I considered this for a LONG time before deciding the right way to
do it is in response to the INIT itself here is why:

We have two basic cases:
1) The peer has restarted
and
2) An attacker or STRAY INIT is arriving.

Now for case 1) I think we all will agree it does not
matter if I wait to send the abort at the cookie-echo or INIT. In
both case the "restart" is NOT going to happen until one of two
things occur:
a) A HB on the now dead association causes an abort (HB from
   Endpoint-Z to Endpoint-A)
<or>
b) The peer drops the extra address and resends its INIT. It could then
   later use add-ip to add the new address (if it supports the
   extension) or shutdown the association properly and restart again
   with the new address.

So just considering case 1) sending the abort from the INIT makes
sense (remember it DOES take a bit of cpu to do the hash .. depending
on the hash algo the implementation is using it can be QUITE a bit
of cpu). And it quickens the recovery time.

Now lets consider case 2) the one in contention..

We have an association:

Endpoint-A                                 Endpoint-Z
IP-1                                        IP-3
    +port5000 <--------------------------->+port6000
IP-2                                        IP-4

Now our evil attacker Endpoint-E sends in an INIT 
after guessing the association is up.

Endpoint-E-----srcip:INIT(Vtag=0,
                          iTag=N, 
                          IPaddress IP-1,
                          IP-2,IP-E)--->Endpoint-Z

Now Endpoint-Z sees a NEW address in the list. It can do one of 
two things:
a) Respond with a ----Abort(Vtag=N)----->src-ip
or
b) send a  -----Cookie()--->src-ip
   
If it does a)
--1) if src-ip is IP-1/IP-2 what happens? Endpoint-A gets the
     abort() checks the V-tag and dumps it on the floor since
     it is a old stale Abort .. not this association.
--2) if src-ip is IP-E then the attacker has failed.

   In either case Endpoint-Z would NOT tell the application
   that anything happened. It would possibly LOG the event to
   a security log. Which would then help point a finger at
   IP-E but it would not tell the app.

If it does b)
--1) It would send a cookie back that it KNOWS that it
     will reject. Thus wasting some CPU cycles to sign
     a bad cookie.
--2) If src-ip is IP-1/IP-2 the peer would get a cookie
     it did not send and discard it.
--3) If src-ip is IP-E the evil endpoint would then send
     the cookie-echo... which would of course be rejected
     with an abort in the same fashion as if we caught
     it early.

So the bottom line is that if you do the detection early 
it just saves CPU cycles on the victim AND CPU cycles on
the peer of the vicitim in the attackers case.

The other neat point is you gain a audit trail of the
events. When sending a Cookie instead of an abort, if the 
attacker uses IP-1/IP-2 as src-ip then we are responding 
normally and create no audit trail for our security people
to track down... if it uses IP-E then we DO have an audit 
trail but have done extra CPU computation to get something 
we could have had with less cost..

In the end the result is the same though..

R




> -----Original Message-----
> From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> Sent: Tuesday, February 12, 2002 5:31 PM
> To: Qiaobing Xie
> Cc: Balani, Sandeep; tsvwg@ietf.org
> Subject: Re: [Tsvwg] SCTP Implementer's guide question...
> 
> Qiaobing,
> 
> On Tue, 12 Feb 2002, Qiaobing Xie wrote:
> 
> > Brian/Sandeep:
> >
> > I believe Sandeep is right. The impl-guide should be corrected. See
> > below for more...
> >
> > "Brian F. G. Bidulock" wrote:
> > >
> > > Sandeep,
> > >
> > > > >  " Before responding the endpoint MUST check to see if the
> > > > >    unexpected INIT adds new addresses to the association. If new
> > > > >    addresses are added to the association, the endpoint MUST
respond
> > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected
INIT
> > >
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > > >    into the 'Verification Tag' of the outbound packet carrying the
> ABORT.
> > >
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > >
> > > The sender of the INIT tag must guess the verification tag of the
> > > existing association and place it in the Initiation Tag field of
> > > the INIT.  It is of no moment that the vTag is zero in the INIT.
> >
> > No. vTag is always zero in the INIT, no matter it's for a fresh
> > association or a re-started one. The endpoint that restarts would most
> > likely lose any memory of its previous life.
> 
> The vTag is not zero in the ABORT.  As the vTag in the received
> ABORT must match the existing association's tag (otherwise the
> ABORT is treated OOTB) the intruder must know the tag of the
> existinging association to abort it.  Not only that, but the
> intruder must also know a source address of the attacked host's
> peer to fake the source address and get it sent there.
> 
> It requires the intruder to know both the existing association
> tag an a valid peer address of the association.  If the intruder
> knows these things, it might as well send an ABORT to the attacked
> host faked from the known source address with the known tag value.
> 
> Sorry, but there is no 'sploit here that is different from the
> normal cases for RFC 2960.
> 
> >
> > .....
> > > > On Tue, 12 Feb 2002, Balani, Sandeep wrote:
> > > >
> > > > > Hi all,
> > > > >   I have a question regarding SCTP implementer's guide. I remember
> this
> > > > > issue was discussed during 2nd Bakeoff.
> > > > >
> > > > > In
> > > >
> http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt
> > > > > section 5.2.2. We are describing the case of association RESTART
> when INIT
> > > > > is received in any state other than CLOSED, COOKIE-ECHOED,
> COOKIE-WAIT and
> > > > > SHUTDOWN-ACK-SENT. The text says
> > > > >
> > > > >  " Before responding the endpoint MUST check to see if the
> > > > >    unexpected INIT adds new addresses to the association. If new
> > > > >    addresses are added to the association, the endpoint MUST
respond
> > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected
INIT
> > > > >    into the 'Verification Tag' of the outbound packet carrying the
> ABORT.
> > > > "
> > > > >
> > > > > I think that we should check the additional IP addresses when we
> receive
> > > > > COOKIE-ECHO from the peer not when we receive INIT from peer,
> >
> > Yes, this check on whether "new addresses are added to the association"
> > is only performed *after* the endpoint received the COOKIE-ECHO *and*
> > determined that this is a re-start case.
> 
> There is no need to wait for COOKIE-ECHO when then ABORT can be sent in
> response to INIT.
> 
> --brian
> 
> >
> > regards,
> > -Qiaobing
> >
> > > > > because if we
> > > > > send ABORT when we receive INIT then an *INTRUDER* can break an
> ESTABLISHED
> > > > > association by sending INIT on an existing association with
addition
> of any
> > > > > IP addresses.
> 
> --
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 13:43:42 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA16836
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 13:43:42 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA22527
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 13:43:44 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA21389;
	Wed, 13 Feb 2002 13:26:18 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA21364
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 13:26:16 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-217.dallas.net [209.44.42.217])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA16230
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 13:26:10 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id MAA09356;
	Wed, 13 Feb 2002 12:25:47 -0600
Date: Wed, 13 Feb 2002 12:25:47 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: "Balani, Sandeep" <s_balani@trillium.com>
Cc: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        Qiaobing Xie <qxie1@email.mot.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP Implementer's guide question...
Message-ID: <20020213122547.H1088@openss7.org>
Reply-To: bidulock@openss7.org
References: <F1CE15E08172D4119247009027AE9D500B5ECF4F@FMSMSX37>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <F1CE15E08172D4119247009027AE9D500B5ECF4F@FMSMSX37>; from s_balani@trillium.com on Wed, Feb 13, 2002 at 10:10:00AM -0800
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Sandeep,

Ivan has put forward a very good reason to not send INIT-ACK
to a potential intruder for an existing association.  This
is because the tie tags in the cookie compromise the security
of the verification tag generator.  (For all associations, not
just the one.)

--brian

On Wed, 13 Feb 2002, Balani, Sandeep wrote:

> Hi Randy,
>  
> Some Comments, 
> 
> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Wednesday, February 13, 2002 6:40 AM
> To: Balani, Sandeep
> Cc: 'bidulock@openss7.org'; Qiaobing Xie; tsvwg@ietf.org
> Subject: Re: [Tsvwg] SCTP Implementer's guide question...
> 
> 
> Balani:
> 
> It is quite unusal for me but I am in complete agreement
> with Brian :-)
> 
> "Balani, Sandeep" wrote:
> > 
> > Hi Brian,
> > 
> >   Somehow i still think that we should do this check when we receive
> > COOKIE-ECHO.
> > Because, even though It send ABORT with vTag=iTag of INIT chunk, It still
> > means that Association is terminated.
> > 
> > When the receiver of INIT respond with ABORT it also send TERMINATION-IND
> to
> > upper user to close the socket. which means that although u may be right
> > that peer won't accept ABORT because of mismatching vTag but this end
> still
> > assume the association to be TERMINATED.
> 
> No it SHOULD NOT send anything up to the upper layer. The ABORT it sends
> with the iTag would NOT be for THIS association. As far as the local
> endpoint that receives the abort is concerned the association has
> NOT restarted, the vtag is incorrect. 
> 
> [ Balani Sandeep] Randy, Can u please explain Why would we not send
> Termination Indication to local user when association is ABORTED. 
> If there was an SCTP exception locally or some illegal event occured in
> which case u send ABORT to peer & terminate the association. 
> 
> One more reason for Terminating the association is in case of RESTART
> because the peer loses all its context it may not remember what addresses it
> sent during first time of association. So in that case even though he
> receives ABORT ( with additional addresses) he has to guess that what
> addresses i sent previously. In that case it will be hit & trial for peer to
> send IP addresses.
>  
> On the other hand if u teminate the assocition he can start with whatever IP
> addresses it wants. Is'nt it ?  
> 
> Brian, I am sorry for my foolishness, but i still don't understand. 
> 
> Lastly the reason to wait for COOKIE-ECHO is because INIT-ACK would be send
> to genuine user because it matched with its association, & that peer will
> dump the INIT-ACK so we would never receive COOKIE-EACHO. that means No need
> to send ABORT.
>  
> Although i agree that extra cpu cycles are utlized. 
> 
> - Sandeep 
> 
> > 
> > On the other hand, If we wait till COOKIE-ECHO,  the intruder can never be
> > able to pass thru the association.
> > 
> > In my understanding getting srcIPaddress and port number (which are
> > standard) is easy.  U r correct that vTag are difficult to guess.
> > 
> 
> I considered this for a LONG time before deciding the right way to
> do it is in response to the INIT itself here is why:
> 
> We have two basic cases:
> 1) The peer has restarted
> and
> 2) An attacker or STRAY INIT is arriving.
> 
> Now for case 1) I think we all will agree it does not
> matter if I wait to send the abort at the cookie-echo or INIT. In
> both case the "restart" is NOT going to happen until one of two
> things occur:
> a) A HB on the now dead association causes an abort (HB from
>    Endpoint-Z to Endpoint-A)
> <or>
> b) The peer drops the extra address and resends its INIT. It could then
>    later use add-ip to add the new address (if it supports the
>    extension) or shutdown the association properly and restart again
>    with the new address.
> 
> So just considering case 1) sending the abort from the INIT makes
> sense (remember it DOES take a bit of cpu to do the hash .. depending
> on the hash algo the implementation is using it can be QUITE a bit
> of cpu). And it quickens the recovery time.
> 
> Now lets consider case 2) the one in contention..
> 
> We have an association:
> 
> Endpoint-A                                 Endpoint-Z
> IP-1                                        IP-3
>     +port5000 <--------------------------->+port6000
> IP-2                                        IP-4
> 
> Now our evil attacker Endpoint-E sends in an INIT 
> after guessing the association is up.
> 
> Endpoint-E-----srcip:INIT(Vtag=0,
>                           iTag=N, 
>                           IPaddress IP-1,
>                           IP-2,IP-E)--->Endpoint-Z
> 
> Now Endpoint-Z sees a NEW address in the list. It can do one of 
> two things:
> a) Respond with a ----Abort(Vtag=N)----->src-ip
> or
> b) send a  -----Cookie()--->src-ip
>    
> If it does a)
> --1) if src-ip is IP-1/IP-2 what happens? Endpoint-A gets the
>      abort() checks the V-tag and dumps it on the floor since
>      it is a old stale Abort .. not this association.
> --2) if src-ip is IP-E then the attacker has failed.
> 
>    In either case Endpoint-Z would NOT tell the application
>    that anything happened. It would possibly LOG the event to
>    a security log. Which would then help point a finger at
>    IP-E but it would not tell the app.
> 
> If it does b)
> --1) It would send a cookie back that it KNOWS that it
>      will reject. Thus wasting some CPU cycles to sign
>      a bad cookie.
> --2) If src-ip is IP-1/IP-2 the peer would get a cookie
>      it did not send and discard it.
> --3) If src-ip is IP-E the evil endpoint would then send
>      the cookie-echo... which would of course be rejected
>      with an abort in the same fashion as if we caught
>      it early.
> 
> So the bottom line is that if you do the detection early 
> it just saves CPU cycles on the victim AND CPU cycles on
> the peer of the vicitim in the attackers case.
> 
> The other neat point is you gain a audit trail of the
> events. When sending a Cookie instead of an abort, if the 
> attacker uses IP-1/IP-2 as src-ip then we are responding 
> normally and create no audit trail for our security people
> to track down... if it uses IP-E then we DO have an audit 
> trail but have done extra CPU computation to get something 
> we could have had with less cost..
> 
> In the end the result is the same though..
> 
> R
> 
> 
> 
> 
> > -----Original Message-----
> > From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > Sent: Tuesday, February 12, 2002 5:31 PM
> > To: Qiaobing Xie
> > Cc: Balani, Sandeep; tsvwg@ietf.org
> > Subject: Re: [Tsvwg] SCTP Implementer's guide question...
> > 
> > Qiaobing,
> > 
> > On Tue, 12 Feb 2002, Qiaobing Xie wrote:
> > 
> > > Brian/Sandeep:
> > >
> > > I believe Sandeep is right. The impl-guide should be corrected. See
> > > below for more...
> > >
> > > "Brian F. G. Bidulock" wrote:
> > > >
> > > > Sandeep,
> > > >
> > > > > >  " Before responding the endpoint MUST check to see if the
> > > > > >    unexpected INIT adds new addresses to the association. If new
> > > > > >    addresses are added to the association, the endpoint MUST
> respond
> > > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected
> INIT
> > > >
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > > > >    into the 'Verification Tag' of the outbound packet carrying the
> > ABORT.
> > > >
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > >
> > > > The sender of the INIT tag must guess the verification tag of the
> > > > existing association and place it in the Initiation Tag field of
> > > > the INIT.  It is of no moment that the vTag is zero in the INIT.
> > >
> > > No. vTag is always zero in the INIT, no matter it's for a fresh
> > > association or a re-started one. The endpoint that restarts would most
> > > likely lose any memory of its previous life.
> > 
> > The vTag is not zero in the ABORT.  As the vTag in the received
> > ABORT must match the existing association's tag (otherwise the
> > ABORT is treated OOTB) the intruder must know the tag of the
> > existinging association to abort it.  Not only that, but the
> > intruder must also know a source address of the attacked host's
> > peer to fake the source address and get it sent there.
> > 
> > It requires the intruder to know both the existing association
> > tag an a valid peer address of the association.  If the intruder
> > knows these things, it might as well send an ABORT to the attacked
> > host faked from the known source address with the known tag value.
> > 
> > Sorry, but there is no 'sploit here that is different from the
> > normal cases for RFC 2960.
> > 
> > >
> > > .....
> > > > > On Tue, 12 Feb 2002, Balani, Sandeep wrote:
> > > > >
> > > > > > Hi all,
> > > > > >   I have a question regarding SCTP implementer's guide. I remember
> > this
> > > > > > issue was discussed during 2nd Bakeoff.
> > > > > >
> > > > > > In
> > > > >
> > http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt
> > > > > > section 5.2.2. We are describing the case of association RESTART
> > when INIT
> > > > > > is received in any state other than CLOSED, COOKIE-ECHOED,
> > COOKIE-WAIT and
> > > > > > SHUTDOWN-ACK-SENT. The text says
> > > > > >
> > > > > >  " Before responding the endpoint MUST check to see if the
> > > > > >    unexpected INIT adds new addresses to the association. If new
> > > > > >    addresses are added to the association, the endpoint MUST
> respond
> > > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected
> INIT
> > > > > >    into the 'Verification Tag' of the outbound packet carrying the
> > ABORT.
> > > > > "
> > > > > >
> > > > > > I think that we should check the additional IP addresses when we
> > receive
> > > > > > COOKIE-ECHO from the peer not when we receive INIT from peer,
> > >
> > > Yes, this check on whether "new addresses are added to the association"
> > > is only performed *after* the endpoint received the COOKIE-ECHO *and*
> > > determined that this is a re-start case.
> > 
> > There is no need to wait for COOKIE-ECHO when then ABORT can be sent in
> > response to INIT.
> > 
> > --brian
> > 
> > >
> > > regards,
> > > -Qiaobing
> > >
> > > > > > because if we
> > > > > > send ABORT when we receive INIT then an *INTRUDER* can break an
> > ESTABLISHED
> > > > > > association by sending INIT on an existing association with
> addition
> > of any
> > > > > > IP addresses.
> > 
> > --
> > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> > 
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 13:48:08 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA17044
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 13:48:08 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA22754
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 13:48:10 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA22351;
	Wed, 13 Feb 2002 13:35:54 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA22277
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 13:35:50 -0500 (EST)
Received: from calliope1.fm.intel.com (fmfdns01.fm.intel.com [132.233.247.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA16622
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 13:35:47 -0500 (EST)
Received: from fmsmsxvs043.fm.intel.com (fmsmsxv043-1.fm.intel.com [132.233.48.128])
	by calliope1.fm.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.50 2002/02/08 23:45:02 root Exp $) with SMTP id SAA27390
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 18:35:48 GMT
Received: from FMSMSX017.fm.intel.com ([132.233.42.196])
 by fmsmsxvs043.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002021310345615408
 ; Wed, 13 Feb 2002 10:34:56 -0800
Received: by fmsmsx017.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <1Y3Z69MG>; Wed, 13 Feb 2002 10:35:47 -0800
Message-ID: <F1CE15E08172D4119247009027AE9D500B5ECF50@FMSMSX37>
From: "Balani, Sandeep" <s_balani@trillium.com>
To: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        "Balani, Sandeep" <s_balani@trillium.com>
Cc: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        Qiaobing Xie<qxie1@email.mot.com>, tsvwg@ietf.org
Subject: RE: [Tsvwg] SCTP Implementer's guide question...
Date: Wed, 13 Feb 2002 10:35:44 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Brian, I never said that we will send INIT-ACK to Intruder.
 
Pls Let me clarify my understanding then u can correct me. 

> We have an association:
> 
> Endpoint-A                                 Endpoint-Z
> IP-1                                        IP-3
>     +port5000 <--------------------------->+port6000
> IP-2                                        IP-4
> 
> Now our evil attacker Endpoint-E sends in an INIT 
> after guessing the association is up.
> 
> Endpoint-E-----srcip:INIT(Vtag=0,
>                           iTag=N, 
>                           IPaddress IP-1,
>                           IP-2,IP-E)--->Endpoint-Z
> 

Now at ENDPOINT Z, because it matches the existing assocaition,with IP-1,
IP-2, Endpoint-Z  will send INIT-ACK not to intruder but to ENDPOINT-A. 

 Endpoint-A     <---------- INIT-ACK-------------- Endpoint-Z 
 ( It dump this
   INIT-ACK )                                      
             
                                                   Endpoint-Z
                                                   ( Never Rcv COOKIE)
  Endpoint-E (Intruder) 
  (IT will never receive 
   any message from Endpoint-Z)

ENDPOINT-A should dump this INIT-ACK so that Endpoint-Z will never receive a
COOKIE-ECHO message. 


Thanks
Sandeep 


-----Original Message-----
From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
Sent: Wednesday, February 13, 2002 10:26 AM
To: Balani, Sandeep
Cc: 'Randall Stewart'; Qiaobing Xie; tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP Implementer's guide question...


Sandeep,

Ivan has put forward a very good reason to not send INIT-ACK
to a potential intruder for an existing association.  This
is because the tie tags in the cookie compromise the security
of the verification tag generator.  (For all associations, not
just the one.)

--brian

On Wed, 13 Feb 2002, Balani, Sandeep wrote:

> Hi Randy,
>  
> Some Comments, 
> 
> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Wednesday, February 13, 2002 6:40 AM
> To: Balani, Sandeep
> Cc: 'bidulock@openss7.org'; Qiaobing Xie; tsvwg@ietf.org
> Subject: Re: [Tsvwg] SCTP Implementer's guide question...
> 
> 
> Balani:
> 
> It is quite unusal for me but I am in complete agreement
> with Brian :-)
> 
> "Balani, Sandeep" wrote:
> > 
> > Hi Brian,
> > 
> >   Somehow i still think that we should do this check when we receive
> > COOKIE-ECHO.
> > Because, even though It send ABORT with vTag=iTag of INIT chunk, It
still
> > means that Association is terminated.
> > 
> > When the receiver of INIT respond with ABORT it also send
TERMINATION-IND
> to
> > upper user to close the socket. which means that although u may be right
> > that peer won't accept ABORT because of mismatching vTag but this end
> still
> > assume the association to be TERMINATED.
> 
> No it SHOULD NOT send anything up to the upper layer. The ABORT it sends
> with the iTag would NOT be for THIS association. As far as the local
> endpoint that receives the abort is concerned the association has
> NOT restarted, the vtag is incorrect. 
> 
> [ Balani Sandeep] Randy, Can u please explain Why would we not send
> Termination Indication to local user when association is ABORTED. 
> If there was an SCTP exception locally or some illegal event occured in
> which case u send ABORT to peer & terminate the association. 
> 
> One more reason for Terminating the association is in case of RESTART
> because the peer loses all its context it may not remember what addresses
it
> sent during first time of association. So in that case even though he
> receives ABORT ( with additional addresses) he has to guess that what
> addresses i sent previously. In that case it will be hit & trial for peer
to
> send IP addresses.
>  
> On the other hand if u teminate the assocition he can start with whatever
IP
> addresses it wants. Is'nt it ?  
> 
> Brian, I am sorry for my foolishness, but i still don't understand. 
> 
> Lastly the reason to wait for COOKIE-ECHO is because INIT-ACK would be
send
> to genuine user because it matched with its association, & that peer will
> dump the INIT-ACK so we would never receive COOKIE-EACHO. that means No
need
> to send ABORT.
>  
> Although i agree that extra cpu cycles are utlized. 
> 
> - Sandeep 
> 
> > 
> > On the other hand, If we wait till COOKIE-ECHO,  the intruder can never
be
> > able to pass thru the association.
> > 
> > In my understanding getting srcIPaddress and port number (which are
> > standard) is easy.  U r correct that vTag are difficult to guess.
> > 
> 
> I considered this for a LONG time before deciding the right way to
> do it is in response to the INIT itself here is why:
> 
> We have two basic cases:
> 1) The peer has restarted
> and
> 2) An attacker or STRAY INIT is arriving.
> 
> Now for case 1) I think we all will agree it does not
> matter if I wait to send the abort at the cookie-echo or INIT. In
> both case the "restart" is NOT going to happen until one of two
> things occur:
> a) A HB on the now dead association causes an abort (HB from
>    Endpoint-Z to Endpoint-A)
> <or>
> b) The peer drops the extra address and resends its INIT. It could then
>    later use add-ip to add the new address (if it supports the
>    extension) or shutdown the association properly and restart again
>    with the new address.
> 
> So just considering case 1) sending the abort from the INIT makes
> sense (remember it DOES take a bit of cpu to do the hash .. depending
> on the hash algo the implementation is using it can be QUITE a bit
> of cpu). And it quickens the recovery time.
> 
> Now lets consider case 2) the one in contention..
> 
> We have an association:
> 
> Endpoint-A                                 Endpoint-Z
> IP-1                                        IP-3
>     +port5000 <--------------------------->+port6000
> IP-2                                        IP-4
> 
> Now our evil attacker Endpoint-E sends in an INIT 
> after guessing the association is up.
> 
> Endpoint-E-----srcip:INIT(Vtag=0,
>                           iTag=N, 
>                           IPaddress IP-1,
>                           IP-2,IP-E)--->Endpoint-Z
> 
> Now Endpoint-Z sees a NEW address in the list. It can do one of 
> two things:
> a) Respond with a ----Abort(Vtag=N)----->src-ip
> or
> b) send a  -----Cookie()--->src-ip
>    
> If it does a)
> --1) if src-ip is IP-1/IP-2 what happens? Endpoint-A gets the
>      abort() checks the V-tag and dumps it on the floor since
>      it is a old stale Abort .. not this association.
> --2) if src-ip is IP-E then the attacker has failed.
> 
>    In either case Endpoint-Z would NOT tell the application
>    that anything happened. It would possibly LOG the event to
>    a security log. Which would then help point a finger at
>    IP-E but it would not tell the app.
> 
> If it does b)
> --1) It would send a cookie back that it KNOWS that it
>      will reject. Thus wasting some CPU cycles to sign
>      a bad cookie.
> --2) If src-ip is IP-1/IP-2 the peer would get a cookie
>      it did not send and discard it.
> --3) If src-ip is IP-E the evil endpoint would then send
>      the cookie-echo... which would of course be rejected
>      with an abort in the same fashion as if we caught
>      it early.
> 
> So the bottom line is that if you do the detection early 
> it just saves CPU cycles on the victim AND CPU cycles on
> the peer of the vicitim in the attackers case.
> 
> The other neat point is you gain a audit trail of the
> events. When sending a Cookie instead of an abort, if the 
> attacker uses IP-1/IP-2 as src-ip then we are responding 
> normally and create no audit trail for our security people
> to track down... if it uses IP-E then we DO have an audit 
> trail but have done extra CPU computation to get something 
> we could have had with less cost..
> 
> In the end the result is the same though..
> 
> R
> 
> 
> 
> 
> > -----Original Message-----
> > From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > Sent: Tuesday, February 12, 2002 5:31 PM
> > To: Qiaobing Xie
> > Cc: Balani, Sandeep; tsvwg@ietf.org
> > Subject: Re: [Tsvwg] SCTP Implementer's guide question...
> > 
> > Qiaobing,
> > 
> > On Tue, 12 Feb 2002, Qiaobing Xie wrote:
> > 
> > > Brian/Sandeep:
> > >
> > > I believe Sandeep is right. The impl-guide should be corrected. See
> > > below for more...
> > >
> > > "Brian F. G. Bidulock" wrote:
> > > >
> > > > Sandeep,
> > > >
> > > > > >  " Before responding the endpoint MUST check to see if the
> > > > > >    unexpected INIT adds new addresses to the association. If new
> > > > > >    addresses are added to the association, the endpoint MUST
> respond
> > > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected
> INIT
> > > >
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > > > >    into the 'Verification Tag' of the outbound packet carrying
the
> > ABORT.
> > > >
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > >
> > > > The sender of the INIT tag must guess the verification tag of the
> > > > existing association and place it in the Initiation Tag field of
> > > > the INIT.  It is of no moment that the vTag is zero in the INIT.
> > >
> > > No. vTag is always zero in the INIT, no matter it's for a fresh
> > > association or a re-started one. The endpoint that restarts would most
> > > likely lose any memory of its previous life.
> > 
> > The vTag is not zero in the ABORT.  As the vTag in the received
> > ABORT must match the existing association's tag (otherwise the
> > ABORT is treated OOTB) the intruder must know the tag of the
> > existinging association to abort it.  Not only that, but the
> > intruder must also know a source address of the attacked host's
> > peer to fake the source address and get it sent there.
> > 
> > It requires the intruder to know both the existing association
> > tag an a valid peer address of the association.  If the intruder
> > knows these things, it might as well send an ABORT to the attacked
> > host faked from the known source address with the known tag value.
> > 
> > Sorry, but there is no 'sploit here that is different from the
> > normal cases for RFC 2960.
> > 
> > >
> > > .....
> > > > > On Tue, 12 Feb 2002, Balani, Sandeep wrote:
> > > > >
> > > > > > Hi all,
> > > > > >   I have a question regarding SCTP implementer's guide. I
remember
> > this
> > > > > > issue was discussed during 2nd Bakeoff.
> > > > > >
> > > > > > In
> > > > >
> > http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt
> > > > > > section 5.2.2. We are describing the case of association RESTART
> > when INIT
> > > > > > is received in any state other than CLOSED, COOKIE-ECHOED,
> > COOKIE-WAIT and
> > > > > > SHUTDOWN-ACK-SENT. The text says
> > > > > >
> > > > > >  " Before responding the endpoint MUST check to see if the
> > > > > >    unexpected INIT adds new addresses to the association. If new
> > > > > >    addresses are added to the association, the endpoint MUST
> respond
> > > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected
> INIT
> > > > > >    into the 'Verification Tag' of the outbound packet carrying
the
> > ABORT.
> > > > > "
> > > > > >
> > > > > > I think that we should check the additional IP addresses when we
> > receive
> > > > > > COOKIE-ECHO from the peer not when we receive INIT from peer,
> > >
> > > Yes, this check on whether "new addresses are added to the
association"
> > > is only performed *after* the endpoint received the COOKIE-ECHO *and*
> > > determined that this is a re-start case.
> > 
> > There is no need to wait for COOKIE-ECHO when then ABORT can be sent in
> > response to INIT.
> > 
> > --brian
> > 
> > >
> > > regards,
> > > -Qiaobing
> > >
> > > > > > because if we
> > > > > > send ABORT when we receive INIT then an *INTRUDER* can break an
> > ESTABLISHED
> > > > > > association by sending INIT on an existing association with
> addition
> > of any
> > > > > > IP addresses.
> > 
> > --
> > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> > 
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 14:03:57 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA17597
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 14:03:57 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA24332
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 14:04:00 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA22906;
	Wed, 13 Feb 2002 13:50:00 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA22875
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 13:49:58 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-217.dallas.net [209.44.42.217])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA17102
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 13:49:51 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id MAA09645;
	Wed, 13 Feb 2002 12:49:29 -0600
Date: Wed, 13 Feb 2002 12:49:29 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: "Balani, Sandeep" <s_balani@trillium.com>
Cc: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        Qiaobing Xie <qxie1@email.mot.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP Implementer's guide question...
Message-ID: <20020213124929.I1088@openss7.org>
Reply-To: bidulock@openss7.org
References: <F1CE15E08172D4119247009027AE9D500B5ECF50@FMSMSX37>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <F1CE15E08172D4119247009027AE9D500B5ECF50@FMSMSX37>; from s_balani@trillium.com on Wed, Feb 13, 2002 at 10:35:44AM -0800
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Sandeep,

All the intruder need do is use source address Endpoint-E
in the INIT.  Then it will get INIT-ACK.  The intruder
uses the tie tags to break both Endpoint A and Endpoint
Z's verification tag generation algorithms.  Then the
intruder sends an ABORT with correct source addresss derived
from the cookie to Endpoint Z and correct verification
tags and destroys a whole bunch of Endpoint Z's existing
associations.  The intruder can do the same for Endpoint
A.

There is no reason to send INIT-ACK at all.  ABORT is
appropriate here as required by the imp guide.

--brian

On Wed, 13 Feb 2002, Balani, Sandeep wrote:

> Brian, I never said that we will send INIT-ACK to Intruder.
>  
> Pls Let me clarify my understanding then u can correct me. 
> 
> > We have an association:
> > 
> > Endpoint-A                                 Endpoint-Z
> > IP-1                                        IP-3
> >     +port5000 <--------------------------->+port6000
> > IP-2                                        IP-4
> > 
> > Now our evil attacker Endpoint-E sends in an INIT 
> > after guessing the association is up.
> > 
> > Endpoint-E-----srcip:INIT(Vtag=0,
> >                           iTag=N, 
> >                           IPaddress IP-1,
> >                           IP-2,IP-E)--->Endpoint-Z
> > 
> 
> Now at ENDPOINT Z, because it matches the existing assocaition,with IP-1,
> IP-2, Endpoint-Z  will send INIT-ACK not to intruder but to ENDPOINT-A. 
> 
>  Endpoint-A     <---------- INIT-ACK-------------- Endpoint-Z 
>  ( It dump this
>    INIT-ACK )                                      
>              
>                                                    Endpoint-Z
>                                                    ( Never Rcv COOKIE)
>   Endpoint-E (Intruder) 
>   (IT will never receive 
>    any message from Endpoint-Z)
> 
> ENDPOINT-A should dump this INIT-ACK so that Endpoint-Z will never receive a
> COOKIE-ECHO message. 
> 
> 
> Thanks
> Sandeep 
> 
> 
> -----Original Message-----
> From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> Sent: Wednesday, February 13, 2002 10:26 AM
> To: Balani, Sandeep
> Cc: 'Randall Stewart'; Qiaobing Xie; tsvwg@ietf.org
> Subject: Re: [Tsvwg] SCTP Implementer's guide question...
> 
> 
> Sandeep,
> 
> Ivan has put forward a very good reason to not send INIT-ACK
> to a potential intruder for an existing association.  This
> is because the tie tags in the cookie compromise the security
> of the verification tag generator.  (For all associations, not
> just the one.)
> 
> --brian
> 
> On Wed, 13 Feb 2002, Balani, Sandeep wrote:
> 
> > Hi Randy,
> >  
> > Some Comments, 
> > 
> > -----Original Message-----
> > From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> > Sent: Wednesday, February 13, 2002 6:40 AM
> > To: Balani, Sandeep
> > Cc: 'bidulock@openss7.org'; Qiaobing Xie; tsvwg@ietf.org
> > Subject: Re: [Tsvwg] SCTP Implementer's guide question...
> > 
> > 
> > Balani:
> > 
> > It is quite unusal for me but I am in complete agreement
> > with Brian :-)
> > 
> > "Balani, Sandeep" wrote:
> > > 
> > > Hi Brian,
> > > 
> > >   Somehow i still think that we should do this check when we receive
> > > COOKIE-ECHO.
> > > Because, even though It send ABORT with vTag=iTag of INIT chunk, It
> still
> > > means that Association is terminated.
> > > 
> > > When the receiver of INIT respond with ABORT it also send
> TERMINATION-IND
> > to
> > > upper user to close the socket. which means that although u may be right
> > > that peer won't accept ABORT because of mismatching vTag but this end
> > still
> > > assume the association to be TERMINATED.
> > 
> > No it SHOULD NOT send anything up to the upper layer. The ABORT it sends
> > with the iTag would NOT be for THIS association. As far as the local
> > endpoint that receives the abort is concerned the association has
> > NOT restarted, the vtag is incorrect. 
> > 
> > [ Balani Sandeep] Randy, Can u please explain Why would we not send
> > Termination Indication to local user when association is ABORTED. 
> > If there was an SCTP exception locally or some illegal event occured in
> > which case u send ABORT to peer & terminate the association. 
> > 
> > One more reason for Terminating the association is in case of RESTART
> > because the peer loses all its context it may not remember what addresses
> it
> > sent during first time of association. So in that case even though he
> > receives ABORT ( with additional addresses) he has to guess that what
> > addresses i sent previously. In that case it will be hit & trial for peer
> to
> > send IP addresses.
> >  
> > On the other hand if u teminate the assocition he can start with whatever
> IP
> > addresses it wants. Is'nt it ?  
> > 
> > Brian, I am sorry for my foolishness, but i still don't understand. 
> > 
> > Lastly the reason to wait for COOKIE-ECHO is because INIT-ACK would be
> send
> > to genuine user because it matched with its association, & that peer will
> > dump the INIT-ACK so we would never receive COOKIE-EACHO. that means No
> need
> > to send ABORT.
> >  
> > Although i agree that extra cpu cycles are utlized. 
> > 
> > - Sandeep 
> > 
> > > 
> > > On the other hand, If we wait till COOKIE-ECHO,  the intruder can never
> be
> > > able to pass thru the association.
> > > 
> > > In my understanding getting srcIPaddress and port number (which are
> > > standard) is easy.  U r correct that vTag are difficult to guess.
> > > 
> > 
> > I considered this for a LONG time before deciding the right way to
> > do it is in response to the INIT itself here is why:
> > 
> > We have two basic cases:
> > 1) The peer has restarted
> > and
> > 2) An attacker or STRAY INIT is arriving.
> > 
> > Now for case 1) I think we all will agree it does not
> > matter if I wait to send the abort at the cookie-echo or INIT. In
> > both case the "restart" is NOT going to happen until one of two
> > things occur:
> > a) A HB on the now dead association causes an abort (HB from
> >    Endpoint-Z to Endpoint-A)
> > <or>
> > b) The peer drops the extra address and resends its INIT. It could then
> >    later use add-ip to add the new address (if it supports the
> >    extension) or shutdown the association properly and restart again
> >    with the new address.
> > 
> > So just considering case 1) sending the abort from the INIT makes
> > sense (remember it DOES take a bit of cpu to do the hash .. depending
> > on the hash algo the implementation is using it can be QUITE a bit
> > of cpu). And it quickens the recovery time.
> > 
> > Now lets consider case 2) the one in contention..
> > 
> > We have an association:
> > 
> > Endpoint-A                                 Endpoint-Z
> > IP-1                                        IP-3
> >     +port5000 <--------------------------->+port6000
> > IP-2                                        IP-4
> > 
> > Now our evil attacker Endpoint-E sends in an INIT 
> > after guessing the association is up.
> > 
> > Endpoint-E-----srcip:INIT(Vtag=0,
> >                           iTag=N, 
> >                           IPaddress IP-1,
> >                           IP-2,IP-E)--->Endpoint-Z
> > 
> > Now Endpoint-Z sees a NEW address in the list. It can do one of 
> > two things:
> > a) Respond with a ----Abort(Vtag=N)----->src-ip
> > or
> > b) send a  -----Cookie()--->src-ip
> >    
> > If it does a)
> > --1) if src-ip is IP-1/IP-2 what happens? Endpoint-A gets the
> >      abort() checks the V-tag and dumps it on the floor since
> >      it is a old stale Abort .. not this association.
> > --2) if src-ip is IP-E then the attacker has failed.
> > 
> >    In either case Endpoint-Z would NOT tell the application
> >    that anything happened. It would possibly LOG the event to
> >    a security log. Which would then help point a finger at
> >    IP-E but it would not tell the app.
> > 
> > If it does b)
> > --1) It would send a cookie back that it KNOWS that it
> >      will reject. Thus wasting some CPU cycles to sign
> >      a bad cookie.
> > --2) If src-ip is IP-1/IP-2 the peer would get a cookie
> >      it did not send and discard it.
> > --3) If src-ip is IP-E the evil endpoint would then send
> >      the cookie-echo... which would of course be rejected
> >      with an abort in the same fashion as if we caught
> >      it early.
> > 
> > So the bottom line is that if you do the detection early 
> > it just saves CPU cycles on the victim AND CPU cycles on
> > the peer of the vicitim in the attackers case.
> > 
> > The other neat point is you gain a audit trail of the
> > events. When sending a Cookie instead of an abort, if the 
> > attacker uses IP-1/IP-2 as src-ip then we are responding 
> > normally and create no audit trail for our security people
> > to track down... if it uses IP-E then we DO have an audit 
> > trail but have done extra CPU computation to get something 
> > we could have had with less cost..
> > 
> > In the end the result is the same though..
> > 
> > R
> > 
> > 
> > 
> > 
> > > -----Original Message-----
> > > From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > > Sent: Tuesday, February 12, 2002 5:31 PM
> > > To: Qiaobing Xie
> > > Cc: Balani, Sandeep; tsvwg@ietf.org
> > > Subject: Re: [Tsvwg] SCTP Implementer's guide question...
> > > 
> > > Qiaobing,
> > > 
> > > On Tue, 12 Feb 2002, Qiaobing Xie wrote:
> > > 
> > > > Brian/Sandeep:
> > > >
> > > > I believe Sandeep is right. The impl-guide should be corrected. See
> > > > below for more...
> > > >
> > > > "Brian F. G. Bidulock" wrote:
> > > > >
> > > > > Sandeep,
> > > > >
> > > > > > >  " Before responding the endpoint MUST check to see if the
> > > > > > >    unexpected INIT adds new addresses to the association. If new
> > > > > > >    addresses are added to the association, the endpoint MUST
> > respond
> > > > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected
> > INIT
> > > > >
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > > > > >    into the 'Verification Tag' of the outbound packet carrying
> the
> > > ABORT.
> > > > >
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > > >
> > > > > The sender of the INIT tag must guess the verification tag of the
> > > > > existing association and place it in the Initiation Tag field of
> > > > > the INIT.  It is of no moment that the vTag is zero in the INIT.
> > > >
> > > > No. vTag is always zero in the INIT, no matter it's for a fresh
> > > > association or a re-started one. The endpoint that restarts would most
> > > > likely lose any memory of its previous life.
> > > 
> > > The vTag is not zero in the ABORT.  As the vTag in the received
> > > ABORT must match the existing association's tag (otherwise the
> > > ABORT is treated OOTB) the intruder must know the tag of the
> > > existinging association to abort it.  Not only that, but the
> > > intruder must also know a source address of the attacked host's
> > > peer to fake the source address and get it sent there.
> > > 
> > > It requires the intruder to know both the existing association
> > > tag an a valid peer address of the association.  If the intruder
> > > knows these things, it might as well send an ABORT to the attacked
> > > host faked from the known source address with the known tag value.
> > > 
> > > Sorry, but there is no 'sploit here that is different from the
> > > normal cases for RFC 2960.
> > > 
> > > >
> > > > .....
> > > > > > On Tue, 12 Feb 2002, Balani, Sandeep wrote:
> > > > > >
> > > > > > > Hi all,
> > > > > > >   I have a question regarding SCTP implementer's guide. I
> remember
> > > this
> > > > > > > issue was discussed during 2nd Bakeoff.
> > > > > > >
> > > > > > > In
> > > > > >
> > > http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt
> > > > > > > section 5.2.2. We are describing the case of association RESTART
> > > when INIT
> > > > > > > is received in any state other than CLOSED, COOKIE-ECHOED,
> > > COOKIE-WAIT and
> > > > > > > SHUTDOWN-ACK-SENT. The text says
> > > > > > >
> > > > > > >  " Before responding the endpoint MUST check to see if the
> > > > > > >    unexpected INIT adds new addresses to the association. If new
> > > > > > >    addresses are added to the association, the endpoint MUST
> > respond
> > > > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected
> > INIT
> > > > > > >    into the 'Verification Tag' of the outbound packet carrying
> the
> > > ABORT.
> > > > > > "
> > > > > > >
> > > > > > > I think that we should check the additional IP addresses when we
> > > receive
> > > > > > > COOKIE-ECHO from the peer not when we receive INIT from peer,
> > > >
> > > > Yes, this check on whether "new addresses are added to the
> association"
> > > > is only performed *after* the endpoint received the COOKIE-ECHO *and*
> > > > determined that this is a re-start case.
> > > 
> > > There is no need to wait for COOKIE-ECHO when then ABORT can be sent in
> > > response to INIT.
> > > 
> > > --brian
> > > 
> > > >
> > > > regards,
> > > > -Qiaobing
> > > >
> > > > > > > because if we
> > > > > > > send ABORT when we receive INIT then an *INTRUDER* can break an
> > > ESTABLISHED
> > > > > > > association by sending INIT on an existing association with
> > addition
> > > of any
> > > > > > > IP addresses.
> > > 
> > > --
> > > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> > >                         ¦ Therefore  all  progress  depends on the ¦
> > >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> > > 
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > 
> > -- 
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 14:14:02 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA17892
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 14:14:01 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA24847
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 14:14:04 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA24080;
	Wed, 13 Feb 2002 14:01:18 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA24052
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 14:01:15 -0500 (EST)
Received: from palrel13.hp.com (palrel13.hp.com [156.153.255.238])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA17492
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 14:01:12 -0500 (EST)
Received: from xparelay2.corp.hp.com (xparelay2.corp.hp.com [15.58.137.112])
	by palrel13.hp.com (Postfix) with ESMTP
	id 159C94004A3; Wed, 13 Feb 2002 11:00:39 -0800 (PST)
Received: from xpabh2.corp.hp.com (xpabh2.corp.hp.com [15.58.136.192])
	by xparelay2.corp.hp.com (Postfix) with ESMTP
	id DFF29B4; Wed, 13 Feb 2002 11:00:28 -0800 (PST)
Received: by xpabh2.corp.hp.com with Internet Mail Service (5.5.2653.19)
	id <1FZ2Y2X2>; Wed, 13 Feb 2002 11:00:28 -0800
Message-ID: <499DC368E25AD411B3F100902740AD650AFC7DB1@xrose03.rose.hp.com>
From: "WENDT,JIM (HP-Roseville,ex1)" <jim_wendt@hp.com>
To: "'Stephen Bailey'" <steph@cs.uchicago.edu>,
        John Hufferd <hufferd@us.ibm.com>
Cc: tsvwg@ietf.org
Subject: RE: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch f
	ramer for TCP
Date: Wed, 13 Feb 2002 11:00:19 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

> The other point to note is that you only play the dice at all if
> resegmentation DOES result from an intermediary, since the source will
> only send segments which are known not to fail false-positively.  This
> technique applies equally to Julian's proposal.

And furthermore, in the case of TUF, only when there has been a
drop of a packet contain a TUF header. Since the length field chains to
the next TUF header, the receiver only has to succumb to probability
when it has lost its way. This would seem to reduce the rate of
occurrence somewhat depending on the packet loss rate.

Jim

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 14:14:59 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA17981
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 14:14:59 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA24923
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 14:15:01 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA24204;
	Wed, 13 Feb 2002 14:02:23 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA24170
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 14:02:19 -0500 (EST)
Received: from sj-msg-core-4.cisco.com (sj-msg-core-4.cisco.com [171.71.163.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA17538
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 14:02:16 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-4.cisco.com (8.11.3/8.9.1) with ESMTP id g1DJ1lt08215;
	Wed, 13 Feb 2002 11:01:47 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABB88948;
	Wed, 13 Feb 2002 11:01:39 -0800 (PST)
Message-ID: <3C6AB811.1D193FAD@stewart.chicago.il.us>
Date: Wed, 13 Feb 2002 13:01:38 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: "Balani, Sandeep" <s_balani@trillium.com>
CC: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        Qiaobing Xie <qxie1@email.mot.com>, tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP Implementer's guide question...
References: <F1CE15E08172D4119247009027AE9D500B5ECF4F@FMSMSX37>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

"Balani, Sandeep" wrote:
> 
> Hi Randy,
> 
> Some Comments,
> 
> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Wednesday, February 13, 2002 6:40 AM
> To: Balani, Sandeep
> Cc: 'bidulock@openss7.org'; Qiaobing Xie; tsvwg@ietf.org
> Subject: Re: [Tsvwg] SCTP Implementer's guide question...
> 
> Balani:
> 
> It is quite unusal for me but I am in complete agreement
> with Brian :-)
> 
> "Balani, Sandeep" wrote:
> >
> > Hi Brian,
> >
> >   Somehow i still think that we should do this check when we receive
> > COOKIE-ECHO.
> > Because, even though It send ABORT with vTag=iTag of INIT chunk, It still
> > means that Association is terminated.
> >
> > When the receiver of INIT respond with ABORT it also send TERMINATION-IND
> to
> > upper user to close the socket. which means that although u may be right
> > that peer won't accept ABORT because of mismatching vTag but this end
> still
> > assume the association to be TERMINATED.
> 
> No it SHOULD NOT send anything up to the upper layer. The ABORT it sends
> with the iTag would NOT be for THIS association. As far as the local
> endpoint that receives the abort is concerned the association has
> NOT restarted, the vtag is incorrect.
> 
> [ Balani Sandeep] Randy, Can u please explain Why would we not send
> Termination Indication to local user when association is ABORTED.
> If there was an SCTP exception locally or some illegal event occured in
> which case u send ABORT to peer & terminate the association.
> 
Nothing has been termianated in the attacker case. Your association
is still up and thus you would not report it (at least in your view).
If it is NOT an attacker you still think the assoc is up even
though the peer has failed and restarted. Either a HB from you
will get you a valid ABORT that WILL get reported in a message
to the ULP. OR the peer will resend a INIT without the new
address and you will send your ULP a restart notification.



> One more reason for Terminating the association is in case of RESTART
> because the peer loses all its context it may not remember what addresses it
> sent during first time of association. So in that case even though he
> receives ABORT ( with additional addresses) he has to guess that what
> addresses i sent previously. In that case it will be hit & trial for peer to
> send IP addresses.

That is the breaks.. there is nothing we can do about this.. Otherwise
we have a serious security hole. The guy that did NOT restart will
eventually get cleared up by the HB.. i.e. the restarting guy
will send an ABORT.

Sending a INIT-ACK with the COOKIE will not change this anway.. and
as Ivan as shown you also show the peer what the tag is of
the association.

> 
> On the other hand if u teminate the assocition he can start with whatever IP
> addresses it wants. Is'nt it ?
> 
> Brian, I am sorry for my foolishness, but i still don't understand.
> 
> Lastly the reason to wait for COOKIE-ECHO is because INIT-ACK would be send
> to genuine user because it matched with its association, & that peer will
> dump the INIT-ACK so we would never receive COOKIE-EACHO. that means No need
> to send ABORT.
> 
> Although i agree that extra cpu cycles are utlized.

I don't know how I can explain it to you any better than my
previous email without you in a room and a white board... sigh..
sorry...

R

> 
> >
> > On the other hand, If we wait till COOKIE-ECHO,  the intruder can never be
> > able to pass thru the association.
> >
> > In my understanding getting srcIPaddress and port number (which are
> > standard) is easy.  U r correct that vTag are difficult to guess.
> >
> 
> I considered this for a LONG time before deciding the right way to
> do it is in response to the INIT itself here is why:
> 
> We have two basic cases:
> 1) The peer has restarted
> and
> 2) An attacker or STRAY INIT is arriving.
> 
> Now for case 1) I think we all will agree it does not
> matter if I wait to send the abort at the cookie-echo or INIT. In
> both case the "restart" is NOT going to happen until one of two
> things occur:
> a) A HB on the now dead association causes an abort (HB from
>    Endpoint-Z to Endpoint-A)
> <or>
> b) The peer drops the extra address and resends its INIT. It could then
>    later use add-ip to add the new address (if it supports the
>    extension) or shutdown the association properly and restart again
>    with the new address.
> 
> So just considering case 1) sending the abort from the INIT makes
> sense (remember it DOES take a bit of cpu to do the hash .. depending
> on the hash algo the implementation is using it can be QUITE a bit
> of cpu). And it quickens the recovery time.
> 
> Now lets consider case 2) the one in contention..
> 
> We have an association:
> 
> Endpoint-A                                 Endpoint-Z
> IP-1                                        IP-3
>     +port5000 <--------------------------->+port6000
> IP-2                                        IP-4
> 
> Now our evil attacker Endpoint-E sends in an INIT
> after guessing the association is up.
> 
> Endpoint-E-----srcip:INIT(Vtag=0,
>                           iTag=N,
>                           IPaddress IP-1,
>                           IP-2,IP-E)--->Endpoint-Z
> 
> Now Endpoint-Z sees a NEW address in the list. It can do one of
> two things:
> a) Respond with a ----Abort(Vtag=N)----->src-ip
> or
> b) send a  -----Cookie()--->src-ip
> 
> If it does a)
> --1) if src-ip is IP-1/IP-2 what happens? Endpoint-A gets the
>      abort() checks the V-tag and dumps it on the floor since
>      it is a old stale Abort .. not this association.
> --2) if src-ip is IP-E then the attacker has failed.
> 
>    In either case Endpoint-Z would NOT tell the application
>    that anything happened. It would possibly LOG the event to
>    a security log. Which would then help point a finger at
>    IP-E but it would not tell the app.
> 
> If it does b)
> --1) It would send a cookie back that it KNOWS that it
>      will reject. Thus wasting some CPU cycles to sign
>      a bad cookie.
> --2) If src-ip is IP-1/IP-2 the peer would get a cookie
>      it did not send and discard it.
> --3) If src-ip is IP-E the evil endpoint would then send
>      the cookie-echo... which would of course be rejected
>      with an abort in the same fashion as if we caught
>      it early.
> 
> So the bottom line is that if you do the detection early
> it just saves CPU cycles on the victim AND CPU cycles on
> the peer of the vicitim in the attackers case.
> 
> The other neat point is you gain a audit trail of the
> events. When sending a Cookie instead of an abort, if the
> attacker uses IP-1/IP-2 as src-ip then we are responding
> normally and create no audit trail for our security people
> to track down... if it uses IP-E then we DO have an audit
> trail but have done extra CPU computation to get something
> we could have had with less cost..
> 
> In the end the result is the same though..
> 
> R
> 
> > -----Original Message-----
> > From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > Sent: Tuesday, February 12, 2002 5:31 PM
> > To: Qiaobing Xie
> > Cc: Balani, Sandeep; tsvwg@ietf.org
> > Subject: Re: [Tsvwg] SCTP Implementer's guide question...
> >
> > Qiaobing,
> >
> > On Tue, 12 Feb 2002, Qiaobing Xie wrote:
> >
> > > Brian/Sandeep:
> > >
> > > I believe Sandeep is right. The impl-guide should be corrected. See
> > > below for more...
> > >
> > > "Brian F. G. Bidulock" wrote:
> > > >
> > > > Sandeep,
> > > >
> > > > > >  " Before responding the endpoint MUST check to see if the
> > > > > >    unexpected INIT adds new addresses to the association. If new
> > > > > >    addresses are added to the association, the endpoint MUST
> respond
> > > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected
> INIT
> > > >
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > > > >    into the 'Verification Tag' of the outbound packet carrying the
> > ABORT.
> > > >
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > >
> > > > The sender of the INIT tag must guess the verification tag of the
> > > > existing association and place it in the Initiation Tag field of
> > > > the INIT.  It is of no moment that the vTag is zero in the INIT.
> > >
> > > No. vTag is always zero in the INIT, no matter it's for a fresh
> > > association or a re-started one. The endpoint that restarts would most
> > > likely lose any memory of its previous life.
> >
> > The vTag is not zero in the ABORT.  As the vTag in the received
> > ABORT must match the existing association's tag (otherwise the
> > ABORT is treated OOTB) the intruder must know the tag of the
> > existinging association to abort it.  Not only that, but the
> > intruder must also know a source address of the attacked host's
> > peer to fake the source address and get it sent there.
> >
> > It requires the intruder to know both the existing association
> > tag an a valid peer address of the association.  If the intruder
> > knows these things, it might as well send an ABORT to the attacked
> > host faked from the known source address with the known tag value.
> >
> > Sorry, but there is no 'sploit here that is different from the
> > normal cases for RFC 2960.
> >
> > >
> > > .....
> > > > > On Tue, 12 Feb 2002, Balani, Sandeep wrote:
> > > > >
> > > > > > Hi all,
> > > > > >   I have a question regarding SCTP implementer's guide. I remember
> > this
> > > > > > issue was discussed during 2nd Bakeoff.
> > > > > >
> > > > > > In
> > > > >
> > http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpguide-03.txt
> > > > > > section 5.2.2. We are describing the case of association RESTART
> > when INIT
> > > > > > is received in any state other than CLOSED, COOKIE-ECHOED,
> > COOKIE-WAIT and
> > > > > > SHUTDOWN-ACK-SENT. The text says
> > > > > >
> > > > > >  " Before responding the endpoint MUST check to see if the
> > > > > >    unexpected INIT adds new addresses to the association. If new
> > > > > >    addresses are added to the association, the endpoint MUST
> respond
> > > > > >    with an ABORT copying the 'Initiation Tag' of the unexpected
> INIT
> > > > > >    into the 'Verification Tag' of the outbound packet carrying the
> > ABORT.
> > > > > "
> > > > > >
> > > > > > I think that we should check the additional IP addresses when we
> > receive
> > > > > > COOKIE-ECHO from the peer not when we receive INIT from peer,
> > >
> > > Yes, this check on whether "new addresses are added to the association"
> > > is only performed *after* the endpoint received the COOKIE-ECHO *and*
> > > determined that this is a re-start case.
> >
> > There is no need to wait for COOKIE-ECHO when then ABORT can be sent in
> > response to INIT.
> >
> > --brian
> >
> > >
> > > regards,
> > > -Qiaobing
> > >
> > > > > > because if we
> > > > > > send ABORT when we receive INIT then an *INTRUDER* can break an
> > ESTABLISHED
> > > > > > association by sending INIT on an existing association with
> addition
> > of any
> > > > > > IP addresses.
> >
> > --
> > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> --
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 14:33:10 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA18494
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 14:33:10 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA26468
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 14:33:13 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA25504;
	Wed, 13 Feb 2002 14:21:27 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA25475
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 14:21:25 -0500 (EST)
Received: from mg03.austin.ibm.com (mg03.austin.ibm.com [192.35.232.20])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA18140
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 14:21:21 -0500 (EST)
Received: from austin.ibm.com (netmail1.austin.ibm.com [9.3.7.138])
	by mg03.austin.ibm.com (AIX4.3/8.9.3/8.9.3) with ESMTP id NAA28552
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 13:18:17 -0600
Received: from austin.ibm.com (death.austin.ibm.com [9.53.216.109])
	by austin.ibm.com (AIX4.3/8.9.3/8.9.3) with ESMTP id NAA16002
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 13:21:21 -0600
Message-ID: <3C6ABA85.6E378890@austin.ibm.com>
Date: Wed, 13 Feb 2002 13:12:05 -0600
From: Jon Grimm <jgrimm@austin.ibm.com>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.4.17lksctp i686)
X-Accept-Language: en
MIME-Version: 1.0
To: "tsvwg@ietf.org" <tsvwg@ietf.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] draft-ietf-tsvwg-sctpsocket-03.txt:
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi all,

In the following (section 5.3.1.1) sac_state event types are named a
little too generically (e.g. RESTART,  SHUTDOWN_COMPLETE).   It would be
nice if those symbolic names were prefixed (e.g.  SAC_ ) to avoid name
collisions in pre-existing programs wishing to exploit SCTP or or
possibly other library headers.    

The same could be said for (section 5.3.1.2) SCTP_PEER_ADDR_CHANGE
events and likely other places too. 

It's a nit, but easily avoidable from the start.  
  
Best Regards,
Jon Grimm




5.3.1.1 SCTP_ASSOC_CHANGE

    Communication notifications inform the ULP that an SCTP association
    has either begun or ended. The identifier for a new association is
    provided by this notificaion. The notification information has the
    following format:

    struct sctp_assoc_change {
       uint16_t sac_type;
       uint16_t sac_flags;
       uint32_t sac_length;
       uint16_t sac_state;
       uint16_t sac_error;
       uint16_t sac_outbound_streams;
       uint16_t sac_inbound_streams;
       sctp_assoc_t sac_assoc_id;
    };

    sac_type:

    It should be SCTP_ASSOC_CHANGE.

    sac_flags: 16 bits (unsigned integer)

    Currently unused.

    sac_length: 32 bits (unsigned integer)

    This field is the total length of the notification data, including
    the notification header.

    sac_state: 32 bits (signed integer)

    This field holds one of a number of values that communicate the
    event that happened to the association.  They include:

    Event Name           Description
    ----------------     ---------------
    COMMUNICATION_UP     A new association is now ready
                         and data may be exchanged with this
                         peer.

    COMMUNICATION_LOST   The association has failed.  The association
                         is now in the closed state.  If SEND FAILED

Stewart et.al.                                                 [Page 23]

Internet Draft       Sockets API extensions for SCTP        January 2002

                         notifications are turned on, a
COMMUNICATION_LOST
                         is followed by a series of SCTP_SEND_FAILED
                         events, one for each outstanding message.

    RESTART              SCTP has detected that the peer has restarted.

    SHUTDOWN_COMPLETE    The association has gracefully closed.

    CANT_START_ASSOC     The association failed to setup. If non
blocking
                         mode is set and data was sent (in the udp
mode),
                         a CANT_START_ASSOC is followed by a series of
                         SCTP_SEND_FAILED events, one for each
outstanding
                         message.

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 14:33:18 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA18513
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 14:33:18 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA26489
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 14:33:21 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA25042;
	Wed, 13 Feb 2002 14:16:53 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA25012
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 14:16:51 -0500 (EST)
Received: from atlrel9.hp.com (atlrel9.hp.com [156.153.255.214])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA18026
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 14:16:48 -0500 (EST)
Received: from xatlrelay1.atl.hp.com (xatlrelay1.atl.hp.com [15.45.89.190])
	by atlrel9.hp.com (Postfix) with ESMTP
	id C0A4080519E; Wed, 13 Feb 2002 14:16:19 -0500 (EST)
Received: from xatlbh1.atl.hp.com (xatlbh1.atl.hp.com [15.45.89.186])
	by xatlrelay1.atl.hp.com (Postfix) with ESMTP
	id 99EB112A; Wed, 13 Feb 2002 14:16:19 -0500 (EST)
Received: by xatlbh1.atl.hp.com with Internet Mail Service (5.5.2653.19)
	id <1RXFQ7FP>; Wed, 13 Feb 2002 14:16:19 -0500
Message-ID: <499DC368E25AD411B3F100902740AD650AFC7DB2@xrose03.rose.hp.com>
From: "WENDT,JIM (HP-Roseville,ex1)" <jim_wendt@hp.com>
To: "'Stephen Bailey'" <steph@cs.uchicago.edu>,
        Julian Satran <Julian_Satran@il.ibm.com>
Cc: rdma@yahoogroups.com, tsvwg@ietf.org
Date: Wed, 13 Feb 2002 14:16:10 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch framer for TC
 P
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

> I see it clearly.  However, this seems to be a case where we can't
> have everything.  If we want to touch the data we can have a
> synchronization technique which can also be used to determine PDU
> alignment (COWS).  If we don't want to touch the data, we can have PDU
> alignment (TUF or NDT) OR synchronization (FIM), but not both.

Not to muddy the waters, but isn't an "aligning FIM" possible
(not that anyone would want to do it), where ULP PDUs (possibly
chunked) are aligned into TCP segments allowing direct placement
of each TCP segment payload, and with FIM to allow location of
subsequent ULP PDUs after a packet drop?

Jim

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 14:43:32 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA18774
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 14:43:32 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA26969
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 14:43:35 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA25231;
	Wed, 13 Feb 2002 14:17:42 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA25199
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 14:17:40 -0500 (EST)
Received: from palrel11.hp.com (palrel11.hp.com [156.153.255.246])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA18050
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 14:17:36 -0500 (EST)
Received: from xparelay1.corp.hp.com (xparelay1.corp.hp.com [15.58.136.173])
	by palrel11.hp.com (Postfix) with ESMTP
	id 05AD5600897; Wed, 13 Feb 2002 11:17:08 -0800 (PST)
Received: from xpabh1.corp.hp.com (xpabh1.corp.hp.com [15.58.136.191])
	by xparelay1.corp.hp.com (Postfix) with ESMTP
	id B69F5E000A4; Wed, 13 Feb 2002 11:17:07 -0800 (PST)
Received: by xpabh1.corp.hp.com with Internet Mail Service (5.5.2653.19)
	id <1SSTSSS5>; Wed, 13 Feb 2002 11:17:07 -0800
Message-ID: <499DC368E25AD411B3F100902740AD650AFC7DB3@xrose03.rose.hp.com>
From: "WENDT,JIM (HP-Roseville,ex1)" <jim_wendt@hp.com>
To: "'Julian Satran'" <julian_satran@il.ibm.com>,
        Stephen Bailey <steph@cs.uchicago.edu>
Cc: John Hufferd <hufferd@us.ibm.com>, rdma@yahoogroups.com, tsvwg@ietf.org
Date: Wed, 13 Feb 2002 11:17:04 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C1B4C3.057E4BF0"
Subject: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch framer for TC
 P
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01C1B4C3.057E4BF0
Content-Type: text/plain;
	charset="iso-8859-1"

The flexibility to synchronize independent of transport segment alignment
would seem
to come at the costs of a significant increase in receiver complexity in
managing the
reconstruction of ULP DDP PDUs across transport segments, and of additional
buffer memory. The ability to directly place individual incoming transport
segments,
while not necessarily providing greater throughput, is a simple, low
meta-data, 
and low buffer-memory approach.
 
Jim

-----Original Message-----
From: Julian Satran [mailto:julian_satran@il.ibm.com]
Sent: Monday, February 11, 2002 11:02 PM
To: Stephen Bailey
Cc: John Hufferd; rdma@yahoogroups.com; tsvwg@ietf.org
Subject: Re: [rdma] a proposal for a different No-Data-Touch framer for TCP



Steph, 

You fail to see that a good synchronization scheme should work with or
without a TCP alignment scheme. 
If the instream alignment can be infered by the ULP the so it can by a
generic DDP. Alignment can somewhat improve performance 
but it is not mandatory neither for a generic nor for a protocol specific
placement scheme.  In other words - steering every TCP packet is the
ultimate effectivenss but not necessarily the only steering scheme.  An IP
steering scheme would work unchanged on every transport. 

Regards, 
Julo 




	Stephen Bailey <steph@cs.uchicago.edu> 


12-02-02 00:41 


        
        To:        "John Hufferd" <hufferd@us.ibm.com> 
        cc:        tsvwg@ietf.org, rdma@yahoogroups.com 
        Subject:        Re: [rdma] a proposal for a different No-Data-Touch
framer for TCP 

       


John,

Thanks for the exhausting analysis.

As Paul points out, you do not want to scan the received segment to
recover synchronization with either of these techniques, since it's
too risky.  The TUF draft specifically makes this point.  Basically,
you get one event per segment.  However, if you want to be really
pessemistic you could say the stream IS segmented into tiny frames
that almost result in scanning.

The TUF draft also says that the use of a TUF-derived PDU containment
property should be discontinued after repeated failure of PDU
containment.  Given Julian's repeated statements that TUF should be
analyzed against an infinite stream, I can conclude that the draft
does not make this point strongly or clearly enough.

The TUF assumption is that there are two distinct, quasi-stable
scenarios: segmentation is preserved or it is not.  Obviously if
segmentation is not preserved, there's no point in trying to use the
PDU containment property because it will never hold.  If a receiver
detect that the PDU containment property does not hold for longer than
would represent a transient condition like PMTU change, it gives up
and stops using it.

Julian's argument that TUF needs to be analyzed against an infinite
stream is a straw man.  You can pick the threshold at which a TUF
receiver gives up to be any number that makes you comfortable (10e6
packets, 10e4, whatever).

I'm not clear on how your analysis incorporates worst case data
pattern distribution assumptions.  There is a ULP running inside the
framing protocol, and if we assume the ULP is efficient, MOST of the
data IT carries is actually controlled by the user of the ULP.  This
user data can just as well be a `search and destroy' sequence that
tries every `eye catcher' in a well-formed context.  This technique
boils down to a blind search for a 32-bit integer, and there's the
additional probability of hitting any particular element of the blind
search pattern (basically like your ex and ey).  Julian's extra 64
bits (salt & digest) are not doing anything for you in this case.
It's as simple as the probability of the resegmented stream hitting
the correct 32-bit number.

Thanks again for your work on this.

Steph

> Steph, Julian,
> The following is an attempt to determine the probability of a false
> positive, both with the TUF approach as defined in
> draft-ietf-tsvwg-tcp-ulp-frame-01.txt, and the semi draft that Julian
> proposed.
> 

------------------------ Yahoo! Groups Sponsor ---------------------~-->
Sponsored by VeriSign - The Value of Trust
Do you need to encrypt all your online transactions? Find
the perfect solution in this FREE Guide from VeriSign.
http://us.click.yahoo.com/vCuuSA/UdiDAA/yigFAA/W6uqlB/TM
---------------------------------------------------------------------~->

To unsubscribe from this group, send an email to:
rdma-unsubscribe@egroups.com



Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ 






Yahoo! Groups Sponsor	

ADVERTISEMENT
 
<http://rd.yahoo.com/M=221000.1882886.3382503.1261774/D=egroupweb/S=17057010
14:HM/A=965714/R=0/O=1/I=brandr-promo-flowersale-alerts-lrecm/*http://shoppi
ng.yahoo.com/promotions/flowers/index.html> 	
 
<http://us.adserver.yahoo.com/l?M=221000.1882886.3382503.1261774/D=egroupmai
l/S=1705701014:HM/A=965714/rand=391276857> 	

To unsubscribe from this group, send an email to:
rdma-unsubscribe@egroups.com



Your use of Yahoo! Groups is subject to the Yahoo!
<http://docs.yahoo.com/info/terms/> Terms of Service. 



------_=_NextPart_001_01C1B4C3.057E4BF0
Content-Type: text/html;
	charset="iso-8859-1"

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">


<META content="MSHTML 5.00.3105.105" name=GENERATOR></HEAD>
<BODY>
<DIV><FONT color=#0000ff face=Arial size=2><SPAN class=640050119-13022002>The 
flexibility to synchronize independent of transport segment alignment would 
seem</SPAN></FONT></DIV>
<DIV><FONT color=#0000ff face=Arial size=2><SPAN class=640050119-13022002>to 
come at the costs of a significant increase in receiver complexity in managing 
the</SPAN></FONT></DIV>
<DIV><FONT color=#0000ff face=Arial size=2><SPAN 
class=640050119-13022002>reconstruction of ULP DDP PDUs across transport 
segments, and of additional</SPAN></FONT></DIV>
<DIV><FONT color=#0000ff face=Arial size=2><SPAN class=640050119-13022002>buffer 
memory. The ability to directly place&nbsp;individual incoming transport 
segments,</SPAN></FONT></DIV>
<DIV><FONT color=#0000ff face=Arial size=2><SPAN class=640050119-13022002>while 
not necessarily providing greater throughput, </SPAN></FONT><FONT color=#0000ff 
face=Arial size=2><SPAN class=640050119-13022002>is a simple, low meta-data, 
</SPAN></FONT></DIV>
<DIV><FONT color=#0000ff face=Arial size=2><SPAN class=640050119-13022002>and 
low buffer-memory approach.</SPAN></FONT></DIV>
<DIV><FONT color=#0000ff face=Arial size=2><SPAN 
class=640050119-13022002></SPAN></FONT>&nbsp;</DIV>
<DIV><FONT color=#0000ff face=Arial size=2><SPAN 
class=640050119-13022002>Jim</SPAN></FONT></DIV>
<BLOCKQUOTE 
style="BORDER-LEFT: #0000ff 2px solid; MARGIN-LEFT: 5px; PADDING-LEFT: 5px">
  <DIV align=left class=OutlookMessageHeader dir=ltr><FONT face=Tahoma 
  size=2>-----Original Message-----<BR><B>From:</B> Julian Satran 
  [mailto:julian_satran@il.ibm.com]<BR><B>Sent:</B> Monday, February 11, 2002 
  11:02 PM<BR><B>To:</B> Stephen Bailey<BR><B>Cc:</B> John Hufferd; 
  rdma@yahoogroups.com; tsvwg@ietf.org<BR><B>Subject:</B> Re: [rdma] a proposal 
  for a different No-Data-Touch framer for TCP<BR><BR></DIV></FONT><BR><FONT 
  face=sans-serif size=2>Steph,</FONT> <BR><BR><FONT face=sans-serif size=2>You 
  fail to see that a good synchronization scheme should work with or without a 
  TCP alignment scheme.</FONT> <BR><FONT face=sans-serif size=2>If the instream 
  alignment can be infered by the ULP the so it can by a generic DDP. Alignment 
  can somewhat improve performance</FONT> <BR><FONT face=sans-serif size=2>but 
  it is not mandatory neither for a generic nor for a protocol specific 
  placement scheme. &nbsp;In other words - steering every TCP packet is the 
  ultimate effectivenss but not necessarily the only steering scheme. &nbsp;An 
  IP steering scheme would work unchanged on every transport.</FONT> 
  <BR><BR><FONT face=sans-serif size=2>Regards,</FONT> <BR><FONT face=sans-serif 
  size=2>Julo</FONT> <BR><BR><BR><BR>
  <TABLE width="100%">
    <TBODY>
    <TR vAlign=top>
      <TD>
      <TD><FONT face=sans-serif size=1><B>Stephen Bailey 
        &lt;steph@cs.uchicago.edu&gt;</B></FONT> 
        <P><FONT face=sans-serif size=1>12-02-02 00:41</FONT> <BR></P>
      <TD><FONT face=Arial size=1>&nbsp; &nbsp; &nbsp; &nbsp; </FONT><BR><FONT 
        face=sans-serif size=1>&nbsp; &nbsp; &nbsp; &nbsp; To: &nbsp; &nbsp; 
        &nbsp; &nbsp;"John Hufferd" &lt;hufferd@us.ibm.com&gt;</FONT> <BR><FONT 
        face=sans-serif size=1>&nbsp; &nbsp; &nbsp; &nbsp; cc: &nbsp; &nbsp; 
        &nbsp; &nbsp;tsvwg@ietf.org, rdma@yahoogroups.com</FONT> <BR><FONT 
        face=sans-serif size=1>&nbsp; &nbsp; &nbsp; &nbsp; Subject: &nbsp; 
        &nbsp; &nbsp; &nbsp;Re: [rdma] a proposal for a different No-Data-Touch 
        framer for TCP</FONT> <BR><BR><FONT face=Arial size=1>&nbsp; &nbsp; 
        &nbsp; &nbsp;</FONT></TR></TBODY></TABLE><BR><BR><FONT face="Courier New" 
  size=2>John,<BR><BR>Thanks for the exhausting analysis.<BR><BR>As Paul points 
  out, you do not want to scan the received segment to<BR>recover 
  synchronization with either of these techniques, since it's<BR>too risky. 
  &nbsp;The TUF draft specifically makes this point. &nbsp;Basically,<BR>you get 
  one event per segment. &nbsp;However, if you want to be really<BR>pessemistic 
  you could say the stream IS segmented into tiny frames<BR>that almost result 
  in scanning.<BR><BR>The TUF draft also says that the use of a TUF-derived PDU 
  containment<BR>property should be discontinued after repeated failure of 
  PDU<BR>containment. &nbsp;Given Julian's repeated statements that TUF should 
  be<BR>analyzed against an infinite stream, I can conclude that the 
  draft<BR>does not make this point strongly or clearly enough.<BR><BR>The TUF 
  assumption is that there are two distinct, quasi-stable<BR>scenarios: 
  segmentation is preserved or it is not. &nbsp;Obviously if<BR>segmentation is 
  not preserved, there's no point in trying to use the<BR>PDU containment 
  property because it will never hold. &nbsp;If a receiver<BR>detect that the 
  PDU containment property does not hold for longer than<BR>would represent a 
  transient condition like PMTU change, it gives up<BR>and stops using 
  it.<BR><BR>Julian's argument that TUF needs to be analyzed against an 
  infinite<BR>stream is a straw man. &nbsp;You can pick the threshold at which a 
  TUF<BR>receiver gives up to be any number that makes you comfortable 
  (10e6<BR>packets, 10e4, whatever).<BR><BR>I'm not clear on how your analysis 
  incorporates worst case data<BR>pattern distribution assumptions. &nbsp;There 
  is a ULP running inside the<BR>framing protocol, and if we assume the ULP is 
  efficient, MOST of the<BR>data IT carries is actually controlled by the user 
  of the ULP. &nbsp;This<BR>user data can just as well be a `search and destroy' 
  sequence that<BR>tries every `eye catcher' in a well-formed context. 
  &nbsp;This technique<BR>boils down to a blind search for a 32-bit integer, and 
  there's the<BR>additional probability of hitting any particular element of the 
  blind<BR>search pattern (basically like your ex and ey). &nbsp;Julian's extra 
  64<BR>bits (salt &amp; digest) are not doing anything for you in this 
  case.<BR>It's as simple as the probability of the resegmented stream 
  hitting<BR>the correct 32-bit number.<BR><BR>Thanks again for your work on 
  this.<BR><BR>Steph<BR><BR>&gt; Steph, Julian,<BR>&gt; The following is an 
  attempt to determine the probability of a false<BR>&gt; positive, both with 
  the TUF approach as defined in<BR>&gt; draft-ietf-tsvwg-tcp-ulp-frame-01.txt, 
  and the semi draft that Julian<BR>&gt; proposed.<BR>&gt; 
  <BR><BR>------------------------ Yahoo! Groups Sponsor 
  ---------------------~--&gt;<BR>Sponsored by VeriSign - The Value of 
  Trust<BR>Do you need to encrypt all your online transactions? Find<BR>the 
  perfect solution in this FREE Guide from 
  VeriSign.<BR>http://us.click.yahoo.com/vCuuSA/UdiDAA/yigFAA/W6uqlB/TM<BR>---------------------------------------------------------------------~-&gt;<BR><BR>To 
  unsubscribe from this group, send an email 
  to:<BR>rdma-unsubscribe@egroups.com<BR><BR><BR><BR>Your use of Yahoo! Groups 
  is subject to http://docs.yahoo.com/info/terms/ <BR></FONT><BR><FONT 
  face="Courier New" size=2><BR></FONT><BR><BR><BR><!-- |**|begin egp html banner|**| -->
  <TABLE border=0 cellPadding=2 cellSpacing=0>
    <TBODY>
    <TR bgColor=#ffffcc>
      <TD align=middle><FONT color=#003399 size=-1><B>Yahoo! Groups 
        Sponsor</B></FONT></TD></TR>
    <TR bgColor=#ffffff>
      <TD align=middle width=470>
        <TABLE border=0 cellPadding=0 cellSpacing=0>
          <TBODY>
          <TR>
            <TD align=middle><FONT face=arial 
              size=-2>ADVERTISEMENT</FONT><BR><A 
              href="http://rd.yahoo.com/M=221000.1882886.3382503.1261774/D=egroupweb/S=1705701014:HM/A=965714/R=0/O=1/I=brandr-promo-flowersale-alerts-lrecm/*http://shopping.yahoo.com/promotions/flowers/index.html" 
              target=_top><IMG alt="" border=0 height=250 
              src="http://us.a1.yimg.com/us.yimg.com/a/an/anchor/shopping/flowers_02/flowers_lrec_new.gif" 
              width=300 NOSEND="1"></A></TD></TR></TBODY></TABLE></TD></TR>
    <TR>
      <TD><IMG alt="" height=1 
        src="http://us.adserver.yahoo.com/l?M=221000.1882886.3382503.1261774/D=egroupmail/S=1705701014:HM/A=965714/rand=391276857" 
        width=1 NOSEND="1"></TD></TR></TBODY></TABLE><!-- |**|end egp html banner|**| --><BR><TT>To 
  unsubscribe from this group, send an email 
  to:<BR>rdma-unsubscribe@egroups.com<BR><BR></TT><BR><BR><TT>Your use of Yahoo! 
  Groups is subject to the <A href="http://docs.yahoo.com/info/terms/">Yahoo! 
  Terms of Service</A>.</TT> <BR></BLOCKQUOTE></BODY></HTML>

------_=_NextPart_001_01C1B4C3.057E4BF0--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 14:50:26 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA18976
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 14:50:25 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA27268
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 14:50:28 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA25185;
	Wed, 13 Feb 2002 14:17:37 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA25109
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 14:17:32 -0500 (EST)
Received: from mgw-x3.nokia.com (mgw-x3.nokia.com [131.228.20.26])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA18047
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 14:17:28 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir01nok.ntc.nokia.com (esvir01nokt.ntc.nokia.com [172.21.143.33])
	by mgw-x3.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1DJHri02204
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 21:17:53 +0200 (EET)
Received: from esebh004.NOE.Nokia.com (unverified) by esvir01nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T590dc42a29ac158f21083@esvir01nok.ntc.nokia.com>;
 Wed, 13 Feb 2002 21:17:29 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh004.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Wed, 13 Feb 2002 21:17:29 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] SCTP Implementer's guide question...
Date: Wed, 13 Feb 2002 21:17:28 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF263257A67@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] SCTP Implementer's guide question...
Thread-Index: AcG0uboBjutVcJJMTTewm857trk1dwAB4rAA
To: <s_balani@trillium.com>, <randall@stewart.chicago.il.us>
Cc: <bidulock@openss7.org>, <qxie1@email.mot.com>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 13 Feb 2002 19:17:29.0119 (UTC) FILETIME=[140E52F0:01C1B4C3]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id OAA25110
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi Balani and the rest,

... snip...

> No it SHOULD NOT send anything up to the upper layer. The 
> ABORT it sends
> with the iTag would NOT be for THIS association. As far as the local
> endpoint that receives the abort is concerned the association has
> NOT restarted, the vtag is incorrect. 
> 
> [ Balani Sandeep] Randy, Can u please explain Why would we not send
> Termination Indication to local user when association is ABORTED. 
> If there was an SCTP exception locally or some illegal event 
> occured in
> which case u send ABORT to peer & terminate the association. 

	You are an endpoint who has a normal association with another endpoint. Suddenly you receive an INIT chunk that seems to be part of this association (but it could be an attacker, and old packet...). You don't terminate or abort anything until you are sure that the peer restarted. That's the reason why you normally make the whole 4-way handshake before you considere that the peer restarted, because you need to check that you are not in any of the other situations (i.e., attacker, old packet...)

	Besides, if you are going to abort an association as soon as you receive an INIT, which kind of protocol do you have? The first guy wanting to close your association does not have more to do than sending you an INIT forging the source address... :-0

> One more reason for Terminating the association is in case of RESTART
> because the peer loses all its context it may not remember 
> what addresses it

	You named it, in case of RESTART... But you don't consider that your peer has restarted until you finish the whole 4 way algorithm. If you finish it, then you say, ok, the peer restarted, and you tell it to the upper user...

	However, in this case you will never terminate the 4-way handshake, because it is risky... You send the ABORT to the peer... But you send the abort to the one who sent you the INIT... If the peer restarted, nothing changes... you will notice it with a HB or other way. If the INIT was sent by an attacker, then nothing changes either, but you don't give it any more information about things such as your VT and stuff...

> sent during first time of association. So in that case even though he
> receives ABORT ( with additional addresses) he has to guess that what
> addresses i sent previously. In that case it will be hit & 

	The ABORT is telling you what are the extra addresses... You just simply have to subtract them from your INIT chunk (the one you have just sent).

> trial for peer to
> send IP addresses.
>  
> On the other hand if u teminate the assocition he can start 
> with whatever IP
> addresses it wants. Is'nt it ?  

	But you must be sure that it is your peer and nobody else the one closing your association...

> Brian, I am sorry for my foolishness, but i still don't understand. 
> 
> Lastly the reason to wait for COOKIE-ECHO is because INIT-ACK 
> would be send
> to genuine user because it matched with its association, & 
> that peer will
> dump the INIT-ACK so we would never receive COOKIE-EACHO. 

	But, you send your INIT ACK to the source of the INIT chunk. If the attacker included an address parameter with one of the addresses or your peer, you will recognize the INIT chunk as being part of that association. Then, if you would send back the INIT ACK, it will go directly to the attacker hands... The attacker may use the information in the INIT ACK to know more about your association that it should...

	I hope this helps a little...

	BR Iván Arias Rodríguez

> that means No need
> to send ABORT.
>  
> Although i agree that extra cpu cycles are utlized. 
> 
> - Sandeep 
> 
> > 
> > On the other hand, If we wait till COOKIE-ECHO,  the 
> intruder can never be
> > able to pass thru the association.
> > 
> > In my understanding getting srcIPaddress and port number (which are
> > standard) is easy.  U r correct that vTag are difficult to guess.
> > 
> 
> I considered this for a LONG time before deciding the right way to
> do it is in response to the INIT itself here is why:
> 
> We have two basic cases:
> 1) The peer has restarted
> and
> 2) An attacker or STRAY INIT is arriving.
> 
> Now for case 1) I think we all will agree it does not
> matter if I wait to send the abort at the cookie-echo or INIT. In
> both case the "restart" is NOT going to happen until one of two
> things occur:
> a) A HB on the now dead association causes an abort (HB from
>    Endpoint-Z to Endpoint-A)
> <or>
> b) The peer drops the extra address and resends its INIT. It 
> could then
>    later use add-ip to add the new address (if it supports the
>    extension) or shutdown the association properly and restart again
>    with the new address.
> 
> So just considering case 1) sending the abort from the INIT makes
> sense (remember it DOES take a bit of cpu to do the hash .. depending
> on the hash algo the implementation is using it can be QUITE a bit
> of cpu). And it quickens the recovery time.
> 
> Now lets consider case 2) the one in contention..
> 
> We have an association:
> 
> Endpoint-A                                 Endpoint-Z
> IP-1                                        IP-3
>     +port5000 <--------------------------->+port6000
> IP-2                                        IP-4
> 
> Now our evil attacker Endpoint-E sends in an INIT 
> after guessing the association is up.
> 
> Endpoint-E-----srcip:INIT(Vtag=0,
>                           iTag=N, 
>                           IPaddress IP-1,
>                           IP-2,IP-E)--->Endpoint-Z
> 
> Now Endpoint-Z sees a NEW address in the list. It can do one of 
> two things:
> a) Respond with a ----Abort(Vtag=N)----->src-ip
> or
> b) send a  -----Cookie()--->src-ip
>    
> If it does a)
> --1) if src-ip is IP-1/IP-2 what happens? Endpoint-A gets the
>      abort() checks the V-tag and dumps it on the floor since
>      it is a old stale Abort .. not this association.
> --2) if src-ip is IP-E then the attacker has failed.
> 
>    In either case Endpoint-Z would NOT tell the application
>    that anything happened. It would possibly LOG the event to
>    a security log. Which would then help point a finger at
>    IP-E but it would not tell the app.
> 
> If it does b)
> --1) It would send a cookie back that it KNOWS that it
>      will reject. Thus wasting some CPU cycles to sign
>      a bad cookie.
> --2) If src-ip is IP-1/IP-2 the peer would get a cookie
>      it did not send and discard it.
> --3) If src-ip is IP-E the evil endpoint would then send
>      the cookie-echo... which would of course be rejected
>      with an abort in the same fashion as if we caught
>      it early.
> 
> So the bottom line is that if you do the detection early 
> it just saves CPU cycles on the victim AND CPU cycles on
> the peer of the vicitim in the attackers case.
> 
> The other neat point is you gain a audit trail of the
> events. When sending a Cookie instead of an abort, if the 
> attacker uses IP-1/IP-2 as src-ip then we are responding 
> normally and create no audit trail for our security people
> to track down... if it uses IP-E then we DO have an audit 
> trail but have done extra CPU computation to get something 
> we could have had with less cost..
> 
> In the end the result is the same though..
> 
> R
> 
> 
> 
> 
> > -----Original Message-----
> > From: Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > Sent: Tuesday, February 12, 2002 5:31 PM
> > To: Qiaobing Xie
> > Cc: Balani, Sandeep; tsvwg@ietf.org
> > Subject: Re: [Tsvwg] SCTP Implementer's guide question...
> > 
> > Qiaobing,
> > 
> > On Tue, 12 Feb 2002, Qiaobing Xie wrote:
> > 
> > > Brian/Sandeep:
> > >
> > > I believe Sandeep is right. The impl-guide should be 
> corrected. See
> > > below for more...
> > >
> > > "Brian F. G. Bidulock" wrote:
> > > >
> > > > Sandeep,
> > > >
> > > > > >  " Before responding the endpoint MUST check to see if the
> > > > > >    unexpected INIT adds new addresses to the 
> association. If new
> > > > > >    addresses are added to the association, the endpoint MUST
> respond
> > > > > >    with an ABORT copying the 'Initiation Tag' of 
> the unexpected
> INIT
> > > >
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > > > >    into the 'Verification Tag' of the outbound 
> packet carrying the
> > ABORT.
> > > >
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > >
> > > > The sender of the INIT tag must guess the verification 
> tag of the
> > > > existing association and place it in the Initiation Tag field of
> > > > the INIT.  It is of no moment that the vTag is zero in the INIT.
> > >
> > > No. vTag is always zero in the INIT, no matter it's for a fresh
> > > association or a re-started one. The endpoint that 
> restarts would most
> > > likely lose any memory of its previous life.
> > 
> > The vTag is not zero in the ABORT.  As the vTag in the received
> > ABORT must match the existing association's tag (otherwise the
> > ABORT is treated OOTB) the intruder must know the tag of the
> > existinging association to abort it.  Not only that, but the
> > intruder must also know a source address of the attacked host's
> > peer to fake the source address and get it sent there.
> > 
> > It requires the intruder to know both the existing association
> > tag an a valid peer address of the association.  If the intruder
> > knows these things, it might as well send an ABORT to the attacked
> > host faked from the known source address with the known tag value.
> > 
> > Sorry, but there is no 'sploit here that is different from the
> > normal cases for RFC 2960.
> > 
> > >
> > > .....
> > > > > On Tue, 12 Feb 2002, Balani, Sandeep wrote:
> > > > >
> > > > > > Hi all,
> > > > > >   I have a question regarding SCTP implementer's 
> guide. I remember
> > this
> > > > > > issue was discussed during 2nd Bakeoff.
> > > > > >
> > > > > > In
> > > > >
> > 
> http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-sctpimpgu
> ide-03.txt
> > > > > > section 5.2.2. We are describing the case of 
> association RESTART
> > when INIT
> > > > > > is received in any state other than CLOSED, COOKIE-ECHOED,
> > COOKIE-WAIT and
> > > > > > SHUTDOWN-ACK-SENT. The text says
> > > > > >
> > > > > >  " Before responding the endpoint MUST check to see if the
> > > > > >    unexpected INIT adds new addresses to the 
> association. If new
> > > > > >    addresses are added to the association, the endpoint MUST
> respond
> > > > > >    with an ABORT copying the 'Initiation Tag' of 
> the unexpected
> INIT
> > > > > >    into the 'Verification Tag' of the outbound 
> packet carrying the
> > ABORT.
> > > > > "
> > > > > >
> > > > > > I think that we should check the additional IP 
> addresses when we
> > receive
> > > > > > COOKIE-ECHO from the peer not when we receive INIT 
> from peer,
> > >
> > > Yes, this check on whether "new addresses are added to 
> the association"
> > > is only performed *after* the endpoint received the 
> COOKIE-ECHO *and*
> > > determined that this is a re-start case.
> > 
> > There is no need to wait for COOKIE-ECHO when then ABORT 
> can be sent in
> > response to INIT.
> > 
> > --brian
> > 
> > >
> > > regards,
> > > -Qiaobing
> > >
> > > > > > because if we
> > > > > > send ABORT when we receive INIT then an *INTRUDER* 
> can break an
> > ESTABLISHED
> > > > > > association by sending INIT on an existing association with
> addition
> > of any
> > > > > > IP addresses.
> > 
> > --
> > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> > 
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 16:11:43 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA20802
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 16:11:43 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA02009
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 16:11:44 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA00246;
	Wed, 13 Feb 2002 15:39:46 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA00166
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 15:39:42 -0500 (EST)
Received: from d12lmsgate.de.ibm.com (d12lmsgate.de.ibm.com [195.212.91.199])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA20117
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 15:39:37 -0500 (EST)
Received: from d12relay01.de.ibm.com (d12relay01.de.ibm.com [9.165.215.22])
	by d12lmsgate.de.ibm.com (1.0.0) with ESMTP id VAA176878;
	Wed, 13 Feb 2002 21:39:03 +0100
Received: from d10hubm1.telaviv.ibm.com (d10ml001.telaviv.ibm.com [9.148.216.55])
	by d12relay01.de.ibm.com (8.11.1m3/NCO v5.01) with ESMTP id g1DKeZM77602;
	Wed, 13 Feb 2002 21:40:36 +0100
To: "WENDT,JIM (HP-Roseville,ex1)" <Jim_Wendt@hp.com>
Cc: John Hufferd <hufferd@us.ibm.com>, rdma@yahoogroups.com,
        Stephen Bailey <steph@cs.uchicago.edu>, tsvwg@ietf.org
MIME-Version: 1.0
X-Mailer: Lotus Notes Release 5.0.9  November 16, 2001
From: "Julian Satran" <Julian_Satran@il.ibm.com>
Message-ID: <OFD27E7D9D.F2B3CD38-ONC2256B5F.006F1B87@telaviv.ibm.com>
Date: Wed, 13 Feb 2002 22:40:37 +0200
X-MIMETrack: Serialize by Router on D10ML001/10/M/IBM(Release 5.0.8 |June 18, 2001) at
 13/02/2002 22:40:40,
	Serialize complete at 13/02/2002 22:40:40
Content-Type: multipart/alternative; boundary="=_alternative 006F4638C2256B5F_="
Subject: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch framer for TC	P
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

This is a multipart message in MIME format.
--=_alternative 006F4638C2256B5F_=
Content-Type: text/plain; charset="us-ascii"

For non-generic hardware DDP and alignment are not synomins.
Neither are they necessarily for generic hardware.

Julo




"WENDT,JIM (HP-Roseville,ex1)" <Jim_Wendt@hp.com>
13-02-02 21:17

 
        To:     Julian Satran/Haifa/IBM@IBMIL, Stephen Bailey <steph@cs.uchicago.edu>
        cc:     John Hufferd/San Jose/IBM@IBMUS, rdma@yahoogroups.com, tsvwg@ietf.org
        Subject:        RE: [rdma] a proposal for a different No-Data-Touch framer for TC       P

 

The flexibility to synchronize independent of transport segment alignment 
would seem
to come at the costs of a significant increase in receiver complexity in 
managing the
reconstruction of ULP DDP PDUs across transport segments, and of 
additional
buffer memory. The ability to directly place individual incoming transport 
segments,
while not necessarily providing greater throughput, is a simple, low 
meta-data, 
and low buffer-memory approach.
 
Jim
-----Original Message-----
From: Julian Satran [mailto:julian_satran@il.ibm.com]
Sent: Monday, February 11, 2002 11:02 PM
To: Stephen Bailey
Cc: John Hufferd; rdma@yahoogroups.com; tsvwg@ietf.org
Subject: Re: [rdma] a proposal for a different No-Data-Touch framer for TCP


Steph, 

You fail to see that a good synchronization scheme should work with or 
without a TCP alignment scheme. 
If the instream alignment can be infered by the ULP the so it can by a 
generic DDP. Alignment can somewhat improve performance 
but it is not mandatory neither for a generic nor for a protocol specific 
placement scheme.  In other words - steering every TCP packet is the 
ultimate effectivenss but not necessarily the only steering scheme.  An IP 
steering scheme would work unchanged on every transport. 

Regards, 
Julo 




Stephen Bailey <steph@cs.uchicago.edu> 
12-02-02 00:41 
        
        To:        "John Hufferd" <hufferd@us.ibm.com> 
        cc:        tsvwg@ietf.org, rdma@yahoogroups.com 
        Subject:        Re: [rdma] a proposal for a different 
No-Data-Touch framer for TCP 

 


John,

Thanks for the exhausting analysis.

As Paul points out, you do not want to scan the received segment to
recover synchronization with either of these techniques, since it's
too risky.  The TUF draft specifically makes this point.  Basically,
you get one event per segment.  However, if you want to be really
pessemistic you could say the stream IS segmented into tiny frames
that almost result in scanning.

The TUF draft also says that the use of a TUF-derived PDU containment
property should be discontinued after repeated failure of PDU
containment.  Given Julian's repeated statements that TUF should be
analyzed against an infinite stream, I can conclude that the draft
does not make this point strongly or clearly enough.

The TUF assumption is that there are two distinct, quasi-stable
scenarios: segmentation is preserved or it is not.  Obviously if
segmentation is not preserved, there's no point in trying to use the
PDU containment property because it will never hold.  If a receiver
detect that the PDU containment property does not hold for longer than
would represent a transient condition like PMTU change, it gives up
and stops using it.

Julian's argument that TUF needs to be analyzed against an infinite
stream is a straw man.  You can pick the threshold at which a TUF
receiver gives up to be any number that makes you comfortable (10e6
packets, 10e4, whatever).

I'm not clear on how your analysis incorporates worst case data
pattern distribution assumptions.  There is a ULP running inside the
framing protocol, and if we assume the ULP is efficient, MOST of the
data IT carries is actually controlled by the user of the ULP.  This
user data can just as well be a `search and destroy' sequence that
tries every `eye catcher' in a well-formed context.  This technique
boils down to a blind search for a 32-bit integer, and there's the
additional probability of hitting any particular element of the blind
search pattern (basically like your ex and ey).  Julian's extra 64
bits (salt & digest) are not doing anything for you in this case.
It's as simple as the probability of the resegmented stream hitting
the correct 32-bit number.

Thanks again for your work on this.

Steph

> Steph, Julian,
> The following is an attempt to determine the probability of a false
> positive, both with the TUF approach as defined in
> draft-ietf-tsvwg-tcp-ulp-frame-01.txt, and the semi draft that Julian
> proposed.
> 

------------------------ Yahoo! Groups Sponsor ---------------------~-->
Sponsored by VeriSign - The Value of Trust
Do you need to encrypt all your online transactions? Find
the perfect solution in this FREE Guide from VeriSign.
http://us.click.yahoo.com/vCuuSA/UdiDAA/yigFAA/W6uqlB/TM
---------------------------------------------------------------------~->

To unsubscribe from this group, send an email to:
rdma-unsubscribe@egroups.com



Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ 






To unsubscribe from this group, send an email to:
rdma-unsubscribe@egroups.com



Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service. 

Yahoo! Groups Sponsor

ADVERTISEMENT




To unsubscribe from this group, send an email to:
rdma-unsubscribe@egroups.com



Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service. 


--=_alternative 006F4638C2256B5F_=
Content-Type: text/html; charset="us-ascii"


<br><font size=2 face="sans-serif">For non-generic hardware DDP and alignment are not synomins.</font>
<br><font size=2 face="sans-serif">Neither are they necessarily for generic hardware.</font>
<br>
<br><font size=2 face="sans-serif">Julo</font>
<br>
<br>
<br>
<table width=100%>
<tr valign=top>
<td>
<td><font size=1 face="sans-serif"><b>&quot;WENDT,JIM (HP-Roseville,ex1)&quot; &lt;Jim_Wendt@hp.com&gt;</b></font>
<p><font size=1 face="sans-serif">13-02-02 21:17</font>
<br>
<td><font size=1 face="Arial">&nbsp; &nbsp; &nbsp; &nbsp; </font>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; To: &nbsp; &nbsp; &nbsp; &nbsp;Julian Satran/Haifa/IBM@IBMIL, Stephen Bailey &lt;steph@cs.uchicago.edu&gt;</font>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; cc: &nbsp; &nbsp; &nbsp; &nbsp;John Hufferd/San Jose/IBM@IBMUS, rdma@yahoogroups.com, tsvwg@ietf.org</font>
<br><font size=1 face="sans-serif">&nbsp; &nbsp; &nbsp; &nbsp; Subject: &nbsp; &nbsp; &nbsp; &nbsp;RE: [rdma] a proposal for a different No-Data-Touch framer for TC &nbsp; &nbsp; &nbsp; &nbsp;P</font>
<br>
<br><font size=1 face="Arial">&nbsp; &nbsp; &nbsp; &nbsp;</font></table>
<br>
<br><font size=2 color=blue face="Arial">The flexibility to synchronize independent of transport segment alignment would seem</font>
<br><font size=2 color=blue face="Arial">to come at the costs of a significant increase in receiver complexity in managing the</font>
<br><font size=2 color=blue face="Arial">reconstruction of ULP DDP PDUs across transport segments, and of additional</font>
<br><font size=2 color=blue face="Arial">buffer memory. The ability to directly place individual incoming transport segments,</font>
<br><font size=2 color=blue face="Arial">while not necessarily providing greater throughput, is a simple, low meta-data, </font>
<br><font size=2 color=blue face="Arial">and low buffer-memory approach.</font>
<br><font size=3 face="Times New Roman">&nbsp;</font>
<br><font size=2 color=blue face="Arial">Jim</font>
<br><font size=2 face="Tahoma">-----Original Message-----<b><br>
From:</b> Julian Satran [mailto:julian_satran@il.ibm.com]<b><br>
Sent:</b> Monday, February 11, 2002 11:02 PM<b><br>
To:</b> Stephen Bailey<b><br>
Cc:</b> John Hufferd; rdma@yahoogroups.com; tsvwg@ietf.org<b><br>
Subject:</b> Re: [rdma] a proposal for a different No-Data-Touch framer for TCP<br>
</font>
<br><font size=2 face="sans-serif"><br>
Steph,</font><font size=3 face="Times New Roman"> <br>
</font><font size=2 face="sans-serif"><br>
You fail to see that a good synchronization scheme should work with or without a TCP alignment scheme.</font><font size=3 face="Times New Roman"> </font><font size=2 face="sans-serif"><br>
If the instream alignment can be infered by the ULP the so it can by a generic DDP. Alignment can somewhat improve performance</font><font size=3 face="Times New Roman"> </font><font size=2 face="sans-serif"><br>
but it is not mandatory neither for a generic nor for a protocol specific placement scheme. &nbsp;In other words - steering every TCP packet is the ultimate effectivenss but not necessarily the only steering scheme. &nbsp;An IP steering scheme would work unchanged on every transport.</font><font size=3 face="Times New Roman"> <br>
</font><font size=2 face="sans-serif"><br>
Regards,</font><font size=3 face="Times New Roman"> </font><font size=2 face="sans-serif"><br>
Julo</font><font size=3 face="Times New Roman"> <br>
<br>
<br>
</font>
<table width=100%>
<tr valign=top>
<td width=1%>
<td width=35%><font size=1 face="sans-serif"><b>Stephen Bailey &lt;steph@cs.uchicago.edu&gt;</b></font><font size=3 face="Times New Roman"> </font>
<p><font size=1 face="sans-serif">12-02-02 00:41</font><font size=3 face="Times New Roman"> </font>
<td width=62%><font size=1 face="Arial">&nbsp; &nbsp; &nbsp; &nbsp; </font><font size=1 face="sans-serif"><br>
 &nbsp; &nbsp; &nbsp; &nbsp;To: &nbsp; &nbsp; &nbsp; &nbsp;&quot;John Hufferd&quot; &lt;hufferd@us.ibm.com&gt;</font><font size=3 face="Times New Roman"> </font><font size=1 face="sans-serif"><br>
 &nbsp; &nbsp; &nbsp; &nbsp;cc: &nbsp; &nbsp; &nbsp; &nbsp;tsvwg@ietf.org, rdma@yahoogroups.com</font><font size=3 face="Times New Roman"> </font><font size=1 face="sans-serif"><br>
 &nbsp; &nbsp; &nbsp; &nbsp;Subject: &nbsp; &nbsp; &nbsp; &nbsp;Re: [rdma] a proposal for a different No-Data-Touch framer for TCP</font><font size=3 face="Times New Roman"> <br>
</font><font size=1 face="Arial"><br>
 &nbsp; &nbsp; &nbsp; </font></table>
<br><font size=3 face="Times New Roman"><br>
</font><font size=2 face="Courier New"><br>
John,<br>
<br>
Thanks for the exhausting analysis.<br>
<br>
As Paul points out, you do not want to scan the received segment to<br>
recover synchronization with either of these techniques, since it's<br>
too risky. &nbsp;The TUF draft specifically makes this point. &nbsp;Basically,<br>
you get one event per segment. &nbsp;However, if you want to be really<br>
pessemistic you could say the stream IS segmented into tiny frames<br>
that almost result in scanning.<br>
<br>
The TUF draft also says that the use of a TUF-derived PDU containment<br>
property should be discontinued after repeated failure of PDU<br>
containment. &nbsp;Given Julian's repeated statements that TUF should be<br>
analyzed against an infinite stream, I can conclude that the draft<br>
does not make this point strongly or clearly enough.<br>
<br>
The TUF assumption is that there are two distinct, quasi-stable<br>
scenarios: segmentation is preserved or it is not. &nbsp;Obviously if<br>
segmentation is not preserved, there's no point in trying to use the<br>
PDU containment property because it will never hold. &nbsp;If a receiver<br>
detect that the PDU containment property does not hold for longer than<br>
would represent a transient condition like PMTU change, it gives up<br>
and stops using it.<br>
<br>
Julian's argument that TUF needs to be analyzed against an infinite<br>
stream is a straw man. &nbsp;You can pick the threshold at which a TUF<br>
receiver gives up to be any number that makes you comfortable (10e6<br>
packets, 10e4, whatever).<br>
<br>
I'm not clear on how your analysis incorporates worst case data<br>
pattern distribution assumptions. &nbsp;There is a ULP running inside the<br>
framing protocol, and if we assume the ULP is efficient, MOST of the<br>
data IT carries is actually controlled by the user of the ULP. &nbsp;This<br>
user data can just as well be a `search and destroy' sequence that<br>
tries every `eye catcher' in a well-formed context. &nbsp;This technique<br>
boils down to a blind search for a 32-bit integer, and there's the<br>
additional probability of hitting any particular element of the blind<br>
search pattern (basically like your ex and ey). &nbsp;Julian's extra 64<br>
bits (salt &amp; digest) are not doing anything for you in this case.<br>
It's as simple as the probability of the resegmented stream hitting<br>
the correct 32-bit number.<br>
<br>
Thanks again for your work on this.<br>
<br>
Steph<br>
<br>
&gt; Steph, Julian,<br>
&gt; The following is an attempt to determine the probability of a false<br>
&gt; positive, both with the TUF approach as defined in<br>
&gt; draft-ietf-tsvwg-tcp-ulp-frame-01.txt, and the semi draft that Julian<br>
&gt; proposed.<br>
&gt; <br>
<br>
------------------------ Yahoo! Groups Sponsor ---------------------~--&gt;<br>
Sponsored by VeriSign - The Value of Trust<br>
Do you need to encrypt all your online transactions? Find<br>
the perfect solution in this FREE Guide from VeriSign.<br>
http://us.click.yahoo.com/vCuuSA/UdiDAA/yigFAA/W6uqlB/TM<br>
---------------------------------------------------------------------~-&gt;<br>
<br>
To unsubscribe from this group, send an email to:<br>
rdma-unsubscribe@egroups.com<br>
<br>
<br>
<br>
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ </font><font size=3 face="Times New Roman"><br>
</font><font size=2 face="Courier New"><br>
</font><font size=3 face="Times New Roman"><br>
<br>
<br>
<br>
</font><font size=3 face="Courier New"><br>
To unsubscribe from this group, send an email to:<br>
rdma-unsubscribe@egroups.com<br>
</font><font size=3 face="Times New Roman"><br>
<br>
</font><font size=3 face="Courier New"><br>
Your use of Yahoo! Groups is subject to the </font><a href=http://docs.yahoo.com/info/terms/><font size=3 color=blue face="Courier New"><u>Yahoo! Terms of Service</u></font></a><font size=3 face="Courier New">.</font><font size=3 face="Times New Roman"> </font>
<br>
<table>
<tr>
<td bgcolor=#ffffcc><font size=2 color=#003399 face="Times New Roman"><b>Yahoo! Groups Sponsor</b></font>
<tr>
<td bgcolor=white>
<table>
<tr>
<td><font size=1 face="Arial">ADVERTISEMENT</font><font size=3 color=blue face="Times New Roman"><u><br>
</u></font><a href="http://rd.yahoo.com/M=218512.1856015.3360178.1269402/D=egroupweb/S=1705701014:HM/A=928901/R=0/*http://ad.doubleclick.net/clk;3788034;6800959;v%3fhttp://www.verisign.com/cgi-bin/go.cgi?a=b228473240057000" target=_blank></a></table>
<br>
<tr>
<td></table>
<br><font size=3 face="Courier New"><br>
To unsubscribe from this group, send an email to:<br>
rdma-unsubscribe@egroups.com<br>
</font><font size=3 face="Times New Roman"><br>
<br>
</font><font size=3 face="Courier New"><br>
Your use of Yahoo! Groups is subject to the </font><a href=http://docs.yahoo.com/info/terms/><font size=3 color=blue face="Courier New"><u>Yahoo! Terms of Service</u></font></a><font size=3 face="Courier New">.</font><font size=3 face="Times New Roman"> </font>
<br>
<br>
--=_alternative 006F4638C2256B5F_=--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 16:15:33 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA20912
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 16:15:33 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA02073
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 16:15:35 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA01566;
	Wed, 13 Feb 2002 16:01:13 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA01511
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 16:01:09 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-217.dallas.net [209.44.42.217])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA20621;
	Wed, 13 Feb 2002 16:01:06 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id PAA11885;
	Wed, 13 Feb 2002 15:00:44 -0600
Date: Wed, 13 Feb 2002 15:00:44 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Jacques Rajchgod <c-Jacques.Rajchgod@wcom.com>
Cc: Sigtran <sigtran@ietf.org>, tsvwg@ietf.org
Message-ID: <20020213150044.M1088@openss7.org>
Reply-To: bidulock@openss7.org
Mail-Followup-To: Jacques Rajchgod <c-Jacques.Rajchgod@wcom.com>,
	Sigtran <sigtran@ietf.org>, tsvwg@ietf.org
References: <NEBBLDAOEDIKLJPELCGEMEHECIAA.c-Jacques.Rajchgod@wcom.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2i
In-Reply-To: <NEBBLDAOEDIKLJPELCGEMEHECIAA.c-Jacques.Rajchgod@wcom.com>; from c-Jacques.Rajchgod@wcom.com on Wed, Feb 13, 2002 at 01:08:11PM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Subject: [Tsvwg] Re: [Sigtran] Issues with SCTP TSN and Cum TSN
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Jacques,

You might check on tsvwg@ietf.org where more SCTP'ers hang
out, but, this is normal behaviour.

The receiver of a SACK with Cummulative Ack of Y-1 should
discard the SACK and do nothing further with it other than
updating a_rwnd because it preceeds the Initial TSN.  This
is how an implementation acks nothing yet updates a_rwnd.
I provided you with a quote from RFC 2960 that the
implementation is allowed to send SACK whenever it wishes
to update a_rwnd.  That includes before data is sent.  If
the receiver of the SACK is aborting the association, IMO
the problem is with the receiver.

--brian


Jacques Rajchgod wrote:                                   (Wed, 13 Feb 2002 13:08:11)
> 
>    Hi, SIGTRAN...
>    
>    
>    
>    Am looking at a following scenario:
>    
>    
>    
>    1) An MGC (a) INIT Chunk is sent.
>    
>    2) The INIT contains an INIT_TSN=X.
>    
>            This is the TSN that a's first DATA pack to a SG (b) will
>    contain.
>    
>    3) The INIT is ACK'ed by a SG (b).
>    
>    4) b's INIT_ACK contains an INIT_TSN=Y.
>    
>            This is the TSN that b's first DATA pack to "a", will
>    contain.
>    
>    5) An exchange of COOKIE + ACK, followed by a couple of HEARTBEATS,
>    take place OK.
>    
>    6) "a" then sends an SCTP Pack where a SACK chunk is bundled with a
>    DATA chunk:
>    
>            The SACK has its TSN set to "X", per a's indication advertised
>    in a's INIT at "1".
>    
>            However, the SCTP pack is found to contain a Payload
>    CUMULATIVE_TSN.
>    
>            The CUM_TSN is set to "Y-1" (b's INIT_TSN minus 1)
>    
>    
>    
>    I'll try to describe the scenario graphically below:
>    
>    
>    
>    1) a SENDS:                        ------------- INIT, INITIAL_TSN=X
>    ----------->
>    
>    2) a RECEIVES from b:        <----------- INIT_ACK, INITIAL_TSN= Y
>    ---------
>    
>    6) a SENDS                        ------------- SACK, INITIAL_TSN=X
>    for SACK, ASP_UP (bundled), CUM_TSN=Y-1 for DATA -------->
>    
>    
>    
>    There is a number of issues I see in this scenario:
>    
>    
>    
>    1) A CUM_TSN is supposed to be sent only AFTER having received data -
>    not before.
>    
>        (Ref: RFC 2960, Section 1.4, Page 10/11, 5th bullet from the top
>    of parag. Section 1.4 has a potential conflict with Section 3.3.4 -->
>    SACK must contain a
>    
>            CUM_TSN_ACK).
>    
>        (Ref: Section 6.2, Page 69 - usage of acknowledgement)
>    
>    2) Since there hadn't been any DATA exchanged yet at the point being
>    described in above scenario, there is no business having a CUM_TSN
>    present in the SCTP pack
>    
>    for PAYLOAD.
>    
>    3) Since there hasn't been any DATA exchanged yet, at the point being
>    described in above scenario, A CUM_TSN present should be ignored by
>    the receiver.
>    
>    
>    
>    Please kindly comment (True/False, or otherwise)...
>    
>    My appreciation,
>    
>    Jacques Rajchgod
>    MCI Worldcom
>    Next Generation System Evaluation
>    Tel: [972] 729-7180
>    E-mail: c-Jacques.Rajchgod@wcom.com

-- 
Brian F. G. Bidulock
bidulock@openss7.org
http://www.openss7.org/

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 13 18:37:56 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA24262
	for <tsvwg-archive@odin.ietf.org>; Wed, 13 Feb 2002 18:37:56 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id SAA09847
	for tsvwg-archive@odin.ietf.org; Wed, 13 Feb 2002 18:37:59 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA08838;
	Wed, 13 Feb 2002 18:22:26 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA08812
	for <tsvwg@optimus.ietf.org>; Wed, 13 Feb 2002 18:22:24 -0500 (EST)
Received: from motgate.mot.com (motgate.mot.com [129.188.136.100])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA24032
	for <tsvwg@ietf.org>; Wed, 13 Feb 2002 18:22:20 -0500 (EST)
Received: [from pobox2.mot.com (pobox2.mot.com [136.182.15.8]) by motgate.mot.com (motgate 2.1) with ESMTP id QAA16059; Wed, 13 Feb 2002 16:22:20 -0700 (MST)]
Received: [from relay2.cig.mot.com (relay2.cig.mot.com [136.182.15.24]) by pobox2.mot.com (MOT-pobox2 2.0) with ESMTP id QAA23169; Wed, 13 Feb 2002 16:22:19 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay2.cig.mot.com (8.11.4/8.11.4) with ESMTP id g1DNM8r16840;
	Wed, 13 Feb 2002 17:22:08 -0600 (CST)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id RAA06705;
	Wed, 13 Feb 2002 17:27:01 -0600 (CST)
Message-ID: <3C6AF5B6.9D47673B@email.mot.com>
Date: Wed, 13 Feb 2002 17:24:38 -0600
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: bidulock@openss7.org, Qiaobing Xie <Qiaobing.Xie@motorola.com>,
        s_balani@trillium.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP Implementer's guide question...
References: <58AEDE27D48F884285007ABF827FF26308EE97@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi, Randy, Ivan and others

I guess I must have missed those discussions overhauling the unexpected
INIT handling in RFC2960.

A couple of comments here:

 - The section title

  "5.2.2 Unexpected INIT in States Other than CLOSED, COOKIE-ECHOED, 
         COOKIE-WAIT and SHUTDOWN-ACK-SENT" 

seems really cryptic now. If it means what it says, would it be better
to change it to:

  "5.2.2 Unexpected INIT in ESTABLISHED, SHUTDOWN-PENDING, SHUTDOWN-
         SENT, or SHUTDOWN-RECEIVED State"

 - The impl-guide new text says:

  "Note: Only when a TCB exists and the association is not in a COOKIE-
   WAIT, COOKIE-ECHOED or SHUTDOWN-ACK-SENT state are the Tie-Tags
   populated with a value other than 0. For a normal association INIT
   (i.e. the endpoint is in the CLOSED state), the Tie-Tags MUST be set
   to 0 (indicating that no previous TCB existed)."

This becomes quite cryptic, too. Would it be better to say:

  "Note: The Tie-Tags in the state cookie MUST be set to 0 in an 
   outbound INIT ACK when the association is in CLOSED, COOKIE-WAIT, 
   COOKIE-ECHOED or SHUTDOWN-ACK-SENT state. In other states, the 
   Tie-Tags MUST be populated with a value other than 0."

-Qiaobing

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 05:33:10 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA14768
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 05:33:10 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id FAA21197
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 05:33:13 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA20370;
	Thu, 14 Feb 2002 05:17:43 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA20341
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 05:17:41 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA14616
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 05:17:37 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir03nok.nokia.com (esvir03nokt.ntc.nokia.com [172.21.143.35])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1E9WaZ16446
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 11:32:36 +0200 (EET)
Received: from esebh004.NOE.Nokia.com (unverified) by esvir03nok.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5910d2c913ac158f23076@esvir03nok.nokia.com>;
 Thu, 14 Feb 2002 11:32:18 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh004.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Thu, 14 Feb 2002 11:32:17 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] SCTP Implementer's guide question...
Date: Thu, 14 Feb 2002 11:32:16 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF26308EE9F@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] SCTP Implementer's guide question...
Thread-Index: AcG05WIjIrB/9GJ/SZaEW7+eCNkRJgAVC38Q
To: <qxie1@email.mot.com>
Cc: <bidulock@openss7.org>, <Qiaobing.Xie@motorola.com>,
        <s_balani@trillium.com>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 14 Feb 2002 09:32:17.0444 (UTC) FILETIME=[7E47A640:01C1B53A]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id FAA20342
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi Qiaobing!

> Hi, Randy, Ivan and others
> 
> I guess I must have missed those discussions overhauling the 
> unexpected
> INIT handling in RFC2960.

	They were offline comments... :-0

> A couple of comments here:
> 
>  - The section title
> 
>   "5.2.2 Unexpected INIT in States Other than CLOSED, COOKIE-ECHOED, 
>          COOKIE-WAIT and SHUTDOWN-ACK-SENT" 
> 
> seems really cryptic now. If it means what it says, would it be better
> to change it to:
> 
>   "5.2.2 Unexpected INIT in ESTABLISHED, SHUTDOWN-PENDING, SHUTDOWN-
>          SENT, or SHUTDOWN-RECEIVED State"

	Probably... :-/

	The problem with the title is that it has been growing... I think initially the CLOSED and SHUTDOWN-RECEIVED state didn't appear...

>  - The impl-guide new text says:
> 
>   "Note: Only when a TCB exists and the association is not in 
> a COOKIE-
>    WAIT, COOKIE-ECHOED or SHUTDOWN-ACK-SENT state are the Tie-Tags
>    populated with a value other than 0. For a normal association INIT
>    (i.e. the endpoint is in the CLOSED state), the Tie-Tags 
> MUST be set
>    to 0 (indicating that no previous TCB existed)."
> 
> This becomes quite cryptic, too. Would it be better to say:
> 
>   "Note: The Tie-Tags in the state cookie MUST be set to 0 in an 
>    outbound INIT ACK when the association is in CLOSED, COOKIE-WAIT, 
>    COOKIE-ECHOED or SHUTDOWN-ACK-SENT state. In other states, the 
>    Tie-Tags MUST be populated with a value other than 0."

	Maybe too... I prefer this one... Or maybe better something like "... INIT ACK when the INIT chunk was received in CLOSED...". An association being in the CLOSED state is maybe a little bit tricky thing... :-/

	Thanks!

	BR Iván Arias Rodríguez

> -Qiaobing
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 06:53:57 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA15552
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 06:53:57 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA23827
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 06:53:58 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA22811;
	Thu, 14 Feb 2002 06:27:28 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA22786
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 06:27:24 -0500 (EST)
Received: from irmgard.exp-math.uni-essen.de (irmgard.exp-math.uni-essen.de [132.252.150.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA15233
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 06:27:21 -0500 (EST)
Received: from there (biene.exp-math.uni-essen.de [132.252.150.215])
	by irmgard.exp-math.uni-essen.de (8.12.2/8.12.2) with SMTP id g1EBRLj9030126;
	Thu, 14 Feb 2002 12:27:22 +0100
Message-Id: <200202141127.g1EBRLj9030126@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain;
  charset="iso-8859-1"
From: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
Organization: University Essen
To: discussion@sctp.de, Hollaus Thomas <thomas.hollaus@siemens.at>,
        "'discussion@sctp.de'" <discussion@sctp.de>
Date: Thu, 14 Feb 2002 13:26:56 +0100
X-Mailer: KMail [version 1.3.1]
References: <D9F2B9CD7BD5D21196BC0800060D9ED6039956EB@vies186a.sie.siemens.at>
In-Reply-To: <D9F2B9CD7BD5D21196BC0800060D9ED6039956EB@vies186a.sie.siemens.at>
Cc: tsvwg@ietf.org
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Subject: [Tsvwg] Possible DoS in SCTP ?  (was: [SCTP-Discussion] receive of unexpected chunks with invalid parameters)
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello Thomas, dear all,

please see my comments below.
I am CC'ing TSVWG to discuss this: does this open up a possibility
for DoS attacks ?

On Thursday, 14. February 2002 08:56, Hollaus Thomas wrote:
> Dear all,
>
> Question to unexpected chunks with invalid mandatory parameters (e.g.
> MIS=0, OS=0, Init-Tag=0):
> As far as I understand the sctp.de implementation, it is possible to attack
> a running association with e.g. an INIT and invalid mandatory parameters.
> After receive of that unexpected chunks the prototype does delete the
> running association (pre 16: sctp-control.c, func scr_init, line 677). Is
> that a correct behaviour according to RFC?

The implementation probably should return an ABORT with an Error Chunk
that specifies the appropriate error code, as per RFC 2960:
      Cause Code
      Value           Cause Code
      ---------      ----------------
       1              Invalid Stream Identifier
       2              Missing Mandatory Parameter
       3              Stale Cookie Error
       4              Out of Resource
       5              Unresolvable Address
       6              Unrecognized Chunk Type
   --> 7              Invalid Mandatory Parameter
       8              Unrecognized Parameters
       9              No User Data
      10              Cookie Received While Shutting Down

However, I am not actually sure that this is a good idea, since that
would make it easily possible (with IP spoofing) to start 
denial-of-service attacks. 
I think the rationale behind this in the RFC was to stop a broken 
client from keeping retransmitting wrong INITs and tell it directly 
WHAT is wrong, and that something is wrong.

Regards,
Andreas

- -- 
- ------------------------------------------------------------------------
Dipl.-Ing. Andreas Jungmaier              |                         _
Computer Networking Technology Group      | ASCII ribbon campaign  ( )
University of Essen                       |  - against HTML email   X
http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8a60YXAsLBNOCSsARAox5AJsE6ZOjMxIdUyH/nsvpuokbGdjeMQCfaTU9
6vbLwcJXTjKeqSx/0zTrepA=
=wK09
-----END PGP SIGNATURE-----

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 07:04:56 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA15810
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 07:04:56 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA24751
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 07:04:58 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA23580;
	Thu, 14 Feb 2002 06:45:13 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA23556
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 06:45:11 -0500 (EST)
Received: from beamer.mchh.siemens.de (beamer.mchh.siemens.de [194.138.158.163])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA15398
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 06:45:08 -0500 (EST)
Received: from moody.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.226])
	by beamer.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id MAA22772;
	Thu, 14 Feb 2002 12:45:03 +0100 (MET)
Received: from mchh246e.demchh201e.icn.siemens.de ([139.21.200.56])
	by moody.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id MAA05024;
	Thu, 14 Feb 2002 12:45:02 +0100 (MET)
Received: by MCHH246E with Internet Mail Service (5.5.2653.19)
	id <1ZZHMR3J>; Thu, 14 Feb 2002 12:45:06 +0100
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF757@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'Andreas Jungmaier'" <ajung@exp-math.uni-essen.de>, discussion@sctp.de,
        Hollaus Thomas <thomas.hollaus@siemens.at>,
        "'discussion@sctp.de'"
	 <discussion@sctp.de>
Cc: tsvwg@ietf.org
Subject: RE: [Tsvwg] Possible DoS in SCTP ?  (was: [SCTP-Discussion] recei
	ve of unexpected chunks with invalid parameters)
Date: Thu, 14 Feb 2002 12:45:04 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Dear all,

if a attacker wants to bring down an association by sending
this kind of INITs he has to guess the verification tag. This
is difficult. And if he knows the verification tag he can do anything.

So I do not see the possibility of a DOS here.

Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Andreas Jungmaier [SMTP:ajung@exp-math.uni-essen.de]
> Sent:	Thursday, February 14, 2002 1:27 PM
> To:	discussion@sctp.de; Hollaus Thomas; 'discussion@sctp.de'
> Cc:	tsvwg@ietf.org
> Subject:	[Tsvwg] Possible DoS in SCTP ?  (was: [SCTP-Discussion] receive of unexpected chunks with invalid parameters)
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hello Thomas, dear all,
> 
> please see my comments below.
> I am CC'ing TSVWG to discuss this: does this open up a possibility
> for DoS attacks ?
> 
> On Thursday, 14. February 2002 08:56, Hollaus Thomas wrote:
> > Dear all,
> >
> > Question to unexpected chunks with invalid mandatory parameters (e.g.
> > MIS=0, OS=0, Init-Tag=0):
> > As far as I understand the sctp.de implementation, it is possible to attack
> > a running association with e.g. an INIT and invalid mandatory parameters.
> > After receive of that unexpected chunks the prototype does delete the
> > running association (pre 16: sctp-control.c, func scr_init, line 677). Is
> > that a correct behaviour according to RFC?
> 
> The implementation probably should return an ABORT with an Error Chunk
> that specifies the appropriate error code, as per RFC 2960:
>       Cause Code
>       Value           Cause Code
>       ---------      ----------------
>        1              Invalid Stream Identifier
>        2              Missing Mandatory Parameter
>        3              Stale Cookie Error
>        4              Out of Resource
>        5              Unresolvable Address
>        6              Unrecognized Chunk Type
>    --> 7              Invalid Mandatory Parameter
>        8              Unrecognized Parameters
>        9              No User Data
>       10              Cookie Received While Shutting Down
> 
> However, I am not actually sure that this is a good idea, since that
> would make it easily possible (with IP spoofing) to start 
> denial-of-service attacks. 
> I think the rationale behind this in the RFC was to stop a broken 
> client from keeping retransmitting wrong INITs and tell it directly 
> WHAT is wrong, and that something is wrong.
> 
> Regards,
> Andreas
> 
> - -- 
> - ------------------------------------------------------------------------
> Dipl.-Ing. Andreas Jungmaier              |                         _
> Computer Networking Technology Group      | ASCII ribbon campaign  ( )
> University of Essen                       |  - against HTML email   X
> http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.0.6 (GNU/Linux)
> Comment: For info see http://www.gnupg.org
> 
> iD8DBQE8a60YXAsLBNOCSsARAox5AJsE6ZOjMxIdUyH/nsvpuokbGdjeMQCfaTU9
> 6vbLwcJXTjKeqSx/0zTrepA=
> =wK09
> -----END PGP SIGNATURE-----
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 08:53:46 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA18107
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 08:53:46 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA28960
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 08:53:49 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA27394;
	Thu, 14 Feb 2002 08:16:46 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA27369
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 08:16:44 -0500 (EST)
Received: from sj-msg-core-3.cisco.com (sj-msg-core-3.cisco.com [171.70.157.152])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA17284
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 08:16:41 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-3.cisco.com (8.11.3/8.9.1) with ESMTP id g1EDG4Z08082;
	Thu, 14 Feb 2002 05:16:04 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABC11144;
	Thu, 14 Feb 2002 05:16:11 -0800 (PST)
Message-ID: <3C6BB89A.DEFF6187@stewart.chicago.il.us>
Date: Thu, 14 Feb 2002 07:16:10 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
CC: discussion@sctp.de, Hollaus Thomas <thomas.hollaus@siemens.at>,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] Possible DoS in SCTP ?  (was: [SCTP-Discussion] receive of 
 unexpected chunks with invalid parameters)
References: <D9F2B9CD7BD5D21196BC0800060D9ED6039956EB@vies186a.sie.siemens.at> <200202141127.g1EBRLj9030126@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Andreas:

I don't think so.. first of all the idea is this (correct
me if I am wrong):

Endpoint-A  <------------assoc----------> Endpoint-Z

Now someone else does this (with invalid mandatory parameters):

Evil-Endpoint :(from-E-A)INIT----------------> Endpoint-Z

Now Endpoint-Z SHOULD send back an ABORT to Evil-Endpoint but
it SHOULD NOT touch the association.. since this is an
overlapping case. As Michael said the only way that Endpoint-A
would tear down the association is if Evil-Endpoint guessed
or knew the V-Tag of the original association AND put that
in the I-Tag in the INIT. In that case (if Evil-Endpoint had
this information) it could just as easily:

a) Send a ABORT() directly to Endpoint-A
b) Send a Message in forged to Endpoint-A .. assuming if it
   has the V-Tag it also has the sequence number of the next
   TSN.

All of these presume more of a man-in-the middle type attack
where the Evil one can see what each side has...

I don't see this as a problem..

R


Andreas Jungmaier wrote:
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hello Thomas, dear all,
> 
> please see my comments below.
> I am CC'ing TSVWG to discuss this: does this open up a possibility
> for DoS attacks ?
> 
> On Thursday, 14. February 2002 08:56, Hollaus Thomas wrote:
> > Dear all,
> >
> > Question to unexpected chunks with invalid mandatory parameters (e.g.
> > MIS=0, OS=0, Init-Tag=0):
> > As far as I understand the sctp.de implementation, it is possible to attack
> > a running association with e.g. an INIT and invalid mandatory parameters.
> > After receive of that unexpected chunks the prototype does delete the
> > running association (pre 16: sctp-control.c, func scr_init, line 677). Is
> > that a correct behaviour according to RFC?
> 
> The implementation probably should return an ABORT with an Error Chunk
> that specifies the appropriate error code, as per RFC 2960:
>       Cause Code
>       Value           Cause Code
>       ---------      ----------------
>        1              Invalid Stream Identifier
>        2              Missing Mandatory Parameter
>        3              Stale Cookie Error
>        4              Out of Resource
>        5              Unresolvable Address
>        6              Unrecognized Chunk Type
>    --> 7              Invalid Mandatory Parameter
>        8              Unrecognized Parameters
>        9              No User Data
>       10              Cookie Received While Shutting Down
> 
> However, I am not actually sure that this is a good idea, since that
> would make it easily possible (with IP spoofing) to start
> denial-of-service attacks.
> I think the rationale behind this in the RFC was to stop a broken
> client from keeping retransmitting wrong INITs and tell it directly
> WHAT is wrong, and that something is wrong.
> 
> Regards,
> Andreas
> 
> - --
> - ------------------------------------------------------------------------
> Dipl.-Ing. Andreas Jungmaier              |                         _
> Computer Networking Technology Group      | ASCII ribbon campaign  ( )
> University of Essen                       |  - against HTML email   X
> http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.0.6 (GNU/Linux)
> Comment: For info see http://www.gnupg.org
> 
> iD8DBQE8a60YXAsLBNOCSsARAox5AJsE6ZOjMxIdUyH/nsvpuokbGdjeMQCfaTU9
> 6vbLwcJXTjKeqSx/0zTrepA=
> =wK09
> -----END PGP SIGNATURE-----
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 09:01:07 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA18350
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 09:01:07 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA29518
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 09:01:09 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA28849;
	Thu, 14 Feb 2002 08:50:57 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA28825
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 08:50:55 -0500 (EST)
Received: from irmgard.exp-math.uni-essen.de (irmgard.exp-math.uni-essen.de [132.252.150.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA18014
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 08:50:51 -0500 (EST)
Received: from there (biene.exp-math.uni-essen.de [132.252.150.215])
	by irmgard.exp-math.uni-essen.de (8.12.2/8.12.2) with SMTP id g1EDopj9009266;
	Thu, 14 Feb 2002 14:50:52 +0100
Message-Id: <200202141350.g1EDopj9009266@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain;
  charset="iso-8859-1"
From: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
Organization: University Essen
To: Randall Stewart <randall@stewart.chicago.il.us>
Subject: Re: [Tsvwg] Possible DoS in SCTP ?  (was: reception of invalid parameters)
Date: Thu, 14 Feb 2002 15:50:32 +0100
X-Mailer: KMail [version 1.3.1]
Cc: discussion@sctp.de, Hollaus Thomas <thomas.hollaus@siemens.at>,
        tsvwg@ietf.org
References: <D9F2B9CD7BD5D21196BC0800060D9ED6039956EB@vies186a.sie.siemens.at> <200202141127.g1EBRLj9030126@irmgard.exp-math.uni-essen.de> <3C6BB89A.DEFF6187@stewart.chicago.il.us>
In-Reply-To: <3C6BB89A.DEFF6187@stewart.chicago.il.us>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Dear all,

I am sorry for not explaining more exactly what I meant.
I am assuming an attack with IP-Spoofing. The evil attacker 
(say E) is spoofing poor victim (say V), using oblivious 
server (say) O. 

Now, 

 Evil-Endpoint E: 
  Source IP V(spoofed)
  Dest IP O
  SCTP-INIT (invalid params) ----------------> Server O
                                   /-------------/  
               SCTP-Abort,Error   /
  Poor Victim V   <--------------/                                 

Now, if E can use V, it could send such packets to a large number
of servers. And then V gets bombed with Aborts.

I agree fully, that an established association is safe, but that
was not the issue I was thinking about. Comments ?

Andreas

> Andreas Jungmaier wrote:
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> >
> > Hello Thomas, dear all,
> >
> > please see my comments below.
> > I am CC'ing TSVWG to discuss this: does this open up a possibility
> > for DoS attacks ?
> >
> > On Thursday, 14. February 2002 08:56, Hollaus Thomas wrote:
> > > Dear all,
> > >
> > > Question to unexpected chunks with invalid mandatory parameters (e.g.
> > > MIS=0, OS=0, Init-Tag=0):
> > > As far as I understand the sctp.de implementation, it is possible to
> > > attack a running association with e.g. an INIT and invalid mandatory
> > > parameters. After receive of that unexpected chunks the prototype does
> > > delete the running association (pre 16: sctp-control.c, func scr_init,
> > > line 677). Is that a correct behaviour according to RFC?
> >
> > The implementation probably should return an ABORT with an Error Chunk
> > that specifies the appropriate error code, as per RFC 2960:
> >       Cause Code
> >       Value           Cause Code
> >       ---------      ----------------
> >        1              Invalid Stream Identifier
> >        2              Missing Mandatory Parameter
> >        3              Stale Cookie Error
> >        4              Out of Resource
> >        5              Unresolvable Address
> >        6              Unrecognized Chunk Type
> >    --> 7              Invalid Mandatory Parameter
> >        8              Unrecognized Parameters
> >        9              No User Data
> >       10              Cookie Received While Shutting Down
> >
> > However, I am not actually sure that this is a good idea, since that
> > would make it easily possible (with IP spoofing) to start
> > denial-of-service attacks.
> > I think the rationale behind this in the RFC was to stop a broken
> > client from keeping retransmitting wrong INITs and tell it directly
> > WHAT is wrong, and that something is wrong.
- -- 
- ------------------------------------------------------------------------
Dipl.-Ing. Andreas Jungmaier              |                         _
Computer Networking Technology Group      | ASCII ribbon campaign  ( )
University of Essen                       |  - against HTML email   X
http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8a864XAsLBNOCSsARAlUzAKCYTPXsNH2p23b3DAKM0/Ye7m73XACfZ9I8
a8KoEnC8KTIbBut4WLULsrI=
=6OgC
-----END PGP SIGNATURE-----

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 09:17:12 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA18589
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 09:17:11 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA00044
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 09:17:14 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA29637;
	Thu, 14 Feb 2002 09:05:20 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA29611
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 09:05:18 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA18400
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 09:05:14 -0500 (EST)
From: john.loughney@nokia.com
Received: from esvir04nok.ntc.nokia.com (esvir04nokt.ntc.nokia.com [172.21.143.36])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1EE5MZ02810
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 16:05:22 +0200 (EET)
Received: from esebh003.NOE.Nokia.com (unverified) by esvir04nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5911cca8d4ac158f24077@esvir04nok.ntc.nokia.com>;
 Thu, 14 Feb 2002 16:05:14 +0200
Received: from esebe004.NOE.Nokia.com ([172.21.138.44]) by esebh003.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Thu, 14 Feb 2002 16:05:14 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] Possible DoS in SCTP ?  (was: reception of invalid parameters)
Date: Thu, 14 Feb 2002 16:05:13 +0200
Message-ID: <0C1353ABB1DEB74DB067ADFF749C4EEF09BCBE@esebe004.NOE.Nokia.com>
Thread-Topic: [Tsvwg] Possible DoS in SCTP ?  (was: reception of invalid parameters)
Thread-Index: AcG1XsSXhMCtaHoAQceV+vX42K2pOAAAO4/Q
To: <ajung@exp-math.uni-essen.de>, <randall@stewart.chicago.il.us>
Cc: <discussion@sctp.de>, <thomas.hollaus@siemens.at>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 14 Feb 2002 14:05:14.0641 (UTC) FILETIME=[9FD99810:01C1B560]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id JAA29612
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi Andreas,

> Now, if E can use V, it could send such packets to a large number
> of servers. And then V gets bombed with Aborts.
> 
> I agree fully, that an established association is safe, but that
> was not the issue I was thinking about. Comments ?

If V does not have established SCTP associations, it will just ignore
the aborts.  If it does have established associations, it will need
to at least check to see if it is concerned about the Aborts.  This
will steal some cycles, but is this a really big threat?

John

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 09:59:31 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA23546
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 09:59:31 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA01893
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 09:59:34 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA01240;
	Thu, 14 Feb 2002 09:41:11 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA01214
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 09:41:09 -0500 (EST)
Received: from sj-msg-core-1.cisco.com (sj-msg-core-1.cisco.com [171.71.163.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA22659
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 09:41:05 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-1.cisco.com (8.11.3/8.9.1) with ESMTP id g1EEebh25187;
	Thu, 14 Feb 2002 06:40:37 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABC11945;
	Thu, 14 Feb 2002 06:40:35 -0800 (PST)
Message-ID: <3C6BCC62.347756C0@stewart.chicago.il.us>
Date: Thu, 14 Feb 2002 08:40:34 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
CC: discussion@sctp.de, Hollaus Thomas <thomas.hollaus@siemens.at>,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] Possible DoS in SCTP ?  (was: reception of invalid 
 parameters)
References: <D9F2B9CD7BD5D21196BC0800060D9ED6039956EB@vies186a.sie.siemens.at> <200202141127.g1EBRLj9030126@irmgard.exp-math.uni-essen.de> <3C6BB89A.DEFF6187@stewart.chicago.il.us> <200202141350.g1EDopj9009266@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Andreas:

There is no amplification here though... so it is no
different than if Evil-E sent the aborts itself
to all the "victims" V... it must send N packets
to get N packets to victims... this is no
different than a UDP attack.. in fact a UDP attack
is easier...

R


Andreas Jungmaier wrote:
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Dear all,
> 
> I am sorry for not explaining more exactly what I meant.
> I am assuming an attack with IP-Spoofing. The evil attacker
> (say E) is spoofing poor victim (say V), using oblivious
> server (say) O.
> 
> Now,
> 
>  Evil-Endpoint E:
>   Source IP V(spoofed)
>   Dest IP O
>   SCTP-INIT (invalid params) ----------------> Server O
>                                    /-------------/
>                SCTP-Abort,Error   /
>   Poor Victim V   <--------------/
> 
> Now, if E can use V, it could send such packets to a large number
> of servers. And then V gets bombed with Aborts.
> 
> I agree fully, that an established association is safe, but that
> was not the issue I was thinking about. Comments ?
> 
> Andreas
> 
> > Andreas Jungmaier wrote:
> > > -----BEGIN PGP SIGNED MESSAGE-----
> > > Hash: SHA1
> > >
> > > Hello Thomas, dear all,
> > >
> > > please see my comments below.
> > > I am CC'ing TSVWG to discuss this: does this open up a possibility
> > > for DoS attacks ?
> > >
> > > On Thursday, 14. February 2002 08:56, Hollaus Thomas wrote:
> > > > Dear all,
> > > >
> > > > Question to unexpected chunks with invalid mandatory parameters (e.g.
> > > > MIS=0, OS=0, Init-Tag=0):
> > > > As far as I understand the sctp.de implementation, it is possible to
> > > > attack a running association with e.g. an INIT and invalid mandatory
> > > > parameters. After receive of that unexpected chunks the prototype does
> > > > delete the running association (pre 16: sctp-control.c, func scr_init,
> > > > line 677). Is that a correct behaviour according to RFC?
> > >
> > > The implementation probably should return an ABORT with an Error Chunk
> > > that specifies the appropriate error code, as per RFC 2960:
> > >       Cause Code
> > >       Value           Cause Code
> > >       ---------      ----------------
> > >        1              Invalid Stream Identifier
> > >        2              Missing Mandatory Parameter
> > >        3              Stale Cookie Error
> > >        4              Out of Resource
> > >        5              Unresolvable Address
> > >        6              Unrecognized Chunk Type
> > >    --> 7              Invalid Mandatory Parameter
> > >        8              Unrecognized Parameters
> > >        9              No User Data
> > >       10              Cookie Received While Shutting Down
> > >
> > > However, I am not actually sure that this is a good idea, since that
> > > would make it easily possible (with IP spoofing) to start
> > > denial-of-service attacks.
> > > I think the rationale behind this in the RFC was to stop a broken
> > > client from keeping retransmitting wrong INITs and tell it directly
> > > WHAT is wrong, and that something is wrong.
> - --
> - ------------------------------------------------------------------------
> Dipl.-Ing. Andreas Jungmaier              |                         _
> Computer Networking Technology Group      | ASCII ribbon campaign  ( )
> University of Essen                       |  - against HTML email   X
> http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.0.6 (GNU/Linux)
> Comment: For info see http://www.gnupg.org
> 
> iD8DBQE8a864XAsLBNOCSsARAlUzAKCYTPXsNH2p23b3DAKM0/Ye7m73XACfZ9I8
> a8KoEnC8KTIbBut4WLULsrI=
> =6OgC
> -----END PGP SIGNATURE-----
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 10:34:11 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA24392
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 10:34:11 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA04191
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 10:34:13 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA02855;
	Thu, 14 Feb 2002 10:11:13 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA02785
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 10:11:08 -0500 (EST)
Received: from mgw-x3.nokia.com (mgw-x3.nokia.com [131.228.20.26])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA23821
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 10:11:00 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir01nok.ntc.nokia.com (esvir01nokt.ntc.nokia.com [172.21.143.33])
	by mgw-x3.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1EFBQi19397
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 17:11:26 +0200 (EET)
Received: from esebh001.NOE.Nokia.com (unverified) by esvir01nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T591208e29dac158f21083@esvir01nok.ntc.nokia.com>;
 Thu, 14 Feb 2002 17:11:01 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh001.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Thu, 14 Feb 2002 17:11:01 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] Possible DoS in SCTP ?  (was: reception of invalid parameters)
Date: Thu, 14 Feb 2002 17:11:01 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF263257A82@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] Possible DoS in SCTP ?  (was: reception of invalid parameters)
Thread-Index: AcG1XsSXhMCtaHoAQceV+vX42K2pOAAAO4/QAAJvVpA=
To: <john.loughney@nokia.com>, <ajung@exp-math.uni-essen.de>,
        <randall@stewart.chicago.il.us>
Cc: <discussion@sctp.de>, <thomas.hollaus@siemens.at>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 14 Feb 2002 15:11:01.0516 (UTC) FILETIME=[D05EECC0:01C1B569]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id KAA02786
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi,

	Moreover, if the attacker can make IP spoofing, why using another host as a "reflection"? It could directly send those ABORTs directly to the attacked host...

	BR Iván Arias Rodríguez

> -----Original Message-----
> From: Loughney John (NRC/Helsinki) 
> Sent: 14. February 2002 16:05
> To: ajung@exp-math.uni-essen.de; randall@stewart.chicago.il.us
> Cc: discussion@sctp.de; thomas.hollaus@siemens.at; tsvwg@ietf.org
> Subject: RE: [Tsvwg] Possible DoS in SCTP ? (was: reception of invalid
> parameters)
> 
> 
> Hi Andreas,
> 
> > Now, if E can use V, it could send such packets to a large number
> > of servers. And then V gets bombed with Aborts.
> > 
> > I agree fully, that an established association is safe, but that
> > was not the issue I was thinking about. Comments ?
> 
> If V does not have established SCTP associations, it will just ignore
> the aborts.  If it does have established associations, it will need
> to at least check to see if it is concerned about the Aborts.  This
> will steal some cycles, but is this a really big threat?
> 
> John
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 10:51:46 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA25133
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 10:51:44 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA05542
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 10:51:46 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA04274;
	Thu, 14 Feb 2002 10:36:52 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA04244
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 10:36:50 -0500 (EST)
Received: from irmgard.exp-math.uni-essen.de (irmgard.exp-math.uni-essen.de [132.252.150.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA24566
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 10:36:45 -0500 (EST)
Received: from there (biene.exp-math.uni-essen.de [132.252.150.215])
	by irmgard.exp-math.uni-essen.de (8.12.2/8.12.2) with SMTP id g1EFalj9009562;
	Thu, 14 Feb 2002 16:36:47 +0100
Message-Id: <200202141536.g1EFalj9009562@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain;
  charset="iso-8859-1"
From: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
Organization: University Essen
To: Randall Stewart <randall@stewart.chicago.il.us>, <john.loughney@nokia.com>
Subject: Re: [Tsvwg] Possible DoS in SCTP ? 
Date: Thu, 14 Feb 2002 17:36:28 +0100
X-Mailer: KMail [version 1.3.1]
Cc: discussion@sctp.de, Hollaus Thomas <thomas.hollaus@siemens.at>,
        tsvwg@ietf.org
References: <D9F2B9CD7BD5D21196BC0800060D9ED6039956EB@vies186a.sie.siemens.at> <200202141350.g1EDopj9009266@irmgard.exp-math.uni-essen.de> <3C6BCC62.347756C0@stewart.chicago.il.us>
In-Reply-To: <3C6BCC62.347756C0@stewart.chicago.il.us>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Randy, 

please see my comments inline.

On Thursday, 14. February 2002 15:40, Randall Stewart wrote:
> Andreas:
>
> There is no amplification here though... so it is no
> different than if Evil-E sent the aborts itself
> to all the "victims" V... it must send N packets
> to get N packets to victims... this is no
> different than a UDP attack.. in fact a UDP attack
> is easier...

Well, *almost*. The only problem I see here is that many
network installations by now have already adapted to UDP
attacks, e.g. by filtering excessive amounts (or simply all)
ICMP dest. unreachble packets (which are subsequently 
jamming the targets/victims network).

With SCTP, however, it is not that simple: you get back
a transport layer packet that is much harder to filter. 
Think about it: there have been MANY distributed DoS attacks
based on UDP, why shouldn't there be some based on SCTP some
day ?

BTW, John: I am not concerned about cycles for processing
ABORTs. The problem really arises when an attacker starts a lot
of "Evil Endpoints E", that all have one target: Victim V.
The scarce resource is bandwidth.


> Andreas Jungmaier wrote:
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> >
> > Dear all,
> >
> > I am sorry for not explaining more exactly what I meant.
> > I am assuming an attack with IP-Spoofing. The evil attacker
> > (say E) is spoofing poor victim (say V), using oblivious
> > server (say) O.
> >
> > Now,
> >
> >  Evil-Endpoint E:
> >   Source IP V(spoofed)
> >   Dest IP O
> >   SCTP-INIT (invalid params) ----------------> Server O
> >                                    /-------------/
> >                SCTP-Abort,Error   /
> >   Poor Victim V   <--------------/
> >
> > Now, if E can use V, it could send such packets to a large number
> > of servers. And then V gets bombed with Aborts.
> >
> > I agree fully, that an established association is safe, but that
> > was not the issue I was thinking about. Comments ?
> >

- -- 
- ------------------------------------------------------------------------
Dipl.-Ing. Andreas Jungmaier              |                         _
Computer Networking Technology Group      | ASCII ribbon campaign  ( )
University of Essen                       |  - against HTML email   X
http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8a+eMXAsLBNOCSsARAmPfAJ47qG/Zy9+7omHnCOkbhsNR1b73QwCfZSQz
vIKQ3YdsJjzX46WI8XX/Qbk=
=g5mQ
-----END PGP SIGNATURE-----

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 11:00:20 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA25293
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 11:00:20 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA06013
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 11:00:22 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA05018;
	Thu, 14 Feb 2002 10:44:01 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA04987
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 10:43:58 -0500 (EST)
Received: from sj-msg-core-3.cisco.com (sj-msg-core-3.cisco.com [171.70.157.152])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA24939
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 10:43:55 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-3.cisco.com (8.11.3/8.9.1) with ESMTP id g1EFhJZ05343;
	Thu, 14 Feb 2002 07:43:19 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABC12837;
	Thu, 14 Feb 2002 07:43:25 -0800 (PST)
Message-ID: <3C6BDB1D.C5BE6F4E@stewart.chicago.il.us>
Date: Thu, 14 Feb 2002 09:43:25 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
CC: john.loughney@nokia.com, discussion@sctp.de,
        Hollaus Thomas <thomas.hollaus@siemens.at>, tsvwg@ietf.org
Subject: Re: [Tsvwg] Possible DoS in SCTP ?
References: <D9F2B9CD7BD5D21196BC0800060D9ED6039956EB@vies186a.sie.siemens.at> <200202141350.g1EDopj9009266@irmgard.exp-math.uni-essen.de> <3C6BCC62.347756C0@stewart.chicago.il.us> <200202141536.g1EFalj9009562@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Andreas Jungmaier wrote:
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Randy,
> 
> please see my comments inline.
> 
> On Thursday, 14. February 2002 15:40, Randall Stewart wrote:
> > Andreas:
> >
> > There is no amplification here though... so it is no
> > different than if Evil-E sent the aborts itself
> > to all the "victims" V... it must send N packets
> > to get N packets to victims... this is no
> > different than a UDP attack.. in fact a UDP attack
> > is easier...
> 
> Well, *almost*. The only problem I see here is that many
> network installations by now have already adapted to UDP
> attacks, e.g. by filtering excessive amounts (or simply all)
> ICMP dest. unreachble packets (which are subsequently
> jamming the targets/victims network).
> 
> With SCTP, however, it is not that simple: you get back
> a transport layer packet that is much harder to filter.

How is SCTP harder to filter than UDP.. UDP IS a transport
protocol and the ports in the UDP/TCP header are in the
same place as the SCTP header...


> Think about it: there have been MANY distributed DoS attacks
> based on UDP, why shouldn't there be some based on SCTP some
> day ?

Or TCP.. at this level though your attacker is limited
to the bandwidth/cpu of the Evil-E machine. This is NO
different than TCP or UDP or anything else out there
now...

I fail to see how it is an issue..

R


> 
> BTW, John: I am not concerned about cycles for processing
> ABORTs. The problem really arises when an attacker starts a lot
> of "Evil Endpoints E", that all have one target: Victim V.
> The scarce resource is bandwidth.
> 
> > Andreas Jungmaier wrote:
> > > -----BEGIN PGP SIGNED MESSAGE-----
> > > Hash: SHA1
> > >
> > > Dear all,
> > >
> > > I am sorry for not explaining more exactly what I meant.
> > > I am assuming an attack with IP-Spoofing. The evil attacker
> > > (say E) is spoofing poor victim (say V), using oblivious
> > > server (say) O.
> > >
> > > Now,
> > >
> > >  Evil-Endpoint E:
> > >   Source IP V(spoofed)
> > >   Dest IP O
> > >   SCTP-INIT (invalid params) ----------------> Server O
> > >                                    /-------------/
> > >                SCTP-Abort,Error   /
> > >   Poor Victim V   <--------------/
> > >
> > > Now, if E can use V, it could send such packets to a large number
> > > of servers. And then V gets bombed with Aborts.
> > >
> > > I agree fully, that an established association is safe, but that
> > > was not the issue I was thinking about. Comments ?
> > >
> 
> - --
> - ------------------------------------------------------------------------
> Dipl.-Ing. Andreas Jungmaier              |                         _
> Computer Networking Technology Group      | ASCII ribbon campaign  ( )
> University of Essen                       |  - against HTML email   X
> http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.0.6 (GNU/Linux)
> Comment: For info see http://www.gnupg.org
> 
> iD8DBQE8a+eMXAsLBNOCSsARAmPfAJ47qG/Zy9+7omHnCOkbhsNR1b73QwCfZSQz
> vIKQ3YdsJjzX46WI8XX/Qbk=
> =g5mQ
> -----END PGP SIGNATURE-----

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 11:25:31 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA26074
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 11:25:31 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA08483
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 11:25:33 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA06844;
	Thu, 14 Feb 2002 11:08:36 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA06815
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 11:08:34 -0500 (EST)
Received: from sj-msg-core-3.cisco.com (sj-msg-core-3.cisco.com [171.70.157.152])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA25549
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 11:08:30 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-3.cisco.com (8.11.3/8.9.1) with ESMTP id g1EG7oZ18085;
	Thu, 14 Feb 2002 08:07:50 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABC13345;
	Thu, 14 Feb 2002 08:07:56 -0800 (PST)
Message-ID: <3C6BE0DB.C21D59EC@stewart.chicago.il.us>
Date: Thu, 14 Feb 2002 10:07:56 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
CC: john.loughney@nokia.com, Hollaus Thomas <thomas.hollaus@siemens.at>,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] Possible DoS in SCTP ?
References: <D9F2B9CD7BD5D21196BC0800060D9ED6039956EB@vies186a.sie.siemens.at> <200202141536.g1EFalj9009562@irmgard.exp-math.uni-essen.de> <3C6BDB1D.C5BE6F4E@stewart.chicago.il.us> <200202141603.g1EG3jj9028496@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Andreas Jungmaier wrote:
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hi all,
> please see my comments inline.
> 
> > > > Andreas:
> > > >
> > > > There is no amplification here though... so it is no
> > > > different than if Evil-E sent the aborts itself
> > > > to all the "victims" V... it must send N packets
> > > > to get N packets to victims... this is no
> > > > different than a UDP attack.. in fact a UDP attack
> > > > is easier...
> > >
> > > Well, *almost*. The only problem I see here is that many
> > > network installations by now have already adapted to UDP
> > > attacks, e.g. by filtering excessive amounts (or simply all)
> > > ICMP dest. unreachble packets (which are subsequently
> > > jamming the targets/victims network).
> > >
> > > With SCTP, however, it is not that simple: you get back
> > > a transport layer packet that is much harder to filter.
> >
> > How is SCTP harder to filter than UDP.. UDP IS a transport
> > protocol and the ports in the UDP/TCP header are in the
> > same place as the SCTP header...
> 
> Yes, but the UDP does not cause the DoS -- it is the ICMP
> destination unreachable packets sent in reply to a UDP packet, that
> cause the problems.
> 

Ok.. ICMP may be used so it does NOT get filtered.. but
in SCTP's case if you do the bounce.. you respond with an
ABORT.. which CAN be filtered.. of course it kills the
association in the end .. but that is what filtering would do..

> > > Think about it: there have been MANY distributed DoS attacks
> > > based on UDP, why shouldn't there be some based on SCTP some
> > > day ?
> >
> > Or TCP.. at this level though your attacker is limited
> > to the bandwidth/cpu of the Evil-E machine. This is NO
> > different than TCP or UDP or anything else out there
> > now...
> 
> No, that was my point. In typical DDoS scenarios, there is
> thousands of Evil-E's, directed by one central process. This has
> been done by e.g. the Stacheldraht based DDoS attacks.
> I am not saying this is much of an issue, I am saying that
> SCTP can be used, as other protocols, for (D)DoS attacks.


So what is the issue.. none I think... TCP and UDP have these
same problems... 

SCTP IS a transport protocol and thus it to gets the same
issues as UDP and TCP..

R

> 
> Best regards,
> Andreas
> 
> > > BTW, John: I am not concerned about cycles for processing
> > > ABORTs. The problem really arises when an attacker starts a lot
> > > of "Evil Endpoints E", that all have one target: Victim V.
> > > The scarce resource is bandwidth.
> > >
> > > > Andreas Jungmaier wrote:
> > > > > -----BEGIN PGP SIGNED MESSAGE-----
> > > > > Hash: SHA1
> > > > >
> > > > > Dear all,
> > > > >
> > > > > I am sorry for not explaining more exactly what I meant.
> > > > > I am assuming an attack with IP-Spoofing. The evil attacker
> > > > > (say E) is spoofing poor victim (say V), using oblivious
> > > > > server (say) O.
> > > > >
> > > > > Now,
> > > > >
> > > > >  Evil-Endpoint E:
> > > > >   Source IP V(spoofed)
> > > > >   Dest IP O
> > > > >   SCTP-INIT (invalid params) ----------------> Server O
> > > > >                                    /-------------/
> > > > >                SCTP-Abort,Error   /
> > > > >   Poor Victim V   <--------------/
> > > > >
> > > > > Now, if E can use V, it could send such packets to a large number
> > > > > of servers. And then V gets bombed with Aborts.
> > > > >
> > > > > I agree fully, that an established association is safe, but that
> > > > > was not the issue I was thinking about. Comments ?
> > >
> > > - --
> > > -
> > > ------------------------------------------------------------------------
> > > Dipl.-Ing. Andreas Jungmaier              |                         _
> > > Computer Networking Technology Group      | ASCII ribbon campaign  ( )
> > > University of Essen                       |  - against HTML email   X
> > > http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
> > > -----BEGIN PGP SIGNATURE-----
> > > Version: GnuPG v1.0.6 (GNU/Linux)
> > > Comment: For info see http://www.gnupg.org
> > >
> > > iD8DBQE8a+eMXAsLBNOCSsARAmPfAJ47qG/Zy9+7omHnCOkbhsNR1b73QwCfZSQz
> > > vIKQ3YdsJjzX46WI8XX/Qbk=
> > > =g5mQ
> > > -----END PGP SIGNATURE-----
> 
> - --
> - ------------------------------------------------------------------------
> Dipl.-Ing. Andreas Jungmaier              |                         _
> Computer Networking Technology Group      | ASCII ribbon campaign  ( )
> University of Essen                       |  - against HTML email   X
> http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.0.6 (GNU/Linux)
> Comment: For info see http://www.gnupg.org
> 
> iD8DBQE8a+3fXAsLBNOCSsARAm4GAJ9P8xD5loz/k4p2eSHeYefJj/3uEgCfbM15
> WsAQnnBB9g6F3Vsmyqqj0xI=
> =TK7O
> -----END PGP SIGNATURE-----

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 11:26:02 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA26110
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 11:26:01 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA08524
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 11:26:04 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA06512;
	Thu, 14 Feb 2002 11:03:52 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA06486
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 11:03:48 -0500 (EST)
Received: from irmgard.exp-math.uni-essen.de (irmgard.exp-math.uni-essen.de [132.252.150.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA25367
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 11:03:45 -0500 (EST)
Received: from there (biene.exp-math.uni-essen.de [132.252.150.215])
	by irmgard.exp-math.uni-essen.de (8.12.2/8.12.2) with SMTP id g1EG3jj9028496;
	Thu, 14 Feb 2002 17:03:45 +0100
Message-Id: <200202141603.g1EG3jj9028496@irmgard.exp-math.uni-essen.de>
Content-Type: text/plain;
  charset="iso-8859-1"
From: Andreas Jungmaier <ajung@exp-math.uni-essen.de>
Organization: University Essen
To: Randall Stewart <randall@stewart.chicago.il.us>
Subject: Re: [Tsvwg] Possible DoS in SCTP ?
Date: Thu, 14 Feb 2002 18:03:27 +0100
X-Mailer: KMail [version 1.3.1]
Cc: john.loughney@nokia.com, discussion@sctp.de,
        Hollaus Thomas <thomas.hollaus@siemens.at>, tsvwg@ietf.org
References: <D9F2B9CD7BD5D21196BC0800060D9ED6039956EB@vies186a.sie.siemens.at> <200202141536.g1EFalj9009562@irmgard.exp-math.uni-essen.de> <3C6BDB1D.C5BE6F4E@stewart.chicago.il.us>
In-Reply-To: <3C6BDB1D.C5BE6F4E@stewart.chicago.il.us>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi all,
please see my comments inline.

> > > Andreas:
> > >
> > > There is no amplification here though... so it is no
> > > different than if Evil-E sent the aborts itself
> > > to all the "victims" V... it must send N packets
> > > to get N packets to victims... this is no
> > > different than a UDP attack.. in fact a UDP attack
> > > is easier...
> >
> > Well, *almost*. The only problem I see here is that many
> > network installations by now have already adapted to UDP
> > attacks, e.g. by filtering excessive amounts (or simply all)
> > ICMP dest. unreachble packets (which are subsequently
> > jamming the targets/victims network).
> >
> > With SCTP, however, it is not that simple: you get back
> > a transport layer packet that is much harder to filter.
>
> How is SCTP harder to filter than UDP.. UDP IS a transport
> protocol and the ports in the UDP/TCP header are in the
> same place as the SCTP header...

Yes, but the UDP does not cause the DoS -- it is the ICMP
destination unreachable packets sent in reply to a UDP packet, that
cause the problems.

> > Think about it: there have been MANY distributed DoS attacks
> > based on UDP, why shouldn't there be some based on SCTP some
> > day ?
>
> Or TCP.. at this level though your attacker is limited
> to the bandwidth/cpu of the Evil-E machine. This is NO
> different than TCP or UDP or anything else out there
> now...

No, that was my point. In typical DDoS scenarios, there is
thousands of Evil-E's, directed by one central process. This has
been done by e.g. the Stacheldraht based DDoS attacks.
I am not saying this is much of an issue, I am saying that 
SCTP can be used, as other protocols, for (D)DoS attacks.

Best regards,
Andreas



> > BTW, John: I am not concerned about cycles for processing
> > ABORTs. The problem really arises when an attacker starts a lot
> > of "Evil Endpoints E", that all have one target: Victim V.
> > The scarce resource is bandwidth.
> >
> > > Andreas Jungmaier wrote:
> > > > -----BEGIN PGP SIGNED MESSAGE-----
> > > > Hash: SHA1
> > > >
> > > > Dear all,
> > > >
> > > > I am sorry for not explaining more exactly what I meant.
> > > > I am assuming an attack with IP-Spoofing. The evil attacker
> > > > (say E) is spoofing poor victim (say V), using oblivious
> > > > server (say) O.
> > > >
> > > > Now,
> > > >
> > > >  Evil-Endpoint E:
> > > >   Source IP V(spoofed)
> > > >   Dest IP O
> > > >   SCTP-INIT (invalid params) ----------------> Server O
> > > >                                    /-------------/
> > > >                SCTP-Abort,Error   /
> > > >   Poor Victim V   <--------------/
> > > >
> > > > Now, if E can use V, it could send such packets to a large number
> > > > of servers. And then V gets bombed with Aborts.
> > > >
> > > > I agree fully, that an established association is safe, but that
> > > > was not the issue I was thinking about. Comments ?
> >
> > - --
> > -
> > ------------------------------------------------------------------------
> > Dipl.-Ing. Andreas Jungmaier              |                         _
> > Computer Networking Technology Group      | ASCII ribbon campaign  ( )
> > University of Essen                       |  - against HTML email   X
> > http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
> > -----BEGIN PGP SIGNATURE-----
> > Version: GnuPG v1.0.6 (GNU/Linux)
> > Comment: For info see http://www.gnupg.org
> >
> > iD8DBQE8a+eMXAsLBNOCSsARAmPfAJ47qG/Zy9+7omHnCOkbhsNR1b73QwCfZSQz
> > vIKQ3YdsJjzX46WI8XX/Qbk=
> > =g5mQ
> > -----END PGP SIGNATURE-----

- -- 
- ------------------------------------------------------------------------
Dipl.-Ing. Andreas Jungmaier              |                         _
Computer Networking Technology Group      | ASCII ribbon campaign  ( )
University of Essen                       |  - against HTML email   X
http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8a+3fXAsLBNOCSsARAm4GAJ9P8xD5loz/k4p2eSHeYefJj/3uEgCfbM15
WsAQnnBB9g6F3Vsmyqqj0xI=
=TK7O
-----END PGP SIGNATURE-----

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 11:28:52 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA26238
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 11:28:51 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA08699
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 11:28:54 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA07183;
	Thu, 14 Feb 2002 11:10:16 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA07105
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 11:10:12 -0500 (EST)
Received: from gorilla.mchh.siemens.de (gorilla.mchh.siemens.de [194.138.158.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA25611
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 11:10:08 -0500 (EST)
Received: from blues.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.227])
	by gorilla.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id RAA11162;
	Thu, 14 Feb 2002 17:10:09 +0100 (MET)
Received: from mchh274e.demchh201e.icn.siemens.de ([139.21.200.84])
	by blues.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id RAA12981;
	Thu, 14 Feb 2002 17:10:02 +0100 (MET)
Received: by MCHH274E with Internet Mail Service (5.5.2653.19)
	id <1DHMP208>; Thu, 14 Feb 2002 17:10:07 +0100
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF75C@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'Andreas Jungmaier'" <ajung@exp-math.uni-essen.de>,
        Randall Stewart
	 <randall@stewart.chicago.il.us>
Cc: john.loughney@nokia.com, discussion@sctp.de,
        Hollaus Thomas
	 <thomas.hollaus@siemens.at>, tsvwg@ietf.org
Subject: RE: [Tsvwg] Possible DoS in SCTP ?
Date: Thu, 14 Feb 2002 17:10:07 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Dear all,

if I understand DDoS attacks correcly, they use some mechanism
to generate a lot of packets by sending one. Using the SCTP
feature currently being the discussed you are generating one
packet by sending one. You can send this directly.

Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Andreas Jungmaier [SMTP:ajung@exp-math.uni-essen.de]
> Sent:	Thursday, February 14, 2002 6:03 PM
> To:	Randall Stewart
> Cc:	john.loughney@nokia.com; discussion@sctp.de; Hollaus Thomas; tsvwg@ietf.org
> Subject:	Re: [Tsvwg] Possible DoS in SCTP ?
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hi all,
> please see my comments inline.
> 
> > > > Andreas:
> > > >
> > > > There is no amplification here though... so it is no
> > > > different than if Evil-E sent the aborts itself
> > > > to all the "victims" V... it must send N packets
> > > > to get N packets to victims... this is no
> > > > different than a UDP attack.. in fact a UDP attack
> > > > is easier...
> > >
> > > Well, *almost*. The only problem I see here is that many
> > > network installations by now have already adapted to UDP
> > > attacks, e.g. by filtering excessive amounts (or simply all)
> > > ICMP dest. unreachble packets (which are subsequently
> > > jamming the targets/victims network).
> > >
> > > With SCTP, however, it is not that simple: you get back
> > > a transport layer packet that is much harder to filter.
> >
> > How is SCTP harder to filter than UDP.. UDP IS a transport
> > protocol and the ports in the UDP/TCP header are in the
> > same place as the SCTP header...
> 
> Yes, but the UDP does not cause the DoS -- it is the ICMP
> destination unreachable packets sent in reply to a UDP packet, that
> cause the problems.
> 
> > > Think about it: there have been MANY distributed DoS attacks
> > > based on UDP, why shouldn't there be some based on SCTP some
> > > day ?
> >
> > Or TCP.. at this level though your attacker is limited
> > to the bandwidth/cpu of the Evil-E machine. This is NO
> > different than TCP or UDP or anything else out there
> > now...
> 
> No, that was my point. In typical DDoS scenarios, there is
> thousands of Evil-E's, directed by one central process. This has
> been done by e.g. the Stacheldraht based DDoS attacks.
> I am not saying this is much of an issue, I am saying that 
> SCTP can be used, as other protocols, for (D)DoS attacks.
> 
> Best regards,
> Andreas
> 
> 
> 
> > > BTW, John: I am not concerned about cycles for processing
> > > ABORTs. The problem really arises when an attacker starts a lot
> > > of "Evil Endpoints E", that all have one target: Victim V.
> > > The scarce resource is bandwidth.
> > >
> > > > Andreas Jungmaier wrote:
> > > > > -----BEGIN PGP SIGNED MESSAGE-----
> > > > > Hash: SHA1
> > > > >
> > > > > Dear all,
> > > > >
> > > > > I am sorry for not explaining more exactly what I meant.
> > > > > I am assuming an attack with IP-Spoofing. The evil attacker
> > > > > (say E) is spoofing poor victim (say V), using oblivious
> > > > > server (say) O.
> > > > >
> > > > > Now,
> > > > >
> > > > >  Evil-Endpoint E:
> > > > >   Source IP V(spoofed)
> > > > >   Dest IP O
> > > > >   SCTP-INIT (invalid params) ----------------> Server O
> > > > >                                    /-------------/
> > > > >                SCTP-Abort,Error   /
> > > > >   Poor Victim V   <--------------/
> > > > >
> > > > > Now, if E can use V, it could send such packets to a large number
> > > > > of servers. And then V gets bombed with Aborts.
> > > > >
> > > > > I agree fully, that an established association is safe, but that
> > > > > was not the issue I was thinking about. Comments ?
> > >
> > > - --
> > > -
> > > ------------------------------------------------------------------------
> > > Dipl.-Ing. Andreas Jungmaier              |                         _> 
> > > Computer Networking Technology Group      | ASCII ribbon campaign  ( )
> > > University of Essen                       |  - against HTML email   X
> > > http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
> > > -----BEGIN PGP SIGNATURE-----
> > > Version: GnuPG v1.0.6 (GNU/Linux)
> > > Comment: For info see http://www.gnupg.org
> > >
> > > iD8DBQE8a+eMXAsLBNOCSsARAmPfAJ47qG/Zy9+7omHnCOkbhsNR1b73QwCfZSQz
> > > vIKQ3YdsJjzX46WI8XX/Qbk=
> > > =g5mQ
> > > -----END PGP SIGNATURE-----
> 
> - -- 
> - ------------------------------------------------------------------------
> Dipl.-Ing. Andreas Jungmaier              |                         _
> Computer Networking Technology Group      | ASCII ribbon campaign  ( )
> University of Essen                       |  - against HTML email   X
> http://www.exp-math.uni-essen.de/~ajung   |              & vcards  / \
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.0.6 (GNU/Linux)
> Comment: For info see http://www.gnupg.org
> 
> iD8DBQE8a+3fXAsLBNOCSsARAm4GAJ9P8xD5loz/k4p2eSHeYefJj/3uEgCfbM15
> WsAQnnBB9g6F3Vsmyqqj0xI=
> =TK7O
> -----END PGP SIGNATURE-----
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 11:34:48 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA26567
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 11:34:48 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA09418
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 11:34:51 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA08302;
	Thu, 14 Feb 2002 11:22:06 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA08275
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 11:22:04 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA25981
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 11:22:01 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir02nok.ntc.nokia.com (esvir02nokt.ntc.nokia.com [172.21.143.34])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1EGMAZ29609
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 18:22:10 +0200 (EET)
Received: from esebh002.NOE.Nokia.com (unverified) by esvir02nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T591249e71eac158f22076@esvir02nok.ntc.nokia.com>;
 Thu, 14 Feb 2002 18:22:02 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh002.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Thu, 14 Feb 2002 18:22:02 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] Possible DoS in SCTP ?
Date: Thu, 14 Feb 2002 18:22:01 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF26308EEA5@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] Possible DoS in SCTP ?
Thread-Index: AcG1cvOC/tyOoFxIQN60HeDqA9zc7wAAArtw
To: <Michael.Tuexen@icn.siemens.de>, <ajung@exp-math.uni-essen.de>,
        <randall@stewart.chicago.il.us>
Cc: <john.loughney@nokia.com>, <discussion@sctp.de>,
        <thomas.hollaus@siemens.at>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 14 Feb 2002 16:22:02.0384 (UTC) FILETIME=[BC0BC900:01C1B573]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id LAA08276
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi all,

	This is TCP specific, but it is a very interesting lecture about DoS attacks... Check this out if you are interested:

	http://grc.com/dos/grcdos.htm

	I had to read it until the end when I started. Maybe because I didn't really know that much about DoS attacks... :-)

	Enjoy!

	BR Iván Arias Rodríguez

> -----Original Message-----
> From: ext Tuexen Michael [mailto:Michael.Tuexen@icn.siemens.de]
> Sent: 14. February 2002 18:10
> To: 'Andreas Jungmaier'; Randall Stewart
> Cc: Loughney John (NRC/Helsinki); discussion@sctp.de; Hollaus Thomas;
> tsvwg@ietf.org
> Subject: RE: [Tsvwg] Possible DoS in SCTP ?
> 
> 
> Dear all,
> 
> if I understand DDoS attacks correcly, they use some mechanism
> to generate a lot of packets by sending one. Using the SCTP
> feature currently being the discussed you are generating one
> packet by sending one. You can send this directly.
> 
> Best regards
> Michael
> 
> Michael Tuexen    Tel.:   +49 89 722 47210
> Siemens AG        Fax:    +49 89 722 48212
> ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
> 
> > -----Original Message-----
> > From:	Andreas Jungmaier [SMTP:ajung@exp-math.uni-essen.de]
> > Sent:	Thursday, February 14, 2002 6:03 PM
> > To:	Randall Stewart
> > Cc:	john.loughney@nokia.com; discussion@sctp.de; Hollaus 
> Thomas; tsvwg@ietf.org
> > Subject:	Re: [Tsvwg] Possible DoS in SCTP ?
> > 
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> > 
> > Hi all,
> > please see my comments inline.
> > 
> > > > > Andreas:
> > > > >
> > > > > There is no amplification here though... so it is no
> > > > > different than if Evil-E sent the aborts itself
> > > > > to all the "victims" V... it must send N packets
> > > > > to get N packets to victims... this is no
> > > > > different than a UDP attack.. in fact a UDP attack
> > > > > is easier...
> > > >
> > > > Well, *almost*. The only problem I see here is that many
> > > > network installations by now have already adapted to UDP
> > > > attacks, e.g. by filtering excessive amounts (or simply all)
> > > > ICMP dest. unreachble packets (which are subsequently
> > > > jamming the targets/victims network).
> > > >
> > > > With SCTP, however, it is not that simple: you get back
> > > > a transport layer packet that is much harder to filter.
> > >
> > > How is SCTP harder to filter than UDP.. UDP IS a transport
> > > protocol and the ports in the UDP/TCP header are in the
> > > same place as the SCTP header...
> > 
> > Yes, but the UDP does not cause the DoS -- it is the ICMP
> > destination unreachable packets sent in reply to a UDP packet, that
> > cause the problems.
> > 
> > > > Think about it: there have been MANY distributed DoS attacks
> > > > based on UDP, why shouldn't there be some based on SCTP some
> > > > day ?
> > >
> > > Or TCP.. at this level though your attacker is limited
> > > to the bandwidth/cpu of the Evil-E machine. This is NO
> > > different than TCP or UDP or anything else out there
> > > now...
> > 
> > No, that was my point. In typical DDoS scenarios, there is
> > thousands of Evil-E's, directed by one central process. This has
> > been done by e.g. the Stacheldraht based DDoS attacks.
> > I am not saying this is much of an issue, I am saying that 
> > SCTP can be used, as other protocols, for (D)DoS attacks.
> > 
> > Best regards,
> > Andreas
> > 
> > 
> > 
> > > > BTW, John: I am not concerned about cycles for processing
> > > > ABORTs. The problem really arises when an attacker starts a lot
> > > > of "Evil Endpoints E", that all have one target: Victim V.
> > > > The scarce resource is bandwidth.
> > > >
> > > > > Andreas Jungmaier wrote:
> > > > > > -----BEGIN PGP SIGNED MESSAGE-----
> > > > > > Hash: SHA1
> > > > > >
> > > > > > Dear all,
> > > > > >
> > > > > > I am sorry for not explaining more exactly what I meant.
> > > > > > I am assuming an attack with IP-Spoofing. The evil attacker
> > > > > > (say E) is spoofing poor victim (say V), using oblivious
> > > > > > server (say) O.
> > > > > >
> > > > > > Now,
> > > > > >
> > > > > >  Evil-Endpoint E:
> > > > > >   Source IP V(spoofed)
> > > > > >   Dest IP O
> > > > > >   SCTP-INIT (invalid params) ----------------> Server O
> > > > > >                                    /-------------/
> > > > > >                SCTP-Abort,Error   /
> > > > > >   Poor Victim V   <--------------/
> > > > > >
> > > > > > Now, if E can use V, it could send such packets to 
> a large number
> > > > > > of servers. And then V gets bombed with Aborts.
> > > > > >
> > > > > > I agree fully, that an established association is 
> safe, but that
> > > > > > was not the issue I was thinking about. Comments ?
> > > >
> > > > - --
> > > > -
> > > > 
> --------------------------------------------------------------
> ----------
> > > > Dipl.-Ing. Andreas Jungmaier              |             
>             _> 
> > > > Computer Networking Technology Group      | ASCII 
> ribbon campaign  ( )
> > > > University of Essen                       |  - against 
> HTML email   X
> > > > http://www.exp-math.uni-essen.de/~ajung   |             
>  & vcards  / \
> > > > -----BEGIN PGP SIGNATURE-----
> > > > Version: GnuPG v1.0.6 (GNU/Linux)
> > > > Comment: For info see http://www.gnupg.org
> > > >
> > > > iD8DBQE8a+eMXAsLBNOCSsARAmPfAJ47qG/Zy9+7omHnCOkbhsNR1b73QwCfZSQz
> > > > vIKQ3YdsJjzX46WI8XX/Qbk=
> > > > =g5mQ
> > > > -----END PGP SIGNATURE-----
> > 
> > - -- 
> > - 
> --------------------------------------------------------------
> ----------
> > Dipl.-Ing. Andreas Jungmaier              |                 
>         _
> > Computer Networking Technology Group      | ASCII ribbon 
> campaign  ( )
> > University of Essen                       |  - against HTML 
> email   X
> > http://www.exp-math.uni-essen.de/~ajung   |              & 
> vcards  / \
> > -----BEGIN PGP SIGNATURE-----
> > Version: GnuPG v1.0.6 (GNU/Linux)
> > Comment: For info see http://www.gnupg.org
> > 
> > iD8DBQE8a+3fXAsLBNOCSsARAm4GAJ9P8xD5loz/k4p2eSHeYefJj/3uEgCfbM15
> > WsAQnnBB9g6F3Vsmyqqj0xI=
> > =TK7O
> > -----END PGP SIGNATURE-----
> > 
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 11:37:16 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA26664
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 11:37:16 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA09530
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 11:37:18 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA06929;
	Thu, 14 Feb 2002 11:09:04 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA06900
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 11:09:02 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA25558
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 11:08:46 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir03nok.nokia.com (esvir03nokt.ntc.nokia.com [172.21.143.35])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1EG8uZ23579
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 18:08:56 +0200 (EET)
Received: from esebh004.NOE.Nokia.com (unverified) by esvir03nok.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T59123dc907ac158f23076@esvir03nok.nokia.com>;
 Thu, 14 Feb 2002 18:08:48 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh004.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Thu, 14 Feb 2002 18:08:48 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Date: Thu, 14 Feb 2002 18:08:47 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF263257A84@esebe017.NOE.Nokia.com>
Thread-Topic: [Sigtran] Issues with SCTP TSN and Cum TSN
Thread-Index: AcG1b4DVh0LpKfD8Tv6W6F/Yi0B2xwAAEqZw
To: <c-Jacques.Rajchgod@wcom.com>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 14 Feb 2002 16:08:48.0385 (UTC) FILETIME=[E2C92710:01C1B571]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id LAA06901
Subject: [Tsvwg] RE: [Sigtran] Issues with SCTP TSN and Cum TSN
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi again,

	First of all, I'll use tsvwg and not sigtran as SCTP is discussed in the tsvwg since quite long ago...

> -----Original Message-----
> From: ext Jacques Rajchgod [mailto:c-Jacques.Rajchgod@wcom.com]
> Sent: 14. February 2002 17:49
> To: sigtran@ietf.org
> Subject: RE: [Sigtran] Issues with SCTP TSN and Cum TSN
> 
> 
> 
> Per Brian's past mail on the matter (and Sec 6.2 of RFC 2960, 
> 3rd parag),
> SACKs can be simply sent to adjust a_rwnd values dynamically. 
> Thus, a SACK
> can be used for that control purpose, in addition to the 
> purpose of ACK'ing
> received data. When there is no data to ACK, SACKs would be 
> used to adjust
> the a_rwnd value dynamically.
> The time when that situation is likely to occur is - 
> immediately at the
> conclusion of the INIT sequence, just before data might have 
> begun to flow.

	Why? You have just stated which is your a_rwnd inside the INIT/INIT ACK and few time later you changed it?... In any case this does not really matter... Let's continue...

> At such time:
> 
> 1) The INIT sender "a" has just received an INIT_ACK from "b".

	Yes, in the normal case it received just one INIT ACK and the COOKIE ACK...

> 2) "a" has studied the returned values in the INIT_ACK from 
> "b" and may
> decide to re-adjust its ("a") RX buffer size to adapt to b's 
> just-advertised
> values.

	:-/ Why? What has to do the receiving window of one peer with the other's one? I don't think you should adjust your rwnd once you know your peer's... why is the reason to do this? What am I missing?

> 3) "a" would send a SACK smack immediately to "b" to notify 
> "b" of its ("a")
> new a_rwnd size.

	Well... yes, if for some reason you have changed it, then it is nice to send a SACK to indicate it... But not compulsorily...

> That is why one may frequently see a SACK at the beginning 
> smack after an
> INIT, even though no M3UA PAYLOAD has yet been exchanged.

	I don't know that much about M3UA or how it can affect to SCTP's rwnd... But in any case, I wouldn't say that it would "frequently" happen. I would rather say that it would be a very rare event... Please, tell me why you would modify rwnd right after knowing the rwnd of your peer...

> In the scenario I was investigating, the MGC ("s") had INIT'ed with
> a_rwnd=32,768K. Then after having received the INIT_ACK from 
> the SG ("b")
> and having studied "b's" advertised capabilities, "a" decided 
> to re-adjust
> its a-rwnd and re-advertise it to "b". So it chose to do such 

	I didn't know this happened... ok then. Somewhere in RFC I think it says that should not shrink your receiver window, but let's say then that it would be a normal situation.

> by bundling a
> SACK with an ASP_UP Request (data payload) - a very efficent 
> traffic usage
> (spares bandwidth). The SACK was now advertising 8K for 
> a_rwnd on "a". Note
> that this is a dynamic way of re-adjusting window sizes.

	Yes... 

> The only issue I have with using SACKs as a_rwnd adjusters, 
> is the CUM_TSN
> param present in them. It would have made great sense to 
> dictate that when a
> SACK is used strictly as an a_rwnd adjuster, that no CUM_TSN 
> be present in
> it as it serves no purpose at that moment. Some manufacturers 
> hick-up on
> that:

	But, again, could you tell me what the Cum TSN ACK is doing wrong? I mean, what is the bad thing about receiving it?

> They see a CUM_TSN and start checking to see what data it is that the
> CUM_TSN is ACK'ing at the time. Since there has been no data 
> sent yet, they
> get lost on that computation...Therefore, to prevent such 

	A very bad implementation I would say... A "normal" one should have a variable such as "next TSN to send"... When the SACK is processed, you check if the received Cum TSN ACK is bigger, smaller or equal:

	- If it is smaller, you simply drop the SACK (it is an old one).
	- If it is equal, you check if there is any Gap ACK Block, and then you process them.
	- If it is bigger, the peer is acknowledging TSNs that you should erase from your memory and some other things.

	Also you use the a_rwnd to update your rwnd value...

> UNWARRANTED (<---
> note the CAPS) causes for hangs, it would have meant good thinking to
> dictate that when a SACK is used strictly as an a_rwnd 
> adjuster, that NO
> CUM_TSN be present at that specific time for that specific use.

	One of the two of us is not getting the point. You never tear down an association because a SACK does not acknowledge anything new... I don't see if there is any other problem... Excuse me but I fail to see the problem... The implementation behaviour you have described is wrong... That shouldn't happen...

	Thanks!

	BR Iván Arias Rodríguez

> Tks.
> 
> Jacques Rajchgod
> MCI Worldcom
> Next Generation System Evaluation
> Tel: [972] 729-7180
> E-mail: c-Jacques.Rajchgod@wcom.com
> 
> 
> 
> -----Original Message-----
> From: Ivan.Arias-Rodriguez@nokia.com
> [mailto:Ivan.Arias-Rodriguez@nokia.com]
> Sent: Thursday, February 14, 2002 3:49 AM
> To: c-Jacques.Rajchgod@wcom.com; Michael.Tuexen@lurchi.franken.de
> Cc: sigtran@ietf.org
> Subject: RE: [Sigtran] Issues with SCTP TSN and Cum TSN
> 
> 
> 	Hi again Jacques,
> 
> > CUM_TSN's are used to ACK DATA that has been exchanged.
> 
> 	Well... SACK are used to ack data exchanged... If there 
> has not been data
> exchanged, why then sending any SACK at all? It is like if 
> you don't have
> any data to send to your peer, so, why would you like to send 
> a DATA chunk
> then?
> 
> 	I still continue not getting the point...
> 
> > At Y-1, there hasn't been any data exchanged by anyone yet -
> > neither the MGC
> > nor the SG.
> >
> > INIT, INIT_ACK's, COOKIEs, COOKIE_ACKs - are NOT DATA...These
> > are CONTROL
> > messages. Only DATA payloads may use SACK's - not Control messages.
> 
> 	So, don't use SACKs... :-)
> 
> > That's why the "Y-1" is wrong at that point.
> 
> 	Michael is right... But I guess that it was just a 
> simply mistake when
> writing your example (as I did when reading it, I saw the X 
> and Y-1 and I
> thought they were right)... As he said
> 
> > MGC->SG: INIT(Initial_TSN=X)
> > SG->MGC: INIT-ACK(Initial_TSN=Y)
> > MGC->SG: SACK(Cumualative_TSN_Ack=Y-1); DATA(TSN=X)
> 
> is the right example.
> 
> > Personally, though, I would have preferred to have seen any
> > of the three
> > following options being implemented at the IETF for subject 
> scenario:
> >
> > 1) That the (Y-1) is somehow SCTP-rejected since its
> > existence indicates
> > that its sender wished to ACK (non-existing) data - an
> > illogical action...
> 
> 	Well... The SACK is not acknowledging anything, but it 
> still has some other
> valuable information such as the a_rwnd. But as I have said, 
> if you need a
> SACK at all, you are not pushed to send it at all... It is an 
> option, not
> any mandatory thing...
> 
> > 2) That a SACK not ACKING DATA - do not use a CUM_TSN at
> > all...since the
> > CUM_TSN is not designed for non-data usage...
> 
> 	:-S
> 
> 	I don't know what you want... The SACK has its 
> structure and it should be
> like that... If you don't need the SACK, don't send it...
> 
> > 3) That a SACK not ACKING DATA - be used with a
> > CUM_TSN=All-Zero, so to be
> > understood as "Not_Used".
> 
> 	What is the point of this? The field is there, so... If 
> you tell me that
> your first DATA chunk will carry the TSN X, then it is right 
> to include a
> Cumulative TSN ACK set to X-1 inside a SACK that (for some 
> reason) you send
> to the "not yet data sender". If you do it so that you 
> haven't send anything
> and the Cum TSN ACK is set to any other value than X-1, then 
> there is a bug
> at the SACK sender...
> 
> 	In any case, could you explain me what is the problem 
> about having a
> (right) Cum TSN ACK even if you don't use it?
> 
> 	Thanks!
> 
> 	BR Iván Arias Rodríguez
> 
> 
> 
> _______________________________________________
> Sigtran mailing list
> Sigtran@ietf.org
> https://www1.ietf.org/mailman/listinfo/sigtran
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 11:56:44 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA27520
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 11:56:44 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA10822
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 11:56:47 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA09387;
	Thu, 14 Feb 2002 11:34:42 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA09334
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 11:34:39 -0500 (EST)
Received: from gorilla.mchh.siemens.de (gorilla.mchh.siemens.de [194.138.158.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA26561;
	Thu, 14 Feb 2002 11:34:35 -0500 (EST)
Received: from blues.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.227])
	by gorilla.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id RAA12724;
	Thu, 14 Feb 2002 17:34:35 +0100 (MET)
Received: from mchh273e.demchh201e.icn.siemens.de ([139.21.200.83])
	by blues.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id RAA18320;
	Thu, 14 Feb 2002 17:34:29 +0100 (MET)
Received: by MCHH273E with Internet Mail Service (5.5.2653.19)
	id <1JWX0C8Y>; Thu, 14 Feb 2002 17:34:35 +0100
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF75D@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'Jacques Rajchgod'" <c-Jacques.Rajchgod@wcom.com>, sigtran@ietf.org,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Date: Thu, 14 Feb 2002 17:34:34 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Subject: [Tsvwg] RE: [Sigtran] Issues with SCTP TSN and Cum TSN
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Dear Jacques,

the Cumulative_TSN_Ack is part of any SACK. I do not see
why this isn't good. And you must have a rule what to use
as an inital value because you must be able to handle the
case where the first packtet containing DATA chunks is
lost and the second arrives. You have to acknowledge them but
you must use the initial value as the cumulative_TSN_Ack.

Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Jacques Rajchgod [SMTP:c-Jacques.Rajchgod@wcom.com]
> Sent:	Thursday, February 14, 2002 4:49 PM
> To:	sigtran@ietf.org
> Subject:	RE: [Sigtran] Issues with SCTP TSN and Cum TSN
> 
> 
> Per Brian's past mail on the matter (and Sec 6.2 of RFC 2960, 3rd parag),
> SACKs can be simply sent to adjust a_rwnd values dynamically. Thus, a SACK
> can be used for that control purpose, in addition to the purpose of ACK'ing
> received data. When there is no data to ACK, SACKs would be used to adjust
> the a_rwnd value dynamically.
> The time when that situation is likely to occur is - immediately at the
> conclusion of the INIT sequence, just before data might have begun to flow.
> At such time:
> 
> 1) The INIT sender "a" has just received an INIT_ACK from "b".
> 2) "a" has studied the returned values in the INIT_ACK from "b" and may
> decide to re-adjust its ("a") RX buffer size to adapt to b's just-advertised
> values.
> 3) "a" would send a SACK smack immediately to "b" to notify "b" of its ("a")
> new a_rwnd size.
> 
> That is why one may frequently see a SACK at the beginning smack after an
> INIT, even though no M3UA PAYLOAD has yet been exchanged.
> 
> In the scenario I was investigating, the MGC ("s") had INIT'ed with
> a_rwnd=32,768K. Then after having received the INIT_ACK from the SG ("b")
> and having studied "b's" advertised capabilities, "a" decided to re-adjust
> its a-rwnd and re-advertise it to "b". So it chose to do such by bundling a
> SACK with an ASP_UP Request (data payload) - a very efficent traffic usage
> (spares bandwidth). The SACK was now advertising 8K for a_rwnd on "a". Note
> that this is a dynamic way of re-adjusting window sizes.
> 
> The only issue I have with using SACKs as a_rwnd adjusters, is the CUM_TSN
> param present in them. It would have made great sense to dictate that when a
> SACK is used strictly as an a_rwnd adjuster, that no CUM_TSN be present in
> it as it serves no purpose at that moment. Some manufacturers hick-up on
> that:
> They see a CUM_TSN and start checking to see what data it is that the
> CUM_TSN is ACK'ing at the time. Since there has been no data sent yet, they
> get lost on that computation...Therefore, to prevent such UNWARRANTED (<---
> note the CAPS) causes for hangs, it would have meant good thinking to
> dictate that when a SACK is used strictly as an a_rwnd adjuster, that NO
> CUM_TSN be present at that specific time for that specific use.
> 
> Tks.
> 
> Jacques Rajchgod
> MCI Worldcom
> Next Generation System Evaluation
> Tel: [972] 729-7180
> E-mail: c-Jacques.Rajchgod@wcom.com
> 
> 
> 
> -----Original Message-----
> From: Ivan.Arias-Rodriguez@nokia.com
> [mailto:Ivan.Arias-Rodriguez@nokia.com]
> Sent: Thursday, February 14, 2002 3:49 AM
> To: c-Jacques.Rajchgod@wcom.com; Michael.Tuexen@lurchi.franken.de
> Cc: sigtran@ietf.org
> Subject: RE: [Sigtran] Issues with SCTP TSN and Cum TSN
> 
> 
> 	Hi again Jacques,
> 
> > CUM_TSN's are used to ACK DATA that has been exchanged.
> 
> 	Well... SACK are used to ack data exchanged... If there has not been data
> exchanged, why then sending any SACK at all? It is like if you don't have
> any data to send to your peer, so, why would you like to send a DATA chunk
> then?
> 
> 	I still continue not getting the point...
> 
> > At Y-1, there hasn't been any data exchanged by anyone yet -> 
> > neither the MGC
> > nor the SG.
> >
> > INIT, INIT_ACK's, COOKIEs, COOKIE_ACKs - are NOT DATA...These
> > are CONTROL
> > messages. Only DATA payloads may use SACK's - not Control messages.
> 
> 	So, don't use SACKs... :-)
> 
> > That's why the "Y-1" is wrong at that point.
> 
> 	Michael is right... But I guess that it was just a simply mistake when
> writing your example (as I did when reading it, I saw the X and Y-1 and I
> thought they were right)... As he said
> 
> > MGC->SG: INIT(Initial_TSN=X)
> > SG->MGC: INIT-ACK(Initial_TSN=Y)
> > MGC->SG: SACK(Cumualative_TSN_Ack=Y-1); DATA(TSN=X)
> 
> is the right example.
> 
> > Personally, though, I would have preferred to have seen any
> > of the three
> > following options being implemented at the IETF for subject scenario:
> >
> > 1) That the (Y-1) is somehow SCTP-rejected since its
> > existence indicates
> > that its sender wished to ACK (non-existing) data - an
> > illogical action...
> 
> 	Well... The SACK is not acknowledging anything, but it still has some other
> valuable information such as the a_rwnd. But as I have said, if you need a
> SACK at all, you are not pushed to send it at all... It is an option, not
> any mandatory thing...
> 
> > 2) That a SACK not ACKING DATA - do not use a CUM_TSN at
> > all...since the
> > CUM_TSN is not designed for non-data usage...
> 
> 	:-S
> 
> 	I don't know what you want... The SACK has its structure and it should be
> like that... If you don't need the SACK, don't send it...
> 
> > 3) That a SACK not ACKING DATA - be used with a
> > CUM_TSN=All-Zero, so to be
> > understood as "Not_Used".
> 
> 	What is the point of this? The field is there, so... If you tell me that
> your first DATA chunk will carry the TSN X, then it is right to include a
> Cumulative TSN ACK set to X-1 inside a SACK that (for some reason) you send
> to the "not yet data sender". If you do it so that you haven't send anything
> and the Cum TSN ACK is set to any other value than X-1, then there is a bug
> at the SACK sender...
> 
> 	In any case, could you explain me what is the problem about having a
> (right) Cum TSN ACK even if you don't use it?
> 
> 	Thanks!
> 
> 	BR Iván Arias Rodríguez
> 
> 
> 
> _______________________________________________
> Sigtran mailing list
> Sigtran@ietf.org
> https://www1.ietf.org/mailman/listinfo/sigtran

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 13:25:06 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA05349
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 13:25:06 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA16036
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 13:25:07 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA15613;
	Thu, 14 Feb 2002 13:13:58 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA15562
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 13:13:55 -0500 (EST)
Received: from dgesmtp01.wcom.com (dgesmtp01.wcom.com [199.249.16.16])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA05034;
	Thu, 14 Feb 2002 13:13:53 -0500 (EST)
Received: from CONVERSION-DAEMON by firewall.wcom.com (PMDF V5.2-33 #42260)
 id <0GRJ00G01BYB0A@firewall.wcom.com>; Thu, 14 Feb 2002 18:13:23 +0000 (GMT)
Received: from dgismtp06.wcomnet.com ([166.38.58.89])
 by firewall.wcom.com (PMDF V5.2-33 #42260)
 with ESMTP id <0GRJ00EMMBYBM0@firewall.wcom.com>; Thu,
 14 Feb 2002 18:13:23 +0000 (GMT)
Received: from dgismtp06.wcomnet.com by dgismtp06.wcomnet.com
 (iPlanet Messaging Server 5.1 (built May  7 2001))
 with SMTP id <0GRJ00601BYA41@dgismtp06.wcomnet.com>; Thu,
 14 Feb 2002 18:13:22 +0000 (GMT)
Received: from ws107v7180a ([166.34.96.254])
 by dgismtp06.wcomnet.com (iPlanet Messaging Server 5.1 (built May  7 2001))
 with SMTP id <0GRJ0053RBY8NP@dgismtp06.wcomnet.com>; Thu,
 14 Feb 2002 18:13:22 +0000 (GMT)
Date: Thu, 14 Feb 2002 12:13:13 -0600
From: Jacques Rajchgod <c-Jacques.Rajchgod@wcom.com>
In-reply-to: <F92978223B01D311ACFF0008C71EE19D015DF75D@MCHH225E>
To: Tuexen Michael <Michael.Tuexen@icn.siemens.de>, sigtran@ietf.org,
        tsvwg@ietf.org
Message-id: <NEBBLDAOEDIKLJPELCGECEHOCIAA.c-Jacques.Rajchgod@wcom.com>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Content-type: text/plain; charset=iso-8859-1
Importance: Normal
X-Priority: 3 (Normal)
X-MSMail-priority: Normal
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from QUOTED-PRINTABLE to 8bit by optimus.ietf.org id NAA15565
Subject: [Tsvwg] RE: [Sigtran] Issues with SCTP TSN and Cum TSN
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

OK. Here it is:

1) a SENDS to b:	 -------------- INIT ---------------->
2) a RECEIVES from b	<------------ INI_ACK ------------
3) a SENDS to b	-------------- COOKIE ---------->
4) a RECEIVES from b	<------------ COOKIE_ACK ------------
3) a SENDS to b	--------------SCTP pack:  SACK (CUM_TSN=x) bundled with
ASP_UP ---------->

- What's the "CUM_TSN=x" pointing to? Which DATA chunk is it ACK'ing
CUMULATIVELY??
- There hasn't been any DATA exchanged anywhere yet, to ack!
- A CUM_TSN is only used to ack CUMULATIVELY - DATA chunks. There are no
DATA chunks exchanged at INIT time. The DATA may come only after the INIT.
- The INIT consists only of CONTROL chunks.
- So, what is a Receiver of that CUM_TSN=x ("b") supposed to look for? "b"
hasn't sent any data to be ACK'ed cumulatively yet!
- Take a look at Section 1.4, page 11, the 5th bullet down - CUM_TSN...

Brian recommended yesterday that the recipient of such SACK ignore that SACK
and just update a_rwnd. Here is the statement:

"The receiver of a SACK with Cummulative Ack of Y-1 should
discard the SACK and do nothing further with it other than
updating a_rwnd because it preceeds the Initial TSN.  This
is how an implementation acks nothing yet updates a_rwnd."

Jacques Rajchgod
MCI Worldcom
Next Generation System Evaluation
Tel: [972] 729-7180
E-mail: c-Jacques.Rajchgod@wcom.com



-----Original Message-----
From: Tuexen Michael [mailto:Michael.Tuexen@icn.siemens.de]
Sent: Thursday, February 14, 2002 10:35 AM
To: 'Jacques Rajchgod'; sigtran@ietf.org; 'tsvwg@ietf.org'
Subject: RE: [Sigtran] Issues with SCTP TSN and Cum TSN


Dear Jacques,

the Cumulative_TSN_Ack is part of any SACK. I do not see
why this isn't good. And you must have a rule what to use
as an inital value because you must be able to handle the
case where the first packtet containing DATA chunks is
lost and the second arrives. You have to acknowledge them but
you must use the initial value as the cumulative_TSN_Ack.

Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Jacques Rajchgod [SMTP:c-Jacques.Rajchgod@wcom.com]
> Sent:	Thursday, February 14, 2002 4:49 PM
> To:	sigtran@ietf.org
> Subject:	RE: [Sigtran] Issues with SCTP TSN and Cum TSN
>
>
> Per Brian's past mail on the matter (and Sec 6.2 of RFC 2960, 3rd parag),
> SACKs can be simply sent to adjust a_rwnd values dynamically. Thus, a SACK
> can be used for that control purpose, in addition to the purpose of
ACK'ing
> received data. When there is no data to ACK, SACKs would be used to adjust
> the a_rwnd value dynamically.
> The time when that situation is likely to occur is - immediately at the
> conclusion of the INIT sequence, just before data might have begun to
flow.
> At such time:
>
> 1) The INIT sender "a" has just received an INIT_ACK from "b".
> 2) "a" has studied the returned values in the INIT_ACK from "b" and may
> decide to re-adjust its ("a") RX buffer size to adapt to b's
just-advertised
> values.
> 3) "a" would send a SACK smack immediately to "b" to notify "b" of its
("a")
> new a_rwnd size.
>
> That is why one may frequently see a SACK at the beginning smack after an
> INIT, even though no M3UA PAYLOAD has yet been exchanged.
>
> In the scenario I was investigating, the MGC ("s") had INIT'ed with
> a_rwnd=32,768K. Then after having received the INIT_ACK from the SG ("b")
> and having studied "b's" advertised capabilities, "a" decided to re-adjust
> its a-rwnd and re-advertise it to "b". So it chose to do such by bundling
a
> SACK with an ASP_UP Request (data payload) - a very efficent traffic usage
> (spares bandwidth). The SACK was now advertising 8K for a_rwnd on "a".
Note
> that this is a dynamic way of re-adjusting window sizes.
>
> The only issue I have with using SACKs as a_rwnd adjusters, is the CUM_TSN
> param present in them. It would have made great sense to dictate that when
a
> SACK is used strictly as an a_rwnd adjuster, that no CUM_TSN be present in
> it as it serves no purpose at that moment. Some manufacturers hick-up on
> that:
> They see a CUM_TSN and start checking to see what data it is that the
> CUM_TSN is ACK'ing at the time. Since there has been no data sent yet,
they
> get lost on that computation...Therefore, to prevent such UNWARRANTED
(<---
> note the CAPS) causes for hangs, it would have meant good thinking to
> dictate that when a SACK is used strictly as an a_rwnd adjuster, that NO
> CUM_TSN be present at that specific time for that specific use.
>
> Tks.
>
> Jacques Rajchgod
> MCI Worldcom
> Next Generation System Evaluation
> Tel: [972] 729-7180
> E-mail: c-Jacques.Rajchgod@wcom.com
>
>
>
> -----Original Message-----
> From: Ivan.Arias-Rodriguez@nokia.com
> [mailto:Ivan.Arias-Rodriguez@nokia.com]
> Sent: Thursday, February 14, 2002 3:49 AM
> To: c-Jacques.Rajchgod@wcom.com; Michael.Tuexen@lurchi.franken.de
> Cc: sigtran@ietf.org
> Subject: RE: [Sigtran] Issues with SCTP TSN and Cum TSN
>
>
> 	Hi again Jacques,
>
> > CUM_TSN's are used to ACK DATA that has been exchanged.
>
> 	Well... SACK are used to ack data exchanged... If there has not been data
> exchanged, why then sending any SACK at all? It is like if you don't have
> any data to send to your peer, so, why would you like to send a DATA chunk
> then?
>
> 	I still continue not getting the point...
>
> > At Y-1, there hasn't been any data exchanged by anyone yet ->
> > neither the MGC
> > nor the SG.
> >
> > INIT, INIT_ACK's, COOKIEs, COOKIE_ACKs - are NOT DATA...These
> > are CONTROL
> > messages. Only DATA payloads may use SACK's - not Control messages.
>
> 	So, don't use SACKs... :-)
>
> > That's why the "Y-1" is wrong at that point.
>
> 	Michael is right... But I guess that it was just a simply mistake when
> writing your example (as I did when reading it, I saw the X and Y-1 and I
> thought they were right)... As he said
>
> > MGC->SG: INIT(Initial_TSN=X)
> > SG->MGC: INIT-ACK(Initial_TSN=Y)
> > MGC->SG: SACK(Cumualative_TSN_Ack=Y-1); DATA(TSN=X)
>
> is the right example.
>
> > Personally, though, I would have preferred to have seen any
> > of the three
> > following options being implemented at the IETF for subject scenario:
> >
> > 1) That the (Y-1) is somehow SCTP-rejected since its
> > existence indicates
> > that its sender wished to ACK (non-existing) data - an
> > illogical action...
>
> 	Well... The SACK is not acknowledging anything, but it still has some
other
> valuable information such as the a_rwnd. But as I have said, if you need a
> SACK at all, you are not pushed to send it at all... It is an option, not
> any mandatory thing...
>
> > 2) That a SACK not ACKING DATA - do not use a CUM_TSN at
> > all...since the
> > CUM_TSN is not designed for non-data usage...
>
> 	:-S
>
> 	I don't know what you want... The SACK has its structure and it should be
> like that... If you don't need the SACK, don't send it...
>
> > 3) That a SACK not ACKING DATA - be used with a
> > CUM_TSN=All-Zero, so to be
> > understood as "Not_Used".
>
> 	What is the point of this? The field is there, so... If you tell me that
> your first DATA chunk will carry the TSN X, then it is right to include a
> Cumulative TSN ACK set to X-1 inside a SACK that (for some reason) you
send
> to the "not yet data sender". If you do it so that you haven't send
anything
> and the Cum TSN ACK is set to any other value than X-1, then there is a
bug
> at the SACK sender...
>
> 	In any case, could you explain me what is the problem about having a
> (right) Cum TSN ACK even if you don't use it?
>
> 	Thanks!
>
> 	BR Iván Arias Rodríguez
>
>
>
> _______________________________________________
> Sigtran mailing list
> Sigtran@ietf.org
> https://www1.ietf.org/mailman/listinfo/sigtran



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 13:38:58 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA05650
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 13:38:58 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA17124
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 13:39:00 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA16188;
	Thu, 14 Feb 2002 13:28:11 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA16130
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 13:28:08 -0500 (EST)
Received: from ilsa.franken.de (postfix@ilsa.franken.de [193.175.24.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA05392;
	Thu, 14 Feb 2002 13:28:05 -0500 (EST)
Received: from localhost (p50801AD8.dip.t-dialin.net [80.128.26.216])
	by ilsa.franken.de (Postfix) with ESMTP
	id A6845245C5; Thu, 14 Feb 2002 19:27:50 +0100 (CET)
Date: Thu, 14 Feb 2002 19:25:56 +0100
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Mime-Version: 1.0 (Apple Message framework v480)
Cc: sigtran@ietf.org, tsvwg@ietf.org
To: Jacques Rajchgod <c-Jacques.Rajchgod@wcom.com>
From: Michael Tuexen <Michael.Tuexen@icn.siemens.de>
In-Reply-To: <NEBBLDAOEDIKLJPELCGECEHOCIAA.c-Jacques.Rajchgod@wcom.com>
Message-Id: <49A68A61-2178-11D6-82A5-0030654C1AB6@icn.siemens.de>
X-Mailer: Apple Mail (2.480)
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id NAA16133
Subject: [Tsvwg] Re: [Sigtran] Issues with SCTP TSN and Cum TSN
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Dear Jacques,
see my comments below.
Best regards
Michael
On Thursday, February 14, 2002, at 07:13 PM, Jacques Rajchgod wrote:

> OK. Here it is:
>
> 1) a SENDS to b:	 -------------- INIT ---------------->
> 2) a RECEIVES from b	<------------ INI_ACK ------------
> 3) a SENDS to b	-------------- COOKIE ---------->
> 4) a RECEIVES from b	<------------ COOKIE_ACK ------------
> 3) a SENDS to b	--------------SCTP pack:  SACK (CUM_TSN=x) bundled with
> ASP_UP ---------->
>
> - What's the "CUM_TSN=x" pointing to? Which DATA chunk is it ACK'ing
> CUMULATIVELY??
The value will be (Initial_TSN of the INIT_ACK chunk from b) - 1
> - There hasn't been any DATA exchanged anywhere yet, to ack!
It doesn't matter.
> - A CUM_TSN is only used to ack CUMULATIVELY - DATA chunks. There are no
> DATA chunks exchanged at INIT time. The DATA may come only after the 
> INIT.
It is just the rule. b knows it and understands that nothing is 
acknowledged.
- The INIT consists only of CONTROL chunks.
It is a control chunk.
> - So, what is a Receiver of that CUM_TSN=x ("b") supposed to look for? 
> "b"
> hasn't sent any data to be ACK'ed cumulatively yet!
B is looking for  (Initial_TSN of the INIT_ACK chunk from b) - 1, he 
knows the rule.
> - Take a look at Section 1.4, page 11, the 5th bullet down - CUM_TSN...
In general that is true. But in case of the initialization it is NOT 
true.
That field is in each SACK chunk, so you have to put a value in it, so 
you
have to define an initialization value.
If you do not have anything to acknowledge you do not have to send a 
SACK. But
if the first packets containing DATA chunks are lost and you have to 
acknowledge
the first one you have to tell the peer that the first packets are lost. 
So you
will also use the initialization value for the Cumulative_TSN_Ack.
>
> Brian recommended yesterday that the recipient of such SACK ignore that 
> SACK
> and just update a_rwnd. Here is the statement:
>
> "The receiver of a SACK with Cummulative Ack of Y-1 should
> discard the SACK and do nothing further with it other than
> updating a_rwnd because it preceeds the Initial TSN.  This
> is how an implementation acks nothing yet updates a_rwnd."
>
> Jacques Rajchgod
> MCI Worldcom
> Next Generation System Evaluation
> Tel: [972] 729-7180
> E-mail: c-Jacques.Rajchgod@wcom.com
>
>
>
> -----Original Message-----
> From: Tuexen Michael [mailto:Michael.Tuexen@icn.siemens.de]
> Sent: Thursday, February 14, 2002 10:35 AM
> To: 'Jacques Rajchgod'; sigtran@ietf.org; 'tsvwg@ietf.org'
> Subject: RE: [Sigtran] Issues with SCTP TSN and Cum TSN
>
>
> Dear Jacques,
>
> the Cumulative_TSN_Ack is part of any SACK. I do not see
> why this isn't good. And you must have a rule what to use
> as an inital value because you must be able to handle the
> case where the first packtet containing DATA chunks is
> lost and the second arrives. You have to acknowledge them but
> you must use the initial value as the cumulative_TSN_Ack.
>
> Best regards
> Michael
>
> Michael Tuexen    Tel.:   +49 89 722 47210
> Siemens AG        Fax:    +49 89 722 48212
> ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
>
>> -----Original Message-----
>> From:	Jacques Rajchgod [SMTP:c-Jacques.Rajchgod@wcom.com]
>> Sent:	Thursday, February 14, 2002 4:49 PM
>> To:	sigtran@ietf.org
>> Subject:	RE: [Sigtran] Issues with SCTP TSN and Cum TSN
>>
>>
>> Per Brian's past mail on the matter (and Sec 6.2 of RFC 2960, 3rd 
>> parag),
>> SACKs can be simply sent to adjust a_rwnd values dynamically. Thus, a 
>> SACK
>> can be used for that control purpose, in addition to the purpose of
> ACK'ing
>> received data. When there is no data to ACK, SACKs would be used to 
>> adjust
>> the a_rwnd value dynamically.
>> The time when that situation is likely to occur is - immediately at the
>> conclusion of the INIT sequence, just before data might have begun to
> flow.
>> At such time:
>>
>> 1) The INIT sender "a" has just received an INIT_ACK from "b".
>> 2) "a" has studied the returned values in the INIT_ACK from "b" and may
>> decide to re-adjust its ("a") RX buffer size to adapt to b's
> just-advertised
>> values.
>> 3) "a" would send a SACK smack immediately to "b" to notify "b" of its
> ("a")
>> new a_rwnd size.
>>
>> That is why one may frequently see a SACK at the beginning smack after 
>> an
>> INIT, even though no M3UA PAYLOAD has yet been exchanged.
>>
>> In the scenario I was investigating, the MGC ("s") had INIT'ed with
>> a_rwnd=32,768K. Then after having received the INIT_ACK from the SG 
>> ("b")
>> and having studied "b's" advertised capabilities, "a" decided to 
>> re-adjust
>> its a-rwnd and re-advertise it to "b". So it chose to do such by 
>> bundling
> a
>> SACK with an ASP_UP Request (data payload) - a very efficent traffic 
>> usage
>> (spares bandwidth). The SACK was now advertising 8K for a_rwnd on "a".
> Note
>> that this is a dynamic way of re-adjusting window sizes.
>>
>> The only issue I have with using SACKs as a_rwnd adjusters, is the 
>> CUM_TSN
>> param present in them. It would have made great sense to dictate that 
>> when
> a
>> SACK is used strictly as an a_rwnd adjuster, that no CUM_TSN be 
>> present in
>> it as it serves no purpose at that moment. Some manufacturers hick-up 
>> on
>> that:
>> They see a CUM_TSN and start checking to see what data it is that the
>> CUM_TSN is ACK'ing at the time. Since there has been no data sent yet,
> they
>> get lost on that computation...Therefore, to prevent such UNWARRANTED
> (<---
>> note the CAPS) causes for hangs, it would have meant good thinking to
>> dictate that when a SACK is used strictly as an a_rwnd adjuster, that 
>> NO
>> CUM_TSN be present at that specific time for that specific use.
>>
>> Tks.
>>
>> Jacques Rajchgod
>> MCI Worldcom
>> Next Generation System Evaluation
>> Tel: [972] 729-7180
>> E-mail: c-Jacques.Rajchgod@wcom.com
>>
>>
>>
>> -----Original Message-----
>> From: Ivan.Arias-Rodriguez@nokia.com
>> [mailto:Ivan.Arias-Rodriguez@nokia.com]
>> Sent: Thursday, February 14, 2002 3:49 AM
>> To: c-Jacques.Rajchgod@wcom.com; Michael.Tuexen@lurchi.franken.de
>> Cc: sigtran@ietf.org
>> Subject: RE: [Sigtran] Issues with SCTP TSN and Cum TSN
>>
>>
>> 	Hi again Jacques,
>>
>>> CUM_TSN's are used to ACK DATA that has been exchanged.
>>
>> 	Well... SACK are used to ack data exchanged... If there has not 
>> been data
>> exchanged, why then sending any SACK at all? It is like if you don't 
>> have
>> any data to send to your peer, so, why would you like to send a DATA 
>> chunk
>> then?
>>
>> 	I still continue not getting the point...
>>
>>> At Y-1, there hasn't been any data exchanged by anyone yet ->
>>> neither the MGC
>>> nor the SG.
>>>
>>> INIT, INIT_ACK's, COOKIEs, COOKIE_ACKs - are NOT DATA...These
>>> are CONTROL
>>> messages. Only DATA payloads may use SACK's - not Control messages.
>>
>> 	So, don't use SACKs... :-)
>>
>>> That's why the "Y-1" is wrong at that point.
>>
>> 	Michael is right... But I guess that it was just a simply mistake 
>> when
>> writing your example (as I did when reading it, I saw the X and Y-1 
>> and I
>> thought they were right)... As he said
>>
>>> MGC->SG: INIT(Initial_TSN=X)
>>> SG->MGC: INIT-ACK(Initial_TSN=Y)
>>> MGC->SG: SACK(Cumualative_TSN_Ack=Y-1); DATA(TSN=X)
>>
>> is the right example.
>>
>>> Personally, though, I would have preferred to have seen any
>>> of the three
>>> following options being implemented at the IETF for subject scenario:
>>>
>>> 1) That the (Y-1) is somehow SCTP-rejected since its
>>> existence indicates
>>> that its sender wished to ACK (non-existing) data - an
>>> illogical action...
>>
>> 	Well... The SACK is not acknowledging anything, but it still has some
> other
>> valuable information such as the a_rwnd. But as I have said, if you 
>> need a
>> SACK at all, you are not pushed to send it at all... It is an option, 
>> not
>> any mandatory thing...
>>
>>> 2) That a SACK not ACKING DATA - do not use a CUM_TSN at
>>> all...since the
>>> CUM_TSN is not designed for non-data usage...
>>
>> 	:-S
>>
>> 	I don't know what you want... The SACK has its structure and it 
>> should be
>> like that... If you don't need the SACK, don't send it...
>>
>>> 3) That a SACK not ACKING DATA - be used with a
>>> CUM_TSN=All-Zero, so to be
>>> understood as "Not_Used".
>>
>> 	What is the point of this? The field is there, so... If you tell 
>> me that
>> your first DATA chunk will carry the TSN X, then it is right to 
>> include a
>> Cumulative TSN ACK set to X-1 inside a SACK that (for some reason) you
> send
>> to the "not yet data sender". If you do it so that you haven't send
> anything
>> and the Cum TSN ACK is set to any other value than X-1, then there is a
> bug
>> at the SACK sender...
>>
>> 	In any case, could you explain me what is the problem about having a
>> (right) Cum TSN ACK even if you don't use it?
>>
>> 	Thanks!
>>
>> 	BR Iván Arias Rodríguez
>>
>>
>>
>> _______________________________________________
>> Sigtran mailing list
>> Sigtran@ietf.org
>> https://www1.ietf.org/mailman/listinfo/sigtran
>
>
>
> _______________________________________________
> Sigtran mailing list
> Sigtran@ietf.org
> https://www1.ietf.org/mailman/listinfo/sigtran
>
>
Michael.Tuexen@micmac.franken.de


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 15:02:35 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA08242
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 15:02:30 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA22953
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 15:02:33 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA21790;
	Thu, 14 Feb 2002 14:49:55 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA21699
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 14:49:51 -0500 (EST)
Received: from dgesmtp02.wcom.com (dgesmtp02.wcom.com [199.249.16.17])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA07916;
	Thu, 14 Feb 2002 14:49:47 -0500 (EST)
Received: from CONVERSION-DAEMON by firewall.wcom.com (PMDF V5.2-33 #42261)
 id <0GRJ00G01GCUZF@firewall.wcom.com>; Thu, 14 Feb 2002 19:48:30 +0000 (GMT)
Received: from dgismtp06.wcomnet.com ([166.38.58.89])
 by firewall.wcom.com (PMDF V5.2-33 #42261)
 with ESMTP id <0GRJ00EIYGCTR6@firewall.wcom.com>; Thu,
 14 Feb 2002 19:48:30 +0000 (GMT)
Received: from dgismtp06.wcomnet.com by dgismtp06.wcomnet.com
 (iPlanet Messaging Server 5.1 (built May  7 2001))
 with SMTP id <0GRJ00D01GBG1P@dgismtp06.wcomnet.com>; Thu,
 14 Feb 2002 19:48:29 +0000 (GMT)
Received: from ws107v7180a ([166.34.96.254])
 by dgismtp06.wcomnet.com (iPlanet Messaging Server 5.1 (built May  7 2001))
 with SMTP id <0GRJ00BKLGCGDR@dgismtp06.wcomnet.com>; Thu,
 14 Feb 2002 19:48:18 +0000 (GMT)
Date: Thu, 14 Feb 2002 13:48:10 -0600
From: Jacques Rajchgod <c-Jacques.Rajchgod@wcom.com>
In-reply-to: <49A68A61-2178-11D6-82A5-0030654C1AB6@icn.siemens.de>
To: Michael Tuexen <Michael.Tuexen@icn.siemens.de>
Cc: sigtran@ietf.org, tsvwg@ietf.org
Message-id: <NEBBLDAOEDIKLJPELCGEEEHPCIAA.c-Jacques.Rajchgod@wcom.com>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Content-type: text/plain; charset=iso-8859-1
Content-transfer-encoding: 7bit
Importance: Normal
X-Priority: 3 (Normal)
X-MSMail-priority: Normal
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] RE: [Sigtran] Issues with SCTP TSN and Cum TSN
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Thanks, Michael...I'll run with that.

Still, some questions:

Your first mail on the topic contained the following comment:

"...If an SCTP endpoint wants to send a SACK chunk (from JR: i.e, after an
INIT) it must fill in the
Cumulative_TSN_Ack field. That value is initialized to the Initial_TSN of
the peer minus
one..."

Where does the "minus-1" come from?
Section "3.3.4" (SACK) 3rd parag states "Less-than or equal..."

Please kindky advise...


Jacques Rajchgod
MCI Worldcom
Next Generation System Evaluation
Tel: [972] 729-7180
E-mail: c-Jacques.Rajchgod@wcom.com



-----Original Message-----
From: sigtran-admin@ietf.org [mailto:sigtran-admin@ietf.org]On Behalf Of
Michael Tuexen
Sent: Thursday, February 14, 2002 12:26 PM
To: Jacques Rajchgod
Cc: sigtran@ietf.org; tsvwg@ietf.org
Subject: Re: [Sigtran] Issues with SCTP TSN and Cum TSN


Dear Jacques,
see my comments below.
Best regards
Michael
On Thursday, February 14, 2002, at 07:13 PM, Jacques Rajchgod wrote:

> OK. Here it is:
>
> 1) a SENDS to b:	 -------------- INIT ---------------->
> 2) a RECEIVES from b	<------------ INI_ACK ------------
> 3) a SENDS to b	-------------- COOKIE ---------->
> 4) a RECEIVES from b	<------------ COOKIE_ACK ------------
> 3) a SENDS to b	--------------SCTP pack:  SACK (CUM_TSN=x) bundled with
> ASP_UP ---------->
>
> - What's the "CUM_TSN=x" pointing to? Which DATA chunk is it ACK'ing
> CUMULATIVELY??
The value will be (Initial_TSN of the INIT_ACK chunk from b) - 1
> - There hasn't been any DATA exchanged anywhere yet, to ack!
It doesn't matter.
> - A CUM_TSN is only used to ack CUMULATIVELY - DATA chunks. There are no
> DATA chunks exchanged at INIT time. The DATA may come only after the
> INIT.
It is just the rule. b knows it and understands that nothing is
acknowledged.
- The INIT consists only of CONTROL chunks.
It is a control chunk.
> - So, what is a Receiver of that CUM_TSN=x ("b") supposed to look for?
> "b"
> hasn't sent any data to be ACK'ed cumulatively yet!
B is looking for  (Initial_TSN of the INIT_ACK chunk from b) - 1, he
knows the rule.
> - Take a look at Section 1.4, page 11, the 5th bullet down - CUM_TSN...
In general that is true. But in case of the initialization it is NOT
true.
That field is in each SACK chunk, so you have to put a value in it, so
you
have to define an initialization value.
If you do not have anything to acknowledge you do not have to send a
SACK. But
if the first packets containing DATA chunks are lost and you have to
acknowledge
the first one you have to tell the peer that the first packets are lost.
So you
will also use the initialization value for the Cumulative_TSN_Ack.
>
> Brian recommended yesterday that the recipient of such SACK ignore that
> SACK
> and just update a_rwnd. Here is the statement:
>
> "The receiver of a SACK with Cummulative Ack of Y-1 should
> discard the SACK and do nothing further with it other than
> updating a_rwnd because it preceeds the Initial TSN.  This
> is how an implementation acks nothing yet updates a_rwnd."
>
> Jacques Rajchgod
> MCI Worldcom
> Next Generation System Evaluation
> Tel: [972] 729-7180
> E-mail: c-Jacques.Rajchgod@wcom.com
>
>
>
> -----Original Message-----
> From: Tuexen Michael [mailto:Michael.Tuexen@icn.siemens.de]
> Sent: Thursday, February 14, 2002 10:35 AM
> To: 'Jacques Rajchgod'; sigtran@ietf.org; 'tsvwg@ietf.org'
> Subject: RE: [Sigtran] Issues with SCTP TSN and Cum TSN
>
>
> Dear Jacques,
>
> the Cumulative_TSN_Ack is part of any SACK. I do not see
> why this isn't good. And you must have a rule what to use
> as an inital value because you must be able to handle the
> case where the first packtet containing DATA chunks is
> lost and the second arrives. You have to acknowledge them but
> you must use the initial value as the cumulative_TSN_Ack.
>
> Best regards
> Michael
>
> Michael Tuexen    Tel.:   +49 89 722 47210
> Siemens AG        Fax:    +49 89 722 48212
> ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
>


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 15:27:55 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA08904
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 15:27:55 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA24225
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 15:27:58 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA23593;
	Thu, 14 Feb 2002 15:16:04 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA23541
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 15:16:01 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-217.dallas.net [209.44.42.217])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA08634;
	Thu, 14 Feb 2002 15:15:54 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id OAA31872;
	Thu, 14 Feb 2002 14:15:45 -0600
Date: Thu, 14 Feb 2002 14:15:45 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Jacques Rajchgod <c-Jacques.Rajchgod@wcom.com>
Cc: Michael Tuexen <Michael.Tuexen@icn.siemens.de>, sigtran@ietf.org,
        tsvwg@ietf.org
Message-ID: <20020214141545.C30356@openss7.org>
Reply-To: bidulock@openss7.org
Mail-Followup-To: Jacques Rajchgod <c-Jacques.Rajchgod@wcom.com>,
	Michael Tuexen <Michael.Tuexen@icn.siemens.de>, sigtran@ietf.org,
	tsvwg@ietf.org
References: <49A68A61-2178-11D6-82A5-0030654C1AB6@icn.siemens.de> <NEBBLDAOEDIKLJPELCGEEEHPCIAA.c-Jacques.Rajchgod@wcom.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2i
In-Reply-To: <NEBBLDAOEDIKLJPELCGEEEHPCIAA.c-Jacques.Rajchgod@wcom.com>; from c-Jacques.Rajchgod@wcom.com on Thu, Feb 14, 2002 at 01:48:10PM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Subject: [Tsvwg] Re: [Sigtran] Issues with SCTP TSN and Cum TSN
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Jacques,

Section 3.3.4:

   The SACK MUST contain the Cumulative TSN Ack and Advertised Receiver
   Window Credit (a_rwnd) parameters.

   By definition, the value of the Cumulative TSN Ack parameter is the
   last TSN received before a break in the sequence of received TSNs
   occurs; the next TSN value following this one has not yet been
   received at the endpoint sending the SACK.  This parameter therefore
   acknowledges receipt of all TSNs less than or equal to its value.

That's where the minus-1 comes from.  The value of the Cummulative
TSN in the SACK is the next expected TSN minus one.

--brian

Jacques Rajchgod wrote:                                   (Thu, 14 Feb 2002 13:48:10)
> Thanks, Michael...I'll run with that.
> 
> Still, some questions:
> 
> Your first mail on the topic contained the following comment:
> 
> "...If an SCTP endpoint wants to send a SACK chunk (from JR: i.e, after an
> INIT) it must fill in the
> Cumulative_TSN_Ack field. That value is initialized to the Initial_TSN of
> the peer minus
> one..."
> 
> Where does the "minus-1" come from?
> Section "3.3.4" (SACK) 3rd parag states "Less-than or equal..."
> 
> Please kindky advise...
> 
> 
> Jacques Rajchgod
> MCI Worldcom
> Next Generation System Evaluation
> Tel: [972] 729-7180
> E-mail: c-Jacques.Rajchgod@wcom.com
> 
> 
> 
> -----Original Message-----
> From: sigtran-admin@ietf.org [mailto:sigtran-admin@ietf.org]On Behalf Of
> Michael Tuexen
> Sent: Thursday, February 14, 2002 12:26 PM
> To: Jacques Rajchgod
> Cc: sigtran@ietf.org; tsvwg@ietf.org
> Subject: Re: [Sigtran] Issues with SCTP TSN and Cum TSN
> 
> 
> Dear Jacques,
> see my comments below.
> Best regards
> Michael
> On Thursday, February 14, 2002, at 07:13 PM, Jacques Rajchgod wrote:
> 
> > OK. Here it is:
> >
> > 1) a SENDS to b:	 -------------- INIT ---------------->
> > 2) a RECEIVES from b	<------------ INI_ACK ------------
> > 3) a SENDS to b	-------------- COOKIE ---------->
> > 4) a RECEIVES from b	<------------ COOKIE_ACK ------------
> > 3) a SENDS to b	--------------SCTP pack:  SACK (CUM_TSN=x) bundled with
> > ASP_UP ---------->
> >
> > - What's the "CUM_TSN=x" pointing to? Which DATA chunk is it ACK'ing
> > CUMULATIVELY??
> The value will be (Initial_TSN of the INIT_ACK chunk from b) - 1
> > - There hasn't been any DATA exchanged anywhere yet, to ack!
> It doesn't matter.
> > - A CUM_TSN is only used to ack CUMULATIVELY - DATA chunks. There are no
> > DATA chunks exchanged at INIT time. The DATA may come only after the
> > INIT.
> It is just the rule. b knows it and understands that nothing is
> acknowledged.
> - The INIT consists only of CONTROL chunks.
> It is a control chunk.
> > - So, what is a Receiver of that CUM_TSN=x ("b") supposed to look for?
> > "b"
> > hasn't sent any data to be ACK'ed cumulatively yet!
> B is looking for  (Initial_TSN of the INIT_ACK chunk from b) - 1, he
> knows the rule.
> > - Take a look at Section 1.4, page 11, the 5th bullet down - CUM_TSN...
> In general that is true. But in case of the initialization it is NOT
> true.
> That field is in each SACK chunk, so you have to put a value in it, so
> you
> have to define an initialization value.
> If you do not have anything to acknowledge you do not have to send a
> SACK. But
> if the first packets containing DATA chunks are lost and you have to
> acknowledge
> the first one you have to tell the peer that the first packets are lost.
> So you
> will also use the initialization value for the Cumulative_TSN_Ack.
> >
> > Brian recommended yesterday that the recipient of such SACK ignore that
> > SACK
> > and just update a_rwnd. Here is the statement:
> >
> > "The receiver of a SACK with Cummulative Ack of Y-1 should
> > discard the SACK and do nothing further with it other than
> > updating a_rwnd because it preceeds the Initial TSN.  This
> > is how an implementation acks nothing yet updates a_rwnd."
> >
> > Jacques Rajchgod
> > MCI Worldcom
> > Next Generation System Evaluation
> > Tel: [972] 729-7180
> > E-mail: c-Jacques.Rajchgod@wcom.com
> >
> >
> >
> > -----Original Message-----
> > From: Tuexen Michael [mailto:Michael.Tuexen@icn.siemens.de]
> > Sent: Thursday, February 14, 2002 10:35 AM
> > To: 'Jacques Rajchgod'; sigtran@ietf.org; 'tsvwg@ietf.org'
> > Subject: RE: [Sigtran] Issues with SCTP TSN and Cum TSN
> >
> >
> > Dear Jacques,
> >
> > the Cumulative_TSN_Ack is part of any SACK. I do not see
> > why this isn't good. And you must have a rule what to use
> > as an inital value because you must be able to handle the
> > case where the first packtet containing DATA chunks is
> > lost and the second arrives. You have to acknowledge them but
> > you must use the initial value as the cumulative_TSN_Ack.
> >
> > Best regards
> > Michael
> >
> > Michael Tuexen    Tel.:   +49 89 722 47210
> > Siemens AG        Fax:    +49 89 722 48212
> > ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
> >
> 
> 
> _______________________________________________
> Sigtran mailing list
> Sigtran@ietf.org
> https://www1.ietf.org/mailman/listinfo/sigtran

-- 
Brian F. G. Bidulock
bidulock@openss7.org
http://www.openss7.org/

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 15:43:10 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA09184
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 15:43:10 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA25483
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 15:43:13 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA24175;
	Thu, 14 Feb 2002 15:26:42 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA24051
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 15:26:35 -0500 (EST)
Received: from ilsa.franken.de (postfix@ilsa.franken.de [193.175.24.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA08891;
	Thu, 14 Feb 2002 15:26:27 -0500 (EST)
Received: from localhost (p508016CA.dip.t-dialin.net [80.128.22.202])
	by ilsa.franken.de (Postfix) with ESMTP
	id F1167245C5; Thu, 14 Feb 2002 21:26:13 +0100 (CET)
Date: Thu, 14 Feb 2002 21:24:19 +0100
Content-Type: text/plain; charset=US-ASCII; format=flowed
Mime-Version: 1.0 (Apple Message framework v480)
Cc: sigtran@ietf.org, tsvwg@ietf.org
To: Jacques Rajchgod <c-Jacques.Rajchgod@wcom.com>
From: Michael Tuexen <Michael.Tuexen@icn.siemens.de>
In-Reply-To: <NEBBLDAOEDIKLJPELCGEEEHPCIAA.c-Jacques.Rajchgod@wcom.com>
Message-Id: <D2E332A6-2188-11D6-82A5-0030654C1AB6@icn.siemens.de>
Content-Transfer-Encoding: 7bit
X-Mailer: Apple Mail (2.480)
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] Re: [Sigtran] Issues with SCTP TSN and Cum TSN
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Dear Jacques,
see my comments below.
Best regards
Michael
On Thursday, February 14, 2002, at 08:48 PM, Jacques Rajchgod wrote:

> Thanks, Michael...I'll run with that.
>
That's fine.
> Still, some questions:
>
> Your first mail on the topic contained the following comment:
>
> "...If an SCTP endpoint wants to send a SACK chunk (from JR: i.e, after 
> an
> INIT) it must fill in the
> Cumulative_TSN_Ack field. That value is initialized to the Initial_TSN 
> of
> the peer minus
> one..."
>
> Where does the "minus-1" come from?
> Section "3.3.4" (SACK) 3rd parag states "Less-than or equal..."
>
> Please kindky advise...
>
12.2. says

    Last Rcvd   : This is the last TSN received in sequence.  This value
    TSN         : is set initially by taking the peer's Initial TSN,
                : received in the INIT or INIT ACK chunk, and
                : subtracting one from it.

and the but-last paragraph of section 5.1 says

    When the TCB is created, each endpoint MUST set its internal
    Cumulative TSN Ack Point to the value of its transmitted Initial TSN
    minus one.

This value will be used to fill the field Cumulative_TSN_Ack in SACK 
chunks.

This is 'indicated' by the text in 3.3.4 (which doesn't describe the 
initialization)

    By definition, the value of the Cumulative TSN Ack parameter is the
    last TSN received before a break in the sequence of received TSNs
    occurs; the next TSN value following this one has not yet been
    received at the endpoint sending the SACK.  This parameter therefore
    acknowledges receipt of all TSNs less than or equal to its value.

It is not as clear as it could be but that is the way it is written in 
RFC 2960.
>
> Jacques Rajchgod
> MCI Worldcom
> Next Generation System Evaluation
> Tel: [972] 729-7180
> E-mail: c-Jacques.Rajchgod@wcom.com
>
>
>
> -----Original Message-----
> From: sigtran-admin@ietf.org [mailto:sigtran-admin@ietf.org]On Behalf Of
> Michael Tuexen
> Sent: Thursday, February 14, 2002 12:26 PM
> To: Jacques Rajchgod
> Cc: sigtran@ietf.org; tsvwg@ietf.org
> Subject: Re: [Sigtran] Issues with SCTP TSN and Cum TSN
>
>
> Dear Jacques,
> see my comments below.
> Best regards
> Michael
> On Thursday, February 14, 2002, at 07:13 PM, Jacques Rajchgod wrote:
>
>> OK. Here it is:
>>
>> 1) a SENDS to b:	 -------------- INIT ---------------->
>> 2) a RECEIVES from b	<------------ INI_ACK ------------
>> 3) a SENDS to b	-------------- COOKIE ---------->
>> 4) a RECEIVES from b	<------------ COOKIE_ACK ------------
>> 3) a SENDS to b	--------------SCTP pack:  SACK (CUM_TSN=x) bundled 
>> with
>> ASP_UP ---------->
>>
>> - What's the "CUM_TSN=x" pointing to? Which DATA chunk is it ACK'ing
>> CUMULATIVELY??
> The value will be (Initial_TSN of the INIT_ACK chunk from b) - 1
>> - There hasn't been any DATA exchanged anywhere yet, to ack!
> It doesn't matter.
>> - A CUM_TSN is only used to ack CUMULATIVELY - DATA chunks. There are 
>> no
>> DATA chunks exchanged at INIT time. The DATA may come only after the
>> INIT.
> It is just the rule. b knows it and understands that nothing is
> acknowledged.
> - The INIT consists only of CONTROL chunks.
> It is a control chunk.
>> - So, what is a Receiver of that CUM_TSN=x ("b") supposed to look for?
>> "b"
>> hasn't sent any data to be ACK'ed cumulatively yet!
> B is looking for  (Initial_TSN of the INIT_ACK chunk from b) - 1, he
> knows the rule.
>> - Take a look at Section 1.4, page 11, the 5th bullet down - CUM_TSN...
> In general that is true. But in case of the initialization it is NOT
> true.
> That field is in each SACK chunk, so you have to put a value in it, so
> you
> have to define an initialization value.
> If you do not have anything to acknowledge you do not have to send a
> SACK. But
> if the first packets containing DATA chunks are lost and you have to
> acknowledge
> the first one you have to tell the peer that the first packets are lost.
> So you
> will also use the initialization value for the Cumulative_TSN_Ack.
>>
>> Brian recommended yesterday that the recipient of such SACK ignore that
>> SACK
>> and just update a_rwnd. Here is the statement:
>>
>> "The receiver of a SACK with Cummulative Ack of Y-1 should
>> discard the SACK and do nothing further with it other than
>> updating a_rwnd because it preceeds the Initial TSN.  This
>> is how an implementation acks nothing yet updates a_rwnd."
>>
>> Jacques Rajchgod
>> MCI Worldcom
>> Next Generation System Evaluation
>> Tel: [972] 729-7180
>> E-mail: c-Jacques.Rajchgod@wcom.com
>>
>>
>>
>> -----Original Message-----
>> From: Tuexen Michael [mailto:Michael.Tuexen@icn.siemens.de]
>> Sent: Thursday, February 14, 2002 10:35 AM
>> To: 'Jacques Rajchgod'; sigtran@ietf.org; 'tsvwg@ietf.org'
>> Subject: RE: [Sigtran] Issues with SCTP TSN and Cum TSN
>>
>>
>> Dear Jacques,
>>
>> the Cumulative_TSN_Ack is part of any SACK. I do not see
>> why this isn't good. And you must have a rule what to use
>> as an inital value because you must be able to handle the
>> case where the first packtet containing DATA chunks is
>> lost and the second arrives. You have to acknowledge them but
>> you must use the initial value as the cumulative_TSN_Ack.
>>
>> Best regards
>> Michael
>>
>> Michael Tuexen    Tel.:   +49 89 722 47210
>> Siemens AG        Fax:    +49 89 722 48212
>> ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
>>
>
>
> _______________________________________________
> Sigtran mailing list
> Sigtran@ietf.org
> https://www1.ietf.org/mailman/listinfo/sigtran
>
>
Michael.Tuexen@micmac.franken.de


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 19:21:36 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA13333
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 19:21:36 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id TAA05359
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 19:21:37 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA05157;
	Thu, 14 Feb 2002 19:13:14 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA05130
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 19:13:12 -0500 (EST)
Received: from msgbas1.cos.agilent.com (msgbas1x.cos.agilent.com [192.25.240.36])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA13225
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 19:13:10 -0500 (EST)
Received: from msgrel1.cos.agilent.com (msgrel1.cos.agilent.com [130.29.152.77])
	by msgbas1.cos.agilent.com (Postfix) with ESMTP
	id D61AD3270; Thu, 14 Feb 2002 17:13:11 -0700 (MST)
Received: from axcsbh1.cos.agilent.com (axcsbh1.cos.agilent.com [130.29.152.143])
	by msgrel1.cos.agilent.com (Postfix) with ESMTP
	id 92C302DC; Thu, 14 Feb 2002 17:13:11 -0700 (MST)
Received: by axcsbh1.cos.agilent.com with Internet Mail Service (5.5.2653.19)
	id <17QY8793>; Thu, 14 Feb 2002 17:13:11 -0700
Message-ID: <1BEBA5E8600DD4119A50009027AF54A00A512287@axcs04.cos.agilent.com>
From: "THALER,PAT (A-Roseville,ex1)" <pat_thaler@agilent.com>
To: "WENDT,JIM (HP-Roseville,ex1)" <Jim_Wendt@hp.com>,
        "'Stephen Bailey'" <steph@cs.uchicago.edu>,
        Julian Satran <Julian_Satran@il.ibm.com>
Cc: rdma@yahoogroups.com, tsvwg@ietf.org
Date: Thu, 14 Feb 2002 17:13:09 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Subject: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch framer for TC
 P
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Jim,

I don't understand what you are suggesting. It seems like an "aligning FIM"
approach could only work for a ULP that had fixed length PDUs - which would
make it a pretty narrow solution. If the PDUs aren't fixed length, then the
markers which are at fixed intervals won't necessarily fall into the same
PDU as the aligned packets so it doesn't ensure ability to place the aligned
packet right after a packet drop. Therefore, it isn't clear tha anything is
gained by alignment.

Or do you just mean that FIM and one of the PDU alignment schemes can be
applied to the same stream so that FIM can be used if resegmentation
destroys the alignment? That would be possible, but it is kind of a belt and
suspenders approach. 

Regards,
Pat

-----Original Message-----
From: WENDT,JIM (HP-Roseville,ex1) [mailto:Jim_Wendt@hp.com]
Sent: Wednesday, February 13, 2002 11:16 AM
To: 'Stephen Bailey'; Julian Satran
Cc: rdma@yahoogroups.com; tsvwg@ietf.org
Subject: RE: [rdma] a proposal for a different No-Data-Touch framer for
TC P


> I see it clearly.  However, this seems to be a case where we can't
> have everything.  If we want to touch the data we can have a
> synchronization technique which can also be used to determine PDU
> alignment (COWS).  If we don't want to touch the data, we can have PDU
> alignment (TUF or NDT) OR synchronization (FIM), but not both.

Not to muddy the waters, but isn't an "aligning FIM" possible
(not that anyone would want to do it), where ULP PDUs (possibly
chunked) are aligned into TCP segments allowing direct placement
of each TCP segment payload, and with FIM to allow location of
subsequent ULP PDUs after a packet drop?

Jim

------------------------ Yahoo! Groups Sponsor ---------------------~-->
Sponsored by VeriSign - The Value of Trust
Secure all your Web servers now - with a proven 5-part
strategy. The FREE Server Security Guide shows you how.
http://us.click.yahoo.com/iWSNbC/VdiDAA/yigFAA/W6uqlB/TM
---------------------------------------------------------------------~->

To unsubscribe from this group, send an email to:
rdma-unsubscribe@egroups.com

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ 


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 14 20:42:10 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA14611
	for <tsvwg-archive@odin.ietf.org>; Thu, 14 Feb 2002 20:42:10 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA08343
	for tsvwg-archive@odin.ietf.org; Thu, 14 Feb 2002 20:42:09 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA08216;
	Thu, 14 Feb 2002 20:35:36 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA08185
	for <tsvwg@optimus.ietf.org>; Thu, 14 Feb 2002 20:35:34 -0500 (EST)
Received: from msgbas1.cos.agilent.com (msgbas1x.cos.agilent.com [192.25.240.36])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA14489
	for <tsvwg@ietf.org>; Thu, 14 Feb 2002 20:35:31 -0500 (EST)
Received: from msgrel1.cos.agilent.com (msgrel1.cos.agilent.com [130.29.152.77])
	by msgbas1.cos.agilent.com (Postfix) with ESMTP
	id DD7982FA4; Thu, 14 Feb 2002 18:35:33 -0700 (MST)
Received: from axcsbh1.cos.agilent.com (axcsbh1.cos.agilent.com [130.29.152.143])
	by msgrel1.cos.agilent.com (Postfix) with ESMTP
	id 957EE170; Thu, 14 Feb 2002 18:35:33 -0700 (MST)
Received: by axcsbh1.cos.agilent.com with Internet Mail Service (5.5.2653.19)
	id <17QY804P>; Thu, 14 Feb 2002 18:35:33 -0700
Message-ID: <1BEBA5E8600DD4119A50009027AF54A00A5122B6@axcs04.cos.agilent.com>
From: "THALER,PAT (A-Roseville,ex1)" <pat_thaler@agilent.com>
To: "WENDT,JIM (HP-Roseville,ex1)" <jim_wendt@hp.com>,
        "'Stephen Bailey'" <steph@cs.uchicago.edu>,
        John Hufferd <hufferd@us.ibm.com>
Cc: tsvwg@ietf.org
Subject: RE: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch f
	 ramer for TCP
Date: Thu, 14 Feb 2002 18:35:32 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

All,

If you use this as in the TUF proposal, then false positives are only
important when the first re-segmented and therefore non-aligned packets
arrive just as packets have been dropped. Since the stream may have gone
from not being resegmented to being re-segmented because of a path change,
it may not be unusual for re-segmentation to happen at the same time that
packets are arriving out of order or packet drops have occurred.

Currently the text in the ulp frame documenmt says: 
     "A TUF implementation which detects a very low
     incidence of the ULPDU containment property for a sustained
     interval (>> RTT) may assume that a resegmenting intermediary is in
     operation and SHOULD discontinue the use of ULP control information
     found using the ULPDU containment property.  "
but, one could reduce the vulnerability by setting a low threshold for
turning off use of ULP when packets have failed containment and requiring
detection of containment for multiple segments before relying on containment
again.

Both TUF and ESC have about the same false positive probabiliy when testing
a single instance at the start of a segment. TUF requires 64 bits to match
the segment length and the key and ESC requires its 96 bit header to match
one of 2^32 correct headers. Note that including data not in the header (a
shared "secret") in the CRC calculation doesn't change this probability. It
is still matching 2^96 data patterns against 2^32 correct answers. The only
thing a shared secret would help is the case where the ULP also covered 64
bits with the same 32-bit CRC because that would yield a false positive when
time the first 32 bits was the framing pattern.

If data is randomly distributed, then probability of a false-positive when a
resegmenting middle box has destroyed containment is ~5e-20. For comparison,
failure of a 32-bit CRC to detect an error longer than its burst detection
and hamming distance is ~2e-10. 5e-20 is a sufficient protection against
false positives for a rare event. If one was gaining and losing containment
once a second it would still take about 1e12 years to get a false positive
just as one was losing containment. Data is of course not randomly
distributed and it is possible that the key conincides with a pattern that
occurs more frequently in the data stream then it would occur in a random
data stresm. For instance, if a machine at my company chose a key that
happened to coincide with part of the ASCII string for Agilent. But, if one
is only testing the first words of segment to detect containment, one can
tolerate a reasonable degredation below 5e-20 for a false positive.

On the other hand, the NDT proposal implies that its method could be used
even in the absence of containment. I don't think that the protection
against false positives is high enough to be robust in that situation. It
doesn't provide a length field, so when packets are arriving in order, it
will either have to use a ULP specific field to decide where to search for
the next candidate framing header or it will be checking every 4 bytes for a
96 bit match. In the latter case, it is checking 3e7 times per second if
data is arriving at about 1 Gbit/s. At that rate with random data, it takes
about 1e5 years to get a false positive, but degredation due to
non-randomness of data could pull the number much lower.

It is not clear to me that either method is robust enough to use when
containment has been lost because we can't quantify how often the algorithm
will have to search every 4 bytes for the header due to out-of-order
arrival. Having a length field in the header will be more robust because
once a header is found in an exhaustive search, subsequent headers can be
found based on the length field. It is even possible to reduce the false
positive rate by only accepting a header found by exhaustive search if the
location pointed to by the length field also contains a valid header.

Regards,
Pat Thaler

-----Original Message-----
From: WENDT,JIM (HP-Roseville,ex1) [mailto:jim_wendt@hp.com]
Sent: Wednesday, February 13, 2002 11:00 AM
To: 'Stephen Bailey'; John Hufferd
Cc: tsvwg@ietf.org
Subject: RE: [Tsvwg] Re: [rdma] a proposal for a different No-Data-Touch
f ramer for TCP


> The other point to note is that you only play the dice at all if
> resegmentation DOES result from an intermediary, since the source will
> only send segments which are known not to fail false-positively.  This
> technique applies equally to Julian's proposal.

And furthermore, in the case of TUF, only when there has been a
drop of a packet contain a TUF header. Since the length field chains to
the next TUF header, the receiver only has to succumb to probability
when it has lost its way. This would seem to reduce the rate of
occurrence somewhat depending on the packet loss rate.

Jim

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 15 01:32:09 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA20675
	for <tsvwg-archive@odin.ietf.org>; Fri, 15 Feb 2002 01:32:09 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id BAA00005
	for tsvwg-archive@odin.ietf.org; Fri, 15 Feb 2002 01:32:05 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA28808;
	Fri, 15 Feb 2002 01:22:42 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id BAA28783
	for <tsvwg@optimus.ietf.org>; Fri, 15 Feb 2002 01:22:41 -0500 (EST)
Received: from david.siemens.com.sg (david.siemens.com.sg [194.138.240.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA20449
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 01:22:40 -0500 (EST)
X-Envelope-Sender-Is: ExchAdmin@sae.siemens.com.sg (at relayer david.siemens.com.sg)
Received: from mail.siemens.com.sg (mail.siemens.com.sg [194.138.237.71])
	by david.siemens.com.sg (8.10.2/8.10.2) with ESMTP id g1F6McB09810
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 14:22:39 +0800 (SGT)
Received: from SGPK001A.sae.siemens.com.sg (sgpk001a.sae.siemens.com.sg [140.231.107.1])
	by mail.siemens.com.sg (8.11.4/8.11.4) with ESMTP id g1F6Mcj27254
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 14:22:38 +0800 (SGT)
Received: by sgpk001a.sae.siemens.com.sg with Internet Mail Service (5.5.2653.19)
	id <1VM4Q3VC>; Fri, 15 Feb 2002 14:15:19 +0800
Message-ID: <1FCAB2AB1194A040B37F3035790C6345C84EE3@sgpk001a.sae.siemens.com.sg>
From: Exchange Administrator <ExchAdmin@sae.siemens.com.sg>
To: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Date: Fri, 15 Feb 2002 14:15:19 +0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: multipart/mixed;
	boundary="----_=_NextPart_000_01C1B5E8.245C3F70"
Subject: [Tsvwg] Recovered Mails from 23 Jan 2002
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_000_01C1B5E8.245C3F70
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C1B5E8.245C3F70"


------_=_NextPart_001_01C1B5E8.245C3F70
Content-Type: text/plain;
	charset="iso-8859-1"

 <<D3YP3B02.txt>> 

------_=_NextPart_001_01C1B5E8.245C3F70
Content-Type: text/html;
	charset="iso-8859-1"

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Generator" CONTENT="MS Exchange Server version 5.5.2653.12">
<TITLE>Recovered Mails from 23 Jan 2002</TITLE>
</HEAD>
<BODY>

<P><FONT FACE="Arial" SIZE=2 COLOR="#000000"> &lt;&lt;D3YP3B02.txt&gt;&gt; </FONT>
</P>

</BODY>
</HTML>
------_=_NextPart_001_01C1B5E8.245C3F70--

------_=_NextPart_000_01C1B5E8.245C3F70
Content-Type: text/plain;
	name="D3YP3B02.txt"
Content-Disposition: attachment;
	filename="D3YP3B02.txt"


To: tsvwg@ietf.org
Subject: Re: [Tsvwg] TCP Smart Framing 


Date: Tue, 22 Jan 2002 18:03:32 -0500
From: Stephen Bailey <steph@cs.uchicago.edu>


> My understanding of routers is pretty rough.  
> [...]
> All packets get one slot.

They actually work about every different way you could imagine, so
it's probably sound not to reason about them much one way or the
other.

Most `new' routers (with hardware-based forwarding planes) don't use
the one slot per packet technique because it results in a substantial
waste of buffering space if the packet stream has many small packets.
Slot-based design usually assume they are only queuing a small
amount of data internally, as might be the case in an older Ethernet
switch. 

There are typically some slotish things in the header processing
parts, but they are provisioned to handle minimum sized packets at
line rate.  In some architectures if you start to turn on a lot of
per-packet features (policing, classification, counters, etc.)  you
may see a (mild or severe) degradation in forwarding performance as a
function of packet rate.

Steph

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg

------_=_NextPart_000_01C1B5E8.245C3F70--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 15 03:59:09 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA00448
	for <tsvwg-archive@odin.ietf.org>; Fri, 15 Feb 2002 03:59:08 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id DAA07009
	for tsvwg-archive@odin.ietf.org; Fri, 15 Feb 2002 03:59:10 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA06312;
	Fri, 15 Feb 2002 03:46:46 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA06280
	for <tsvwg@optimus.ietf.org>; Fri, 15 Feb 2002 03:46:44 -0500 (EST)
Received: from mgw-x1.nokia.com (mgw-x1.nokia.com [131.228.20.21])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA00221
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 03:46:41 -0500 (EST)
From: john.loughney@nokia.com
Received: from esvir05nok.ntc.nokia.com (esvir05nokt.ntc.nokia.com [172.21.143.37])
	by mgw-x1.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1F8kF300371
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 10:46:15 +0200 (EET)
Received: from esebh004.NOE.Nokia.com (unverified) by esvir05nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5915cf616dac158f25245@esvir05nok.ntc.nokia.com> for <tsvwg@ietf.org>;
 Fri, 15 Feb 2002 10:46:41 +0200
Received: from esebe004.NOE.Nokia.com ([172.21.138.44]) by esebh004.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Fri, 15 Feb 2002 10:46:41 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Date: Fri, 15 Feb 2002 10:46:40 +0200
Message-ID: <0C1353ABB1DEB74DB067ADFF749C4EEFC648B0@esebe004.NOE.Nokia.com>
Thread-Topic: [Tsvwg] SCTP Implementer's guide question...
Thread-Index: AcGz7S0kJS35y1SmQkmR6bwgHUCEIgCD/NGw
To: <tsvwg@ietf.org>
X-OriginalArrivalTime: 15 Feb 2002 08:46:41.0854 (UTC) FILETIME=[4A277DE0:01C1B5FD]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id DAA06281
Subject: [Tsvwg] SCTP Checksum issue
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi all,

What is the current status of the SCTP Checksum draft?  This has completed
WG last call - will this be sent to the IESG shortly?

Thanks,
John

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 15 04:02:42 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA00509
	for <tsvwg-archive@odin.ietf.org>; Fri, 15 Feb 2002 04:02:42 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id EAA07494
	for tsvwg-archive@odin.ietf.org; Fri, 15 Feb 2002 04:02:43 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA06837;
	Fri, 15 Feb 2002 03:55:30 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id DAA06797
	for <tsvwg@optimus.ietf.org>; Fri, 15 Feb 2002 03:55:28 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA00370
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 03:55:25 -0500 (EST)
From: john.loughney@nokia.com
Received: from esvir04nok.ntc.nokia.com (esvir04nokt.ntc.nokia.com [172.21.143.36])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1F8tXZ22452
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 10:55:33 +0200 (EET)
Received: from esebh004.NOE.Nokia.com (unverified) by esvir04nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5915d760dcac158f24077@esvir04nok.ntc.nokia.com> for <tsvwg@ietf.org>;
 Fri, 15 Feb 2002 10:55:25 +0200
Received: from esebe004.NOE.Nokia.com ([172.21.138.44]) by esebh004.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Fri, 15 Feb 2002 10:55:25 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Date: Fri, 15 Feb 2002 10:55:25 +0200
Message-ID: <0C1353ABB1DEB74DB067ADFF749C4EEFC648B3@esebe004.NOE.Nokia.com>
Thread-Topic: [Tsvwg] SCTP Implementer's guide question...
Thread-Index: AcGz7S0kJS35y1SmQkmR6bwgHUCEIgCD/NGwAABNaXA=
To: <tsvwg@ietf.org>
X-OriginalArrivalTime: 15 Feb 2002 08:55:25.0956 (UTC) FILETIME=[828B1C40:01C1B5FE]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id DAA06798
Subject: [Tsvwg] SCTP Add IP Extension status
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi all,

This draft is looking quite stable - is there a chance that we can get a work group last call after Minneapolis?

John

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 15 07:01:01 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA02460
	for <tsvwg-archive@odin.ietf.org>; Fri, 15 Feb 2002 07:01:00 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA15447
	for tsvwg-archive@odin.ietf.org; Fri, 15 Feb 2002 07:01:01 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA14801;
	Fri, 15 Feb 2002 06:52:08 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA14770
	for <tsvwg@optimus.ietf.org>; Fri, 15 Feb 2002 06:52:05 -0500 (EST)
Received: from mgw-x1.nokia.com (mgw-x1.nokia.com [131.228.20.21])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA02309
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 06:52:03 -0500 (EST)
From: john.loughney@nokia.com
Received: from esvir05nok.ntc.nokia.com (esvir05nokt.ntc.nokia.com [172.21.143.37])
	by mgw-x1.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1FBpa313417
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 13:51:36 +0200 (EET)
Received: from esebh004.NOE.Nokia.com (unverified) by esvir05nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5916791391ac158f257d4@esvir05nok.ntc.nokia.com>;
 Fri, 15 Feb 2002 13:52:02 +0200
Received: from esebe004.NOE.Nokia.com ([172.21.138.44]) by esebh004.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Fri, 15 Feb 2002 13:52:02 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] SCTP Add IP Extension status 
Date: Fri, 15 Feb 2002 13:51:58 +0200
Message-ID: <0C1353ABB1DEB74DB067ADFF749C4EEFC648C7@esebe004.NOE.Nokia.com>
Thread-Topic: [Tsvwg] SCTP Add IP Extension status 
Thread-Index: AcG2FrJ3XtgdUGGQRI2RUrKDUUhc/QAAG08A
To: <mankin@isi.edu>
Cc: <tsvwg@ietf.org>
X-OriginalArrivalTime: 15 Feb 2002 11:52:03.0043 (UTC) FILETIME=[2EE64330:01C1B617]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id GAA14771
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Hi Allison,


> As I said in the meeting in SLC, it needs some reviews - it's
> stable but it has had no discussion.  A review of it and some 
> discussion before WGLC would be desirable.

Time to crank out a review, I guess ...

John

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 15 07:02:31 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA02511
	for <tsvwg-archive@odin.ietf.org>; Fri, 15 Feb 2002 07:02:31 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA15478
	for tsvwg-archive@odin.ietf.org; Fri, 15 Feb 2002 07:02:32 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA14611;
	Fri, 15 Feb 2002 06:48:40 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA14586
	for <tsvwg@optimus.ietf.org>; Fri, 15 Feb 2002 06:48:38 -0500 (EST)
Received: from minotaur.nge.isi.edu (minotaur.nge.isi.edu [65.114.169.202])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA02252
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 06:48:36 -0500 (EST)
Received: from minotaur (mankin@localhost)
	by minotaur.nge.isi.edu (8.11.6/8.11.6) with ESMTP id g1FBcHO14904;
	Fri, 15 Feb 2002 06:38:19 -0500
Message-Id: <200202151138.g1FBcHO14904@minotaur.nge.isi.edu>
To: john.loughney@nokia.com
cc: tsvwg@ietf.org
Reply-To: mankin@isi.edu
Subject: Re: [Tsvwg] SCTP Add IP Extension status 
In-reply-to: Your message of Fri, 15 Feb 2002 10:55:25 +0200.
             <0C1353ABB1DEB74DB067ADFF749C4EEFC648B3@esebe004.NOE.Nokia.com> 
Mime-Version: 1.0 (generated by tm-edit 1.7)
Content-Type: text/plain; charset=US-ASCII
Date: Fri, 15 Feb 2002 06:38:17 -0500
From: Allison Mankin <mankin@isi.edu>
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

John,

As I said in the meeting in SLC, it needs some reviews - it's
stable but it has had no discussion.  A review of it and some discussion
before WGLC would be desirable.

Allison

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 15 07:05:07 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA02723
	for <tsvwg-archive@odin.ietf.org>; Fri, 15 Feb 2002 07:05:07 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA15704
	for tsvwg-archive@odin.ietf.org; Fri, 15 Feb 2002 07:05:08 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA14676;
	Fri, 15 Feb 2002 06:49:22 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA14647
	for <tsvwg@optimus.ietf.org>; Fri, 15 Feb 2002 06:49:20 -0500 (EST)
Received: from minotaur.nge.isi.edu (minotaur.nge.isi.edu [65.114.169.202])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA02257
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 06:49:19 -0500 (EST)
Received: from minotaur (mankin@localhost)
	by minotaur.nge.isi.edu (8.11.6/8.11.6) with ESMTP id g1FBhtm14915;
	Fri, 15 Feb 2002 06:43:55 -0500
Message-Id: <200202151143.g1FBhtm14915@minotaur.nge.isi.edu>
To: john.loughney@nokia.com
cc: tsvwg@ietf.org
Reply-To: mankin@isi.edu
Subject: Re: [Tsvwg] SCTP Checksum issue 
In-reply-to: Your message of Fri, 15 Feb 2002 10:46:40 +0200.
             <0C1353ABB1DEB74DB067ADFF749C4EEFC648B0@esebe004.NOE.Nokia.com> 
Mime-Version: 1.0 (generated by tm-edit 1.7)
Content-Type: text/plain; charset=US-ASCII
Date: Fri, 15 Feb 2002 06:43:55 -0500
From: Allison Mankin <mankin@isi.edu>
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

John,

Thanks for the ping.  An 03 rev is due for this addressing several
WGLC comments.  There may be a state transition stall because
of ADs and authors miscommunicating, so your ping is a good reminder
to refresh among ourselves and get this moving again.

IETF Last Call should start next week.

Allison

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 15 07:16:36 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA02974
	for <tsvwg-archive@odin.ietf.org>; Fri, 15 Feb 2002 07:16:35 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA16116
	for tsvwg-archive@odin.ietf.org; Fri, 15 Feb 2002 07:16:37 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA15889;
	Fri, 15 Feb 2002 07:08:53 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA15862
	for <tsvwg@optimus.ietf.org>; Fri, 15 Feb 2002 07:08:52 -0500 (EST)
Received: from gorilla.mchh.siemens.de (gorilla.mchh.siemens.de [194.138.158.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA02784
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 07:08:49 -0500 (EST)
Received: from blues.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.227])
	by gorilla.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id NAA18058;
	Fri, 15 Feb 2002 13:08:46 +0100 (MET)
Received: from mchh273e.demchh201e.icn.siemens.de ([139.21.200.83])
	by blues.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id NAA05133;
	Fri, 15 Feb 2002 13:08:40 +0100 (MET)
Received: by MCHH273E with Internet Mail Service (5.5.2653.19)
	id <1JWX0V7F>; Fri, 15 Feb 2002 13:08:42 +0100
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF765@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'john.loughney@nokia.com'" <john.loughney@nokia.com>, mankin@isi.edu
Cc: tsvwg@ietf.org
Subject: RE: [Tsvwg] SCTP Add IP Extension status 
Date: Fri, 15 Feb 2002 13:08:32 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

The AddIP extension will be tested at connectathon
(www.connectathon.org) which will take place end of
February. This should result in some discussions, at
least between the participants.

This brings up another question:
Who will participate testing SCTP and the AddIP
draft in particular? 

We will be there with our prototype implementation.

Best regards
Michael 

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	john.loughney@nokia.com [SMTP:john.loughney@nokia.com]
> Sent:	Friday, February 15, 2002 12:52 PM
> To:	mankin@isi.edu
> Cc:	tsvwg@ietf.org
> Subject:	RE: [Tsvwg] SCTP Add IP Extension status 
> 
> Hi Allison,
> 
> 
> > As I said in the meeting in SLC, it needs some reviews - it's
> > stable but it has had no discussion.  A review of it and some 
> > discussion before WGLC would be desirable.
> 
> Time to crank out a review, I guess ...
> 
> John
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 15 08:17:55 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA04494
	for <tsvwg-archive@odin.ietf.org>; Fri, 15 Feb 2002 08:17:55 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA19152
	for tsvwg-archive@odin.ietf.org; Fri, 15 Feb 2002 08:17:56 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA18639;
	Fri, 15 Feb 2002 08:01:36 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA18602
	for <tsvwg@optimus.ietf.org>; Fri, 15 Feb 2002 08:01:33 -0500 (EST)
Received: from hrtades9.atea.be (viap111.atea.be [194.78.143.111])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA04036;
	Fri, 15 Feb 2002 08:01:24 -0500 (EST)
Received: from hrtades10.atea.be (siemens.atea.be [139.10.143.141]) by hrtades9.atea.be with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13)
	id DF4YTJS4; Fri, 15 Feb 2002 14:00:43 +0100
Received: by siemens.atea.be with Internet Mail Service (5.5.2653.19)
	id <19TAJQH8>; Fri, 15 Feb 2002 14:01:05 +0100
Message-ID: <E76F715C0429D5118F2100508BB9EDEE0795BD@hrtades7.atea.be>
From: Coene Lode <Lode.Coene@siemens.atea.be>
To: "Transport Area wg (E-mail)" <tsvwg@ietf.org>,
        "SIGTRAN (E-mail)"
	 <sigtran@ietf.org>
Cc: "Multi6 (E-mail)" <multi6@ops.ietf.org>
Date: Fri, 15 Feb 2002 13:58:09 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: [Tsvwg] updated text SCTP  Multihoming issues
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Folks,

Thanks for all the commentaries received so far. 
The only remaining issue was the text of of paragraph 2.3 "SCTP multihoming
and the size of routing tables"

This is the new version.

    As multihoming means that more than one destination address is used
    on the host, that would mean that a routing descision must be made
    on the host in IP. The host does not know beforehand to which other
    host it is going to send something, so that would in theory require
    that all possible paths to all possible destinations should be known
    on that host. This amounts to the host being a part of the
    distribution of the routing information in the network.

    Possible solutions would require to ask only for the paths to host
    that are actually in use(meaning a association is about to be setup
    with that particular host). This is a viable solution for hosts with
    a small number of associations to different hosts.

    If the host has many associations with a lot of different host then
    then this becomes cumbersome(getting the specific paths from the
    routers and the updates and all) and leads in practice to same
    problem of distributing prefixes from the edge router(s) to the
    host.

    It might be useful to explore ways where no distribution of routing
    information to the host for using multihoming is needed or where the
    interface/link selection is not based on the use of different
    prefixes. Not all hosts have facilities for containing possible
    large routing tables/databases.


The document draft-coene-sctp-multihome-02.txt can be found at:
http://search.ietf.org/internet-drafts/draft-coene-sctp-multihome-02.txt    

Suggestion, clarification, corrections, are welcome.
I'll publish the updated(-03) draft wednesday 20/02/02

Yours sincerly,
Lode

Siemens atea
atealaan 34          2200 Herentals, Belgium
E-mail: lode.coene@siemens.atea.be
Tel: +32-14-252081
Fax: +32-14-253212

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 15 14:41:01 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA17483
	for <tsvwg-archive@odin.ietf.org>; Fri, 15 Feb 2002 14:41:01 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA13176
	for tsvwg-archive@odin.ietf.org; Fri, 15 Feb 2002 14:41:03 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA12732;
	Fri, 15 Feb 2002 14:32:06 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA12703
	for <tsvwg@optimus.ietf.org>; Fri, 15 Feb 2002 14:32:04 -0500 (EST)
Received: from sandmail.sandburst.com (sandmail.sandburst.com [216.57.132.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA17070
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 14:32:01 -0500 (EST)
To: "THALER,PAT (A-Roseville,ex1)" <pat_thaler@agilent.com>
Cc: tsvwg@ietf.org
Subject: Re: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch framer for TC P 
In-Reply-To: Message from "THALER,PAT (A-Roseville,ex1)" <pat_thaler@agilent.com> 
   of "Thu, 14 Feb 2002 17:13:09 MST." <1BEBA5E8600DD4119A50009027AF54A00A512287@axcs04.cos.agilent.com> 
References: <1BEBA5E8600DD4119A50009027AF54A00A512287@axcs04.cos.agilent.com> 
Date: Fri, 15 Feb 2002 14:31:18 -0500
From: Stephen Bailey <steph@cs.uchicago.edu>
Message-Id: <20020215193130.37B024E8F@sandmail.sandburst.com>
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Pat,

> I don't understand what you are suggesting.

Actually, it seems you do understand what he's suggesting, but are not
buying it :^)

The way I'd define it, it's not that the ULP PDUs are fixed length,
but that the framing protocol may add padding within a finite
interval.  For best transport efficiency, you would need to do what
TUF calls `ULPDU packing'.  You may end up with more than one ULPDU
per finite interval.

It actually seems like it might be a pretty good scheme.  Your finite
interval would be EMSS, and the segmentation algorithm would be
similar to the current TUF except to get markers at the beginning of
segments.

This might be a new leading candidate.  It's no touch and completely
safe.  The cost is that you must either do ULPDU packing (the
unpacking is actually the burdensome part, but I personally never
thought it was too horrible), or live with inefficient transport of
small messages.  For bulk data protocols (e.g. storage) the
overwhelming majority of the data you carry is in large
(interval-sized) PDUs, so a small loss of transport efficiency for the
control PDUs won't hurt.

Alright, you've done it again Jim!  By muddying the waters we clarify
them :-?

Steph


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 15 16:41:09 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA20472
	for <tsvwg-archive@odin.ietf.org>; Fri, 15 Feb 2002 16:41:09 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA19781
	for tsvwg-archive@odin.ietf.org; Fri, 15 Feb 2002 16:41:10 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA19457;
	Fri, 15 Feb 2002 16:33:46 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA19423
	for <tsvwg@optimus.ietf.org>; Fri, 15 Feb 2002 16:33:44 -0500 (EST)
Received: from minotaur.nge.isi.edu (minotaur.nge.isi.edu [65.114.169.202])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA20297
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 16:33:42 -0500 (EST)
Received: from minotaur (mankin@localhost)
	by minotaur.nge.isi.edu (8.11.6/8.11.6) with ESMTP id g1FLXgh15440;
	Fri, 15 Feb 2002 16:33:42 -0500
Message-Id: <200202152133.g1FLXgh15440@minotaur.nge.isi.edu>
To: tsvwg@ietf.org
Cc: ekr@rtfm.com, ajung@exp-math.uni-essen.de, Michael.Tuexen@icn.siemens.de
Reply-To: mankin@isi.edu
Mime-Version: 1.0 (generated by tm-edit 1.7)
Content-Type: text/plain; charset=US-ASCII
Date: Fri, 15 Feb 2002 16:33:42 -0500
From: Allison Mankin <mankin@isi.edu>
Subject: [Tsvwg] TSVWG Last Call on draft-tsvwg-tsl-over-sctp-00.txt
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Transport Working Group,

This is a working group last call, as hummed for in SLC, on
the TLS over SCTP draft.

It waited a little for this review, but in the end that may be
reasonable, because a specification of AES ciphersuites for
TLS has passed IESG in the meantime:

  draft-ietf-tls-ciphersuite-06.txt

There's a need for a minimum mandatory to implement ciphersuite
in the TLS over SCTP draft, and so please also consider the
suggestion that it be

TLS_RSA_WITH_AES_128_CBC_SHA

Last call will end February 28.  Target of the spec is Proposed
Standard.


Allison and Scott

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 15 20:16:40 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24224
	for <tsvwg-archive@odin.ietf.org>; Fri, 15 Feb 2002 20:16:39 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA28817
	for tsvwg-archive@odin.ietf.org; Fri, 15 Feb 2002 20:16:40 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA28442;
	Fri, 15 Feb 2002 20:05:44 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA28413
	for <tsvwg@optimus.ietf.org>; Fri, 15 Feb 2002 20:05:42 -0500 (EST)
Received: from palrel13.hp.com (palrel13.hp.com [156.153.255.238])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24128
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 20:05:41 -0500 (EST)
Received: from xparelay2.corp.hp.com (xparelay2.corp.hp.com [15.58.137.112])
	by palrel13.hp.com (Postfix) with ESMTP
	id EB76D40047E; Fri, 15 Feb 2002 17:05:11 -0800 (PST)
Received: from xpabh2.corp.hp.com (xpabh2.corp.hp.com [15.58.136.192])
	by xparelay2.corp.hp.com (Postfix) with ESMTP
	id AD391179; Fri, 15 Feb 2002 17:05:11 -0800 (PST)
Received: by xpabh2.corp.hp.com with Internet Mail Service (5.5.2653.19)
	id <1FZ29LPT>; Fri, 15 Feb 2002 17:05:11 -0800
Message-ID: <499DC368E25AD411B3F100902740AD650AFC7DB9@xrose03.rose.hp.com>
From: "WENDT,JIM (HP-Roseville,ex1)" <jim_wendt@hp.com>
To: "'Stephen Bailey'" <steph@cs.uchicago.edu>,
        "THALER,PAT (A-Roseville,ex1)" <pat_thaler@agilent.com>
Cc: tsvwg@ietf.org
Subject: RE: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch f
	ramer for TC P 
Date: Fri, 15 Feb 2002 17:05:08 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Steph & Pat,
Let's call the new scheme AFIM (Aligning FIM). My thought for AFIM
was that one or more ULP PDUs are aligned with, and packed
into each TCP segment ala TUF's 'ULPDU packing'. Thus, each TCP
segment starts with a new ULP PDU and contains one or more complete
ULP PDUs.

But my initial idea for the re-synchronization markers was to continue
to have the markers occur at a fixed interval in the sequence stream as
in the original FIM. Pat is right that the marker's forward pointer
might point to a ULP PDU that wasn't in the same TCP segment as the
marker. This scheme is no touch (direct placement of each TCP payload)
unless the stream has been resegmented and alignment has been broken,
and a packet drop occurs. I still think this is better than straight
FIM which causes most ULP PDUs to span across TCP segments and requires
more complex processing to reconstruct ULP PDUs before placing them.

Steph's variation, if I understand it, is different in that the FIM
interval is set to the EMSS, the space between each marker contains one
or more complete ULP PDUs followed by padding to fill the EMSS interval,
and TCP segments are constructed such that the FIM markers fall at the
beginning of each TCP segment. I'm not sure what happens across PMTU
changes in this scheme.

Jim


> Pat,
> 
> > I don't understand what you are suggesting.
> 
> Actually, it seems you do understand what he's suggesting, but are not
> buying it :^)
> 
> The way I'd define it, it's not that the ULP PDUs are fixed length,
> but that the framing protocol may add padding within a finite
> interval.  For best transport efficiency, you would need to do what
> TUF calls `ULPDU packing'.  You may end up with more than one ULPDU
> per finite interval.
> 
> It actually seems like it might be a pretty good scheme.  Your finite
> interval would be EMSS, and the segmentation algorithm would be
> similar to the current TUF except to get markers at the beginning of
> segments.
> 
> This might be a new leading candidate.  It's no touch and completely
> safe.  The cost is that you must either do ULPDU packing (the
> unpacking is actually the burdensome part, but I personally never
> thought it was too horrible), or live with inefficient transport of
> small messages.  For bulk data protocols (e.g. storage) the
> overwhelming majority of the data you carry is in large
> (interval-sized) PDUs, so a small loss of transport efficiency for the
> control PDUs won't hurt.
> 
> Alright, you've done it again Jim!  By muddying the waters we clarify
> them :-?
> 
> Steph
> 
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 15 20:17:28 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24243
	for <tsvwg-archive@odin.ietf.org>; Fri, 15 Feb 2002 20:17:28 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA28831
	for tsvwg-archive@odin.ietf.org; Fri, 15 Feb 2002 20:17:29 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA28520;
	Fri, 15 Feb 2002 20:07:50 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA28485
	for <tsvwg@optimus.ietf.org>; Fri, 15 Feb 2002 20:07:48 -0500 (EST)
Received: from sandmail.sandburst.com (sandmail.sandburst.com [216.57.132.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24147
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 20:07:46 -0500 (EST)
To: "THALER,PAT (A-Roseville,ex1)" <pat_thaler@agilent.com>
Cc: tsvwg@ietf.org
Subject: Re: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch framer for TC P 
In-Reply-To: Message from "THALER,PAT (A-Roseville,ex1)" <pat_thaler@agilent.com> 
   of "Thu, 14 Feb 2002 17:13:09 MST." <1BEBA5E8600DD4119A50009027AF54A00A512287@axcs04.cos.agilent.com> 
References: <1BEBA5E8600DD4119A50009027AF54A00A512287@axcs04.cos.agilent.com> 
Date: Fri, 15 Feb 2002 20:07:05 -0500
From: Stephen Bailey <steph@cs.uchicago.edu>
Message-Id: <20020216010716.405284EA6@sandmail.sandburst.com>
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

> I don't understand what you are suggesting. It seems like an "aligning FIM"
> approach could only work for a ULP that had fixed length PDUs - which would
> make it a pretty narrow solution.

The other thing about aligning FIM is that if the ULP is capable of
padding its PDUs, you don't actually need a marker in the stream at
all.  I'm not sure that's a useful property, but it's cute.  You know
for sure that a ULPDU begins every finite interval.

Steph


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 15 20:23:14 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24308
	for <tsvwg-archive@odin.ietf.org>; Fri, 15 Feb 2002 20:23:14 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA29003
	for tsvwg-archive@odin.ietf.org; Fri, 15 Feb 2002 20:23:15 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA28725;
	Fri, 15 Feb 2002 20:13:55 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA28696
	for <tsvwg@optimus.ietf.org>; Fri, 15 Feb 2002 20:13:53 -0500 (EST)
Received: from sandmail.sandburst.com (sandmail.sandburst.com [216.57.132.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA24198
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 20:13:51 -0500 (EST)
To: "WENDT,JIM (HP-Roseville,ex1)" <jim_wendt@hp.com>
Cc: pat_thaler@agilent.com, tsvwg@ietf.org
Subject: Re: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch f ramer for TC P 
In-Reply-To: Message from "WENDT,JIM (HP-Roseville,ex1)" <jim_wendt@hp.com> 
   of "Fri, 15 Feb 2002 17:05:08 PST." <499DC368E25AD411B3F100902740AD650AFC7DB9@xrose03.rose.hp.com> 
References: <499DC368E25AD411B3F100902740AD650AFC7DB9@xrose03.rose.hp.com> 
Date: Fri, 15 Feb 2002 20:13:10 -0500
From: Stephen Bailey <steph@cs.uchicago.edu>
Message-Id: <20020216011321.0D26D4EA6@sandmail.sandburst.com>
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

> I'm not sure what happens across PMTU changes in this scheme.

As with current TUF, the receiver stumbles along (slow path) for a
little while with uncontained ULPDUs until the sender learns of the
new EMSS, signals the new finite interval, and starts sending with the
new interval.  You'll need a synchronous handshake to do the interval
update signalling (clear the wire, signal the new finite interval,
wait for the ack and then start sending with the new finite interval).

Steph

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 15 22:49:05 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA28286
	for <tsvwg-archive@odin.ietf.org>; Fri, 15 Feb 2002 22:49:05 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id WAA04565
	for tsvwg-archive@odin.ietf.org; Fri, 15 Feb 2002 22:49:07 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA04132;
	Fri, 15 Feb 2002 22:39:27 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA04101
	for <tsvwg@optimus.ietf.org>; Fri, 15 Feb 2002 22:39:25 -0500 (EST)
Received: from lycos.com ([211.220.51.242])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id WAA27947
	for <tsvwg@ietf.org>; Fri, 15 Feb 2002 22:39:13 -0500 (EST)
Message-Id: <200202160339.WAA27947@ietf.org>
Reply-To: kjoi_dj33@lycos.com
From: ¸¶ÄÉÅÍ <kjoi_dj33@lycos.com>
To: <tsvwg@ietf.org>
Mime-Version: 1.0
Content-Type: text/html; charset="ks_c_5601-1987"
Date: Sat, 16 Feb 2002 12:39:02 +0900
X-User: 2.63-figjdliv-lilkhn-Ddigq
Subject: [Tsvwg] [±¤°í]»õ·Î¿î »ç¾÷ÀÌ³ª ºÎ¾÷À» Ã£À¸½Ê´Ï±î?
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

<HTML>
<HEAD>
<META content="text/html; charset=ks_c_5601-1987" http-equiv=Content-Type>
<STYLE> p, font, span { line-height:120%; margin-top:0; margin-bottom:0; }</STYLE>
</HEAD><BODY>
<P>¢º Çã¶ô¾øÀÌ È«º¸¸ÞÀÏÀ» º¸³»°Ô µÈ Á¡ »ç°úµå¸³´Ï´Ù. <BR>¢º Á¤º¸Åë½Å¸ÁÀÌ¿ëÃËÁø¹ý±ÔÁ¤À» ÁØ¼öÇÏ¿© ±¤°í¸ÞÀÏÀÓÀ» Ç¥½ÃÇÏ¿´À¾´Ï´Ù. <BR>¢º 
ÀüÀÚ¿ìÆíÁÖ¼Ò´Â ÀÎÅÍ³Ý»ó¿¡¼­ ÃëµæÇÏ¿´À¸¸ç, ÀüÀÚ¿ìÆíÁÖ¼Ò¿Ü ¾î¶°ÇÑ °³ÀÎ Á¤º¸µµ °¡Áö°í ÀÖÁö ¾Ê½À´Ï´Ù<BR></P>
<b><br><marquee><font color="green" size="5"> !!!! 
Áö±Ý ÀÌ¼ø°£ ´Ô²²¼­ ÀúÈñ ´ÙÀÌ³Ê½ºÆ¼¸¦ ¸¸³ª½Å °ÍÀº ´ë´ÜÇÑ Çà¿îÀÌ¶ó È®½ÅÇÕ´Ï´Ù. !!!!</font>            </marquee></FONT></b>
<P>&nbsp;</P>
<P>¹Ý°©½À´Ï´Ù. <BR>Àú´Â ºÎ»ê¿¡¼­ Á÷ÀåÀ» ´Ù´Ï¸é¼­ ºÎ¾÷À¸·Î ÀÌ »ç¾÷À» ÁøÇàÇÏ°í ÀÖ´Â Àå°­¹ÎÀÌ¶õ »ç¶÷ÀÔ´Ï´Ù. <BR>Àû±ØÀûÀÌ°í ¿­Á¤ÀûÀ¸·Î 
»ì¸é¼­ ÀÚ±â ¹ßÀüÀ» À§ÇØ ²÷ÀÓ¾øÀÌ ³ë·ÂÇÏ°í ÀÖ½À´Ï´Ù. </P>
<P>»ç¾÷Àº ¹°·Ð ¸¸³ª¼­ ´ëÈ­¸¦ ³ª´©¾î¾ß °¡Àå ÁÁ°ÚÁö¸¸, ¿ì¼± ¼­¸éÀ¸·Î ´ë·«À» ¾Ë·Áµå¸®´Ï »ì Æìº¸½Ã°í °è¼Ó °ü½ÉÀÌ ÀÖÀ¸½Ã¸é ¿¬¶ôÁÖ½Ã±â 
¹Ù¶ø´Ï´Ù.... </P>
<P>&nbsp;</P>
<P>»ç¾÷ÀÇ °³³äÀÌ ÀÌÇØÇÏ±â½±°í, ³×Æ®&#50916;À» Çü¼ºÇÏ´Â ¹æ¹ýÀÌ on-line, off-line À¸·Î ´Ù¾çÇÏ°í, µè´Â »ç¶÷µµ ³Ê¹«³ª ½±°Ô ³³µæÇÒ ¼ö 
ÀÖ±â ¶§¹®¿¡ ´©±¸³ª ÇÏ½Ç ¼ö°¡ ÀÖ½À´Ï´Ù. </P>
<P>Æ¯È÷ ±âÁ¸¿¡ ¾î¶² Çü½ÄÀÌµç ³×Æ®&#50916; ¸¶ÄÉÆÃÀ» ÇÏ°í °è½Ã´Â ºÐµéÀÌ¸é ´õ¿í Àß ÇÏ½Ç ¼ö <BR>ÀÖ½À´Ï´Ù. </P>
<P>&nbsp;</P>
<P>ÀÌ »ç¾÷Àº Åë½Å »ç¾÷ Áß °¡Àå¸ÕÀú ½ÃÀÛµÈ »õ·Î¿î »ç¾÷À¸·Î¼­ ´ÙÀÌ³Ê½ºÆ¼ ¶ó´Â È¸»ç·Î º°Á¤Åë½Å1È£¿Í 2È£ÀÇ È¸»çÀÔ´Ï´Ù.</P>
<P>ÀÚº» ¾øÀÌµµ ÀÚ±â»ç¾÷À» ÇÏ´Â°Í°ú °°À¸¸ç ½Ã°£ÀÌ Èå¸§¿¡ µû¶ó Á¡Â÷ ±× ±Ô¸ð°¡ Ä¿Á® 1³â ÈÄ È¸¿ø ¼ö¿Í ¸ÅÃâ¾×¸é¿¡¼­ ºñ±³°¡ ¾È µÉ Á¤µµ·Î 
È®ÀåµÇ´Â ±²ÀåÈ÷ ºñÁ¯ÀÌ ÀÖ´Â »ç¾÷ÀÔ´Ï´Ù. </P>
<P>&nbsp;</P>
<P>¾Æ¸¶µµ ÁÖº¯¿¡¼­ ¾Æ´ÂºÐµéÀÌ ´Ù´Ü°è »ç¾÷À» ÇÏ½Ã¸é¼­ °°ÀÌ »ç¾÷ÇØ º¸Áö ¾Ê°Ú³Ä´Â Á¦¾È ÇÑ µÎ ¹øÂë ¾È µé¾îº¸½Å ºÐµé °ÅÀÇ¾øÀ¸¸®¶ó 
»ý°¢µË´Ï´Ù.</P>
<P>±×·¸Áö¸¸ ´Ù´Ü°è°¡ ¹¼´Ï±î. ±âÁ¸ ½ÃÀåÀÇ À¯Åë±¸Á¶¸¦ ¹Ù²Ù¾î Áï Áß°£À¯Åë´Ü°è(ÃÑÆÇ, ´ë¸®Á¡) ¶Ç´Â ±¤°í¾øÀÌ ±¸Àü¿¡ ÀÇÇÑÆÇ¸Å·Î Áß°£´Ü°è¿¡¼­ 
¹ß»ýµÇ´Â ÁöÃâÀ» ÁÙ¿© ±×¼öÀÍÀ» ¼ÒºñÀÚ¿¡°Ô È¯¿ø½ÃÄÑÁÖ´Â ¼±ÁøÇü ÆÇ¸Å¹æ½ÄÀÔ´Ï´Ù.</P>
<P>³×Æ®¿÷ »ç¾÷Àº È¸¿øµéÀÇ Àç±¸¸Å°¡ ºü¸¥ Á¦Ç°À¸·Î ±¸¼ºµÇ¾î¾ß ÇÑ´Ù´Â°ÍÀº Àß¾Æ½Ã°ÚÁÒ? <BR></P>
<P>¿ì¸®°¡ ÇÏ·çµµ ÀüÈ­¸¦ °ÉÁö ¾Ê°í´Â »ì¾Æ°¥ ¼ö ¾ø´Ù´Â °ÍÀ» ´Ô²²¼­µµ ÀÎÁ¤ÇÏ½Ê´Ï±î? <BR>¹Ù·Î ÀÌ ÀüÈ­¿ä±ÝÀ» ÇöÀç ±â°£Åë½Å¿¡ ÈÄºÒ·Î 
³³ºÎÇÏ´Â°ÍÀ» ¼±ºÒ·Î Ä«µå¿¡ ÃæÀüÇØ¼­ ¾²´Â°ÍÀÌ »ç¾÷ÀÔ´Ï´Ù. </P>
<P>&nbsp;</P>
<P>2000³âµµ ±¹³» ¼Òºñ½ÃÀåÁß Åë½Å½ÃÀå±Ô¸ð´Â ÀÚ±×¸¸Ä¡ ¾à24Á¶¿øÀÌ¾úÀ¸¸ç 2001³âµµ´Â ¾ÆÁ÷ Åë°è°¡ ³ª¿ÀÁö ¾Ê¾Æ Àß¸ð¸£Áö¸¸ ÈÎ¾À ´Ã¾î³µÀ¸¸ç 
Âü°í·Î ¸»¾¸µå¸°´Ù¸é ½Ò¼Òºñ½ÃÀåÀº ¾à 1Á¶¿ø, ÀÚµ¿Â÷½ÃÀå-¾à6Á¶, À¯·ù(¼®À¯)½ÃÀå-¾à6Á¶¿øÀÌ¾ú½À´Ï´Ù. <BR>Åë½Å½ÃÀå ±Ô¸ð°¡ ¾ó¸¶³ª Å«½ÃÀåÀÎÁö 
ÁüÀÛÀÌ °¡½Ã°ÚÁÒ? </P>
<P>ÀÌ °Å´ëÅë½Å½ÃÀåÀÌ ÀÌ´ë·Î °³¹æµÈ´Ù¸é¿ì¸®³ª¶ó ½Å°æÀÎ Åë½ÅÀº ±×´ë·Î ¿Ü±¹À¯¼ö¾÷Ã¼¿¡ Àá½Ä ´çÇÒ ¼ö ¹Û¿¡ ¾ø½À´Ï´Ù.<BR></P>
<P>±×·¡¼­ Á¤ºÎ¿¡¼­ º°Á¤Åë½ÅÀ» Çã¿ëÇÏ¿© °³ÀÎÀüÈ­±¹(1È£»ç¾÷ÀÚ)À¸·Î ÇÏ¿©±Ý °³¹æÀÌÀü¿¡ ±¹Á¦°æÀï·ÂÀ» °®Ãß±â À§ÇÑ ±¹Ã¥»ç¾÷ÀÔ´Ï´Ù. </P>
<P>ÅëÇÕ¼±ºÒÄ«µå ÇÏ³ª·Î ½Ã³».¿Ü,°øÁßÀüÈ­ ÅëÈ­´Â ¹°·Ð ÈÞ´ëÆù ±âÁ¾,¹øÈ£º¯°æ¾øÀÌ 011ºÎÅÍ 019±îÁö Ãà±â´ÉÀ¸·Î ¸¹Àº ¹öÆ°À» ´©¸£Áö ¾Ê°í 
¿Â°¡Á·ÀÌ ÇÔ²² °£ÆíÇÏ°Ô »ç¿ëÇÏ½Ç¼ö ÀÖ½À´Ï´Ù.</P>
<P><BR>¹°·Ð »ó±â³»¿ëÀº ¾ÆÁÖ ±âº»ÀûÀÎ ³»¿ëÀÌ±¸¿ä.<BR>ÀÚ±â¹ØÀ¸·Î 2¸í¸¸ µÑ¼ö ÀÖ´Â ¹ÙÀÌ³Ê¸® ±¸Á¶ÀÔ´Ï´Ù</P>
<P>ÀÌ¿Ü¿¡µµ ºÎ°¡»ê¾÷µµ »Ó¸¸¾Æ´Ï¶ó&nbsp; ¼öÀÍ±¸Á¶,ÀüÀÚ»ó°Å·¡µîµî ÇÏ°í½ÍÀº ÀÌ¾ß±â´Â ¸¹À¸³ª ÀÌÁ¤µµ·Î ¼³¸íÀ» ÇØµå¸®±¸¿ä ´õ±Ã±ÀÇÏ½Å°ÍµéÀº Á÷Á¢ 
»ó´ãÀ» ÇÏ½Ã°Å³ª È¸»çÀÇ È¨ÆäÀÌÁö¸¦ µÑ·¯º¸½Ã¸é µÇ°ÚÀ¾´Ï´Ù. </P>
<P>&nbsp;</P>
<P><A href="http://www.dynastytel.co.kr">http://www.dynastytel.co.kr</A>(º°Á¤Åë½Å) 
<BR><A href="http://www.clubaqua.co.kr">http://www.clubaqua.co.kr</A>(¿©Çà»ç)<BR><A 
href="http://www.visiondynasty.com">http://www.visiondynasty.com</A>(ÀüÀÚ»ó°Å·¡)<BR><A 
href="http://www.e-dynasty.co.kr">http://www.e-dynasty.co.kr</A>(È¨ÆäÀÌÁö)<BR>¿¬¶ôÃ³ : 
<A 
href="mailto:codejang@intizen.com">codejang@intizen.com</A><BR>H.P&nbsp;&nbsp;&nbsp; 
: 019-596-9808<BR>¼ö½Å°ÅºÎ : <A 
href="mailto:e-dynasty@simmani.com">e-dynasty@simmani.com</A> ¶Ç´Â ¾Æ·¡ÀÇ ¼ö½Å°ÅºÎ ¹öÆ°À» 
´­·¯ÁÖ¼¼¿ä</P>
<P>&nbsp;</P>
<P>&nbsp;</P>
</BODY>
</HTML>
¿øÄ¡¾ÊÀ¸½Ã¸é <A href="mailto:e-dynasty@simmani.com"> ¼ö½Å°ÅºÎ </A>¸¦ ´­·¯ÁÖ½Ã°í Á¦¸ñ¿¡ ¼ö½Å°ÅºÎ¶ó°í Àû¾îÁÖ½Ã±â ¹Ù¶ø´Ï´Ù.
ºÒÆíÀ» µå·È´Ù¸é ÁË¼ÛÇÕ´Ï´Ù.

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Feb 16 10:36:52 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA13240
	for <tsvwg-archive@odin.ietf.org>; Sat, 16 Feb 2002 10:36:52 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA07392
	for tsvwg-archive@odin.ietf.org; Sat, 16 Feb 2002 10:36:55 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA06347;
	Sat, 16 Feb 2002 10:26:30 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA06315
	for <tsvwg@optimus.ietf.org>; Sat, 16 Feb 2002 10:26:28 -0500 (EST)
Received: from sj-msg-core-4.cisco.com (sj-msg-core-4.cisco.com [171.71.163.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA12957
	for <tsvwg@ietf.org>; Sat, 16 Feb 2002 10:26:22 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-4.cisco.com (8.11.3/8.9.1) with ESMTP id g1GFPtt13379;
	Sat, 16 Feb 2002 07:25:55 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABC72898;
	Sat, 16 Feb 2002 07:25:53 -0800 (PST)
Message-ID: <3C6E7A00.C8450CCE@stewart.chicago.il.us>
Date: Sat, 16 Feb 2002 09:25:53 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
CC: "'john.loughney@nokia.com'" <john.loughney@nokia.com>, mankin@isi.edu,
        tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP Add IP Extension status
References: <F92978223B01D311ACFF0008C71EE19D015DF765@MCHH225E>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Cisco will bring the kernel FreeBSD/NetBSD/OpenBSD implementation
which we will release after Connect-a-thon.

We will be able to test all proposed SCTP extensions..

R

Tuexen Michael wrote:
> 
> The AddIP extension will be tested at connectathon
> (www.connectathon.org) which will take place end of
> February. This should result in some discussions, at
> least between the participants.
> 
> This brings up another question:
> Who will participate testing SCTP and the AddIP
> draft in particular?
> 
> We will be there with our prototype implementation.
> 
> Best regards
> Michael
> 
> Michael Tuexen    Tel.:   +49 89 722 47210
> Siemens AG        Fax:    +49 89 722 48212
> ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
> 
> > -----Original Message-----
> > From: john.loughney@nokia.com [SMTP:john.loughney@nokia.com]
> > Sent: Friday, February 15, 2002 12:52 PM
> > To:   mankin@isi.edu
> > Cc:   tsvwg@ietf.org
> > Subject:      RE: [Tsvwg] SCTP Add IP Extension status
> >
> > Hi Allison,
> >
> >
> > > As I said in the meeting in SLC, it needs some reviews - it's
> > > stable but it has had no discussion.  A review of it and some
> > > discussion before WGLC would be desirable.
> >
> > Time to crank out a review, I guess ...
> >
> > John
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Feb 16 10:38:23 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA13269
	for <tsvwg-archive@odin.ietf.org>; Sat, 16 Feb 2002 10:38:23 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA07426
	for tsvwg-archive@odin.ietf.org; Sat, 16 Feb 2002 10:38:26 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA07194;
	Sat, 16 Feb 2002 10:31:46 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA07163
	for <tsvwg@optimus.ietf.org>; Sat, 16 Feb 2002 10:31:44 -0500 (EST)
Received: from sj-msg-core-4.cisco.com (sj-msg-core-4.cisco.com [171.71.163.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA13144
	for <tsvwg@ietf.org>; Sat, 16 Feb 2002 10:31:40 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-4.cisco.com (8.11.3/8.9.1) with ESMTP id g1GFVDt14452;
	Sat, 16 Feb 2002 07:31:13 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABC72925;
	Sat, 16 Feb 2002 07:31:11 -0800 (PST)
Message-ID: <3C6E7B3E.1EA2FDB0@stewart.chicago.il.us>
Date: Sat, 16 Feb 2002 09:31:10 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Jacques Rajchgod <c-Jacques.Rajchgod@wcom.com>
CC: Michael Tuexen <Michael.Tuexen@icn.siemens.de>, tsvwg@ietf.org
Subject: Re: [Tsvwg] RE: [Sigtran] Issues with SCTP TSN and Cum TSN
References: <NEBBLDAOEDIKLJPELCGEEEHPCIAA.c-Jacques.Rajchgod@wcom.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Jacques:

Right before section 5.1.1 on Page 54 (in RFC2960) we read:

"
   When the TCB is created, each endpoint MUST set its internal
   Cumulative TSN Ack Point to the value of its transmitted Initial TSN
   minus one.
"

Jacques Rajchgod wrote:
> 
> Thanks, Michael...I'll run with that.
> 
> Still, some questions:
> 
> Your first mail on the topic contained the following comment:
> 
> "...If an SCTP endpoint wants to send a SACK chunk (from JR: i.e, after an
> INIT) it must fill in the
> Cumulative_TSN_Ack field. That value is initialized to the Initial_TSN of
> the peer minus
> one..."
> 
> Where does the "minus-1" come from?
> Section "3.3.4" (SACK) 3rd parag states "Less-than or equal..."
> 
> Please kindky advise...
> 
> Jacques Rajchgod
> MCI Worldcom
> Next Generation System Evaluation
> Tel: [972] 729-7180
> E-mail: c-Jacques.Rajchgod@wcom.com
> 
> -----Original Message-----
> From: sigtran-admin@ietf.org [mailto:sigtran-admin@ietf.org]On Behalf Of
> Michael Tuexen
> Sent: Thursday, February 14, 2002 12:26 PM
> To: Jacques Rajchgod
> Cc: sigtran@ietf.org; tsvwg@ietf.org
> Subject: Re: [Sigtran] Issues with SCTP TSN and Cum TSN
> 
> Dear Jacques,
> see my comments below.
> Best regards
> Michael
> On Thursday, February 14, 2002, at 07:13 PM, Jacques Rajchgod wrote:
> 
> > OK. Here it is:
> >
> > 1) a SENDS to b:       -------------- INIT ---------------->
> > 2) a RECEIVES from b  <------------ INI_ACK ------------
> > 3) a SENDS to b       -------------- COOKIE ---------->
> > 4) a RECEIVES from b  <------------ COOKIE_ACK ------------
> > 3) a SENDS to b       --------------SCTP pack:  SACK (CUM_TSN=x) bundled with
> > ASP_UP ---------->
> >
> > - What's the "CUM_TSN=x" pointing to? Which DATA chunk is it ACK'ing
> > CUMULATIVELY??
> The value will be (Initial_TSN of the INIT_ACK chunk from b) - 1
> > - There hasn't been any DATA exchanged anywhere yet, to ack!
> It doesn't matter.
> > - A CUM_TSN is only used to ack CUMULATIVELY - DATA chunks. There are no
> > DATA chunks exchanged at INIT time. The DATA may come only after the
> > INIT.
> It is just the rule. b knows it and understands that nothing is
> acknowledged.
> - The INIT consists only of CONTROL chunks.
> It is a control chunk.
> > - So, what is a Receiver of that CUM_TSN=x ("b") supposed to look for?
> > "b"
> > hasn't sent any data to be ACK'ed cumulatively yet!
> B is looking for  (Initial_TSN of the INIT_ACK chunk from b) - 1, he
> knows the rule.
> > - Take a look at Section 1.4, page 11, the 5th bullet down - CUM_TSN...
> In general that is true. But in case of the initialization it is NOT
> true.
> That field is in each SACK chunk, so you have to put a value in it, so
> you
> have to define an initialization value.
> If you do not have anything to acknowledge you do not have to send a
> SACK. But
> if the first packets containing DATA chunks are lost and you have to
> acknowledge
> the first one you have to tell the peer that the first packets are lost.
> So you
> will also use the initialization value for the Cumulative_TSN_Ack.
> >
> > Brian recommended yesterday that the recipient of such SACK ignore that
> > SACK
> > and just update a_rwnd. Here is the statement:
> >
> > "The receiver of a SACK with Cummulative Ack of Y-1 should
> > discard the SACK and do nothing further with it other than
> > updating a_rwnd because it preceeds the Initial TSN.  This
> > is how an implementation acks nothing yet updates a_rwnd."
> >
> > Jacques Rajchgod
> > MCI Worldcom
> > Next Generation System Evaluation
> > Tel: [972] 729-7180
> > E-mail: c-Jacques.Rajchgod@wcom.com
> >
> >
> >
> > -----Original Message-----
> > From: Tuexen Michael [mailto:Michael.Tuexen@icn.siemens.de]
> > Sent: Thursday, February 14, 2002 10:35 AM
> > To: 'Jacques Rajchgod'; sigtran@ietf.org; 'tsvwg@ietf.org'
> > Subject: RE: [Sigtran] Issues with SCTP TSN and Cum TSN
> >
> >
> > Dear Jacques,
> >
> > the Cumulative_TSN_Ack is part of any SACK. I do not see
> > why this isn't good. And you must have a rule what to use
> > as an inital value because you must be able to handle the
> > case where the first packtet containing DATA chunks is
> > lost and the second arrives. You have to acknowledge them but
> > you must use the initial value as the cumulative_TSN_Ack.
> >
> > Best regards
> > Michael
> >
> > Michael Tuexen    Tel.:   +49 89 722 47210
> > Siemens AG        Fax:    +49 89 722 48212
> > ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
> >
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Feb 16 17:44:48 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA17442
	for <tsvwg-archive@odin.ietf.org>; Sat, 16 Feb 2002 17:44:48 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA22447
	for tsvwg-archive@odin.ietf.org; Sat, 16 Feb 2002 17:44:53 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA22181;
	Sat, 16 Feb 2002 17:36:18 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA22150
	for <tsvwg@optimus.ietf.org>; Sat, 16 Feb 2002 17:36:15 -0500 (EST)
Received: from hotmail.com (f71.law3.hotmail.com [209.185.241.71])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA17345
	for <tsvwg@ietf.org>; Sat, 16 Feb 2002 17:36:09 -0500 (EST)
Received: from mail pickup service by hotmail.com with Microsoft SMTPSVC;
	 Sat, 16 Feb 2002 14:35:42 -0800
Received: from 131.107.3.84 by lw3fd.law3.hotmail.msn.com with HTTP;
	Sat, 16 Feb 2002 22:35:42 GMT
X-Originating-IP: [131.107.3.84]
From: "sanjay kaniyar" <skaniyar@hotmail.com>
To: Michael.Tuexen@icn.siemens.de, ajung@exp-math.uni-essen.de,
        randall@stewart.chicago.il.us
Cc: john.loughney@nokia.com, discussion@sctp.de, thomas.hollaus@siemens.at,
        tsvwg@ietf.org
Date: Sat, 16 Feb 2002 14:35:42 -0800
Mime-Version: 1.0
Content-Type: text/plain; format=flowed
Message-ID: <F71LlBy7Elbrllb15k90001d9e1@hotmail.com>
X-OriginalArrivalTime: 16 Feb 2002 22:35:42.0797 (UTC) FILETIME=[447ACFD0:01C1B73A]
Subject: [Tsvwg] RE: Possible DoS in SCTP: How does this apply to TCP?
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Folks,

From the postings about this topic, I could gather the following about using 
IP address spoofing for DOS attack: This is not considered a major threat 
if:
(1) there is no persistent resource consumption on the victim’s machine
(2) the attacker is sacrificing (pretty much) the same amount (or order) of 
bandwidth as the victim.

Let’s consider an attacker A, victim V and a bunch of available servers on 
the internet S(i), which are available for accepting connections.

If A were to send a SYN to S(1) with source address of V, S(1) will send a 
SYN+ACK to V. Note that, S(1) will retransmit the SYN+ACK for quite a number 
of times before actually giving up (or it gets a RST from V). It may be 
possible to gag V by sending the spoofed SYN to S(1) with different source 
ports or by sending SYNs to S(2), S(3) etc. Once SYN+ACKs start getting 
queued up at the link to V, it is going to multiply rapidly because the RST 
could take longer and longer (because of duplicate SYN+ACKs). The RST 
throttling employed in some operating systems would make this worse.

Won't this constitute an attack as victim might lose 4 times the bandwidth 
(2 SYN+ACK, 2 RST) as the attacker (1 SYN)?

Thanks,
Sanjay


_________________________________________________________________
Chat with friends online, try MSN Messenger: http://messenger.msn.com


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Feb 16 19:13:12 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA18508
	for <tsvwg-archive@odin.ietf.org>; Sat, 16 Feb 2002 19:13:12 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id TAA25632
	for tsvwg-archive@odin.ietf.org; Sat, 16 Feb 2002 19:13:14 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA25241;
	Sat, 16 Feb 2002 19:01:08 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA25216
	for <tsvwg@optimus.ietf.org>; Sat, 16 Feb 2002 19:01:05 -0500 (EST)
Received: from sj-msg-core-1.cisco.com (sj-msg-core-1.cisco.com [171.71.163.11])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA18285
	for <tsvwg@ietf.org>; Sat, 16 Feb 2002 19:01:00 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-1.cisco.com (8.11.3/8.9.1) with ESMTP id g1H00Yh13515;
	Sat, 16 Feb 2002 16:00:34 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABC75409;
	Sat, 16 Feb 2002 16:00:32 -0800 (PST)
Message-ID: <3C6EF29F.C1039172@stewart.chicago.il.us>
Date: Sat, 16 Feb 2002 18:00:32 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: sanjay kaniyar <skaniyar@hotmail.com>
CC: Michael.Tuexen@icn.siemens.de, ajung@exp-math.uni-essen.de,
        john.loughney@nokia.com, thomas.hollaus@siemens.at, tsvwg@ietf.org
Subject: Re: [Tsvwg] RE: Possible DoS in SCTP: How does this apply to TCP?
References: <F71LlBy7Elbrllb15k90001d9e1@hotmail.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Sanjay:
 
I think TCP is TCP. There is nothing that will change
any behavior you note. I will make the observation though
that in SCTP's case there is NO retransmit of SYN-ACK (i.e.
the INIT-ACK in SCTP terms). Since the sender of the INIT-ACK
retains no state and depends on the sender of the INIT's timer
to kick in and send a new INIT if the INIT-ACK is lost.

So for SCTP it is truely the same expense for V as it
is for A...

R

sanjay kaniyar wrote:
> 
> Folks,
> 
> >From the postings about this topic, I could gather the following about using
> IP address spoofing for DOS attack: This is not considered a major threat
> if:
> (1) there is no persistent resource consumption on the victim?s machine
> (2) the attacker is sacrificing (pretty much) the same amount (or order) of
> bandwidth as the victim.
> 
> Let?s consider an attacker A, victim V and a bunch of available servers on
> the internet S(i), which are available for accepting connections.
> 
> If A were to send a SYN to S(1) with source address of V, S(1) will send a
> SYN+ACK to V. Note that, S(1) will retransmit the SYN+ACK for quite a number
> of times before actually giving up (or it gets a RST from V). It may be
> possible to gag V by sending the spoofed SYN to S(1) with different source
> ports or by sending SYNs to S(2), S(3) etc. Once SYN+ACKs start getting
> queued up at the link to V, it is going to multiply rapidly because the RST
> could take longer and longer (because of duplicate SYN+ACKs). The RST
> throttling employed in some operating systems would make this worse.
> 
> Won't this constitute an attack as victim might lose 4 times the bandwidth
> (2 SYN+ACK, 2 RST) as the attacker (1 SYN)?
> 
> Thanks,
> Sanjay
> 
> _________________________________________________________________
> Chat with friends online, try MSN Messenger: http://messenger.msn.com
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Feb 16 23:20:58 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA22670
	for <tsvwg-archive@odin.ietf.org>; Sat, 16 Feb 2002 23:20:58 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id XAA03868
	for tsvwg-archive@odin.ietf.org; Sat, 16 Feb 2002 23:21:02 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id XAA03697;
	Sat, 16 Feb 2002 23:12:35 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id XAA03672
	for <tsvwg@optimus.ietf.org>; Sat, 16 Feb 2002 23:12:33 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-217.dallas.net [209.44.42.217])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA22597
	for <tsvwg@ietf.org>; Sat, 16 Feb 2002 23:12:26 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id WAA09409;
	Sat, 16 Feb 2002 22:11:45 -0600
Date: Sat, 16 Feb 2002 22:11:45 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <randall@stewart.chicago.il.us>
Cc: sanjay kaniyar <skaniyar@hotmail.com>, Michael.Tuexen@icn.siemens.de,
        ajung@exp-math.uni-essen.de, john.loughney@nokia.com,
        thomas.hollaus@siemens.at, tsvwg@ietf.org
Subject: Re: [Tsvwg] RE: Possible DoS in SCTP: How does this apply to TCP?
Message-ID: <20020216221145.A5520@openss7.org>
Reply-To: bidulock@openss7.org
References: <F71LlBy7Elbrllb15k90001d9e1@hotmail.com> <3C6EF29F.C1039172@stewart.chicago.il.us>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3C6EF29F.C1039172@stewart.chicago.il.us>; from randall@stewart.chicago.il.us on Sat, Feb 16, 2002 at 06:00:32PM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Randall,

Actually I think the SCTP situation is even better than
that.  IP spoofed INITs would be a foolish way to mount
such an attack.  S(i) will at least throttle the spoofed
INITs to some degree because it takes some time to generate
the cookie.  The attacker would be better to send its own
INIT-ACKs directly to V spoofed from S(i).  That would give
more bang for the buck in the attack.

When S(i) are made to generate a state cookie, there is a
de-amplification effect.

--brian

On Sat, 16 Feb 2002, Randall Stewart wrote:

> Sanjay:
>  
> I think TCP is TCP. There is nothing that will change
> any behavior you note. I will make the observation though
> that in SCTP's case there is NO retransmit of SYN-ACK (i.e.
> the INIT-ACK in SCTP terms). Since the sender of the INIT-ACK
> retains no state and depends on the sender of the INIT's timer
> to kick in and send a new INIT if the INIT-ACK is lost.
> 
> So for SCTP it is truely the same expense for V as it
> is for A...
> 
> R
> 
> sanjay kaniyar wrote:
> > 
> > Folks,
> > 
> > >From the postings about this topic, I could gather the following about using
> > IP address spoofing for DOS attack: This is not considered a major threat
> > if:
> > (1) there is no persistent resource consumption on the victim?s machine
> > (2) the attacker is sacrificing (pretty much) the same amount (or order) of
> > bandwidth as the victim.
> > 
> > Let?s consider an attacker A, victim V and a bunch of available servers on
> > the internet S(i), which are available for accepting connections.
> > 
> > If A were to send a SYN to S(1) with source address of V, S(1) will send a
> > SYN+ACK to V. Note that, S(1) will retransmit the SYN+ACK for quite a number
> > of times before actually giving up (or it gets a RST from V). It may be
> > possible to gag V by sending the spoofed SYN to S(1) with different source
> > ports or by sending SYNs to S(2), S(3) etc. Once SYN+ACKs start getting
> > queued up at the link to V, it is going to multiply rapidly because the RST
> > could take longer and longer (because of duplicate SYN+ACKs). The RST
> > throttling employed in some operating systems would make this worse.
> > 
> > Won't this constitute an attack as victim might lose 4 times the bandwidth
> > (2 SYN+ACK, 2 RST) as the attacker (1 SYN)?
> > 
> > Thanks,
> > Sanjay
> > 
> > _________________________________________________________________
> > Chat with friends online, try MSN Messenger: http://messenger.msn.com
> > 
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 18 00:41:06 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA18900
	for <tsvwg-archive@odin.ietf.org>; Mon, 18 Feb 2002 00:41:06 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id AAA19303
	for tsvwg-archive@odin.ietf.org; Mon, 18 Feb 2002 00:41:08 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA18459;
	Mon, 18 Feb 2002 00:19:05 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id AAA18434
	for <tsvwg@optimus.ietf.org>; Mon, 18 Feb 2002 00:19:03 -0500 (EST)
Received: from boreas.isi.edu (boreas.isi.edu [128.9.160.161])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA18728
	for <tsvwg@ietf.org>; Mon, 18 Feb 2002 00:19:01 -0500 (EST)
Received: from isi.edu (c1-vpn4.isi.edu [128.9.176.34])
	by boreas.isi.edu (8.11.6/8.11.2) with ESMTP id g1I5J1Q19105;
	Sun, 17 Feb 2002 21:19:01 -0800 (PST)
Message-ID: <3C708EBA.6040504@isi.edu>
Date: Sun, 17 Feb 2002 21:18:50 -0800
From: Joe Touch <touch@ISI.EDU>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.6) Gecko/20011120
X-Accept-Language: en-us
MIME-Version: 1.0
To: Stephen Bailey <steph@cs.uchicago.edu>
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] Status of TCP ULP Framing
References: <01081411492001.07738@biene.exp-math.uni-essen.de> <200108150056.f7F0uxT24071@chmls16.mediaone.net> <01081513251602.08955@biene.exp-math.uni-essen.de> <200108160223.f7G2NIT19768@chmls16.mediaone.net> <3B875222.7D682F34@isi.edu> <200108251448.f7PEmsZ29084@chmls05.mediaone.net>  <3B87DF46.F09A5254@isi.edu> <200109031310.f83DARm10868@chmls06.mediaone.net>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Stephen Bailey wrote:

> Joe,
> 
> 
>>A protocol, in this case TCP, is more than a wire specification; 
>>it defines the behavior of the endpoints, as well as the API.
>>
>>At least two of those three are changing.
>>
> 
> To the best of our understanding, the behavior described for a
> framing-aware TCP supporting PDU alignment-mode is not a change from
> what is allowed by the TCP protocol.  In fact, if the application
> knew, or guessed at the EMSS, and tiled PDUs appropriately, some
> current TCPs would be highly likely to behave exactly like a
> framing-aware TCP.  In the cases where the two would differ, the
> framing-aware TCP behavior is a change from current implementation
> practice, but it is not a change from what is specified.


There is absolutely no correlation between ADUs and PDUs in TCP. A given 
TCP implementation is free to 'tile' as it sees fit, including overlapping.

What you really have is a way to embed a new protocol inside TCP, using 
TCP's retransmission/etc., but defining whatever new semantics you 
desire, including alignment. The protocol 'backs off' to TCP when the 
receiver verification fails. This reduces the overhead of connection 
establishment for your protocol when you succeed, at the expense of 
complicating an existing protocol for no other good reason.

Any 'escape' mechanism that attempts to circumvent this characteristic 
is, at least IMO, a new protocol. Embedding it in TCP to accelerate its 
acceptance (vs., e.g., using a new protocol ID) does not make it a 
feature of TCP.


Joe


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 18 14:53:03 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA15698
	for <tsvwg-archive@odin.ietf.org>; Mon, 18 Feb 2002 14:53:03 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA08251
	for tsvwg-archive@odin.ietf.org; Mon, 18 Feb 2002 14:53:07 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA06823;
	Mon, 18 Feb 2002 14:23:53 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA06795
	for <tsvwg@optimus.ietf.org>; Mon, 18 Feb 2002 14:23:51 -0500 (EST)
Received: from localhost ([211.109.171.65])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id OAA14604
	for <tsvwg@ietf.org>; Mon, 18 Feb 2002 14:23:46 -0500 (EST)
Message-Id: <200202181923.OAA14604@ietf.org>
Reply-To: 16ioio@hanmail.net
From: 16ioio<16ioio@hanmail.net>
To: tsvwg@ietf.org
Mime-Version: 1.0
Content-Type: text/html; charset="ks_c_5601-1987"
Date: Tue, 19 Feb 2002 04:23:59 +0900
Subject: [Tsvwg] ¿Ü·Î¿òÀ» ´Þ·¡ÁÙ ¾ÖÀÎÀÌ ÇÊ¿äÇÏ½ÅºÐ (¼ºÀÎ-±¤°í)
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

<HTML>
  <HEAD>
      <TITLE></TITLE>
  </HEAD>
  <BODY>

<IMG alt="" hspace=0 src="c:\mail\7747.gif" align=left border=0>

<pre>

<h1><font color = red>
¿Ü·Î¿òÀ» ´Þ·¡ÁÙ ¾ÖÀÎÀÌ ÇÊ¿äÇÒ¶§ 
24½Ã°£ ±â´Ù¸®°í ÀÖ¾î¿ä
               </font>
<h4>
1.  ÀüÈ­±â¸¦ µé°í
   <font color = blue size = 18>060-700-7747</font>   ¸¦ ´©¸£¼¼¿ä.
   
   (Àü±¹ Àý´ë ½Ã³»¿ä±Ý) 


<font color = blue size = 18>°¡ÀÔºñ ¾ø½¿</font>

2. ¿Ü·Î¿òÀ» ´Þ·¡ÁÙ ´ëÈ­»ó´ë°¡ ÇÊ¿äÇÏ½ÅºÐ.

3. Áö±Ý ÀüÈ­¸¦ µé°í 060-700-7747¸¦ ´­·¯ÁÖ¼¼¿ä

4. ±×´ë°¡ ¿øÇÏ´Â°ÍÀ» ÇØ°áÇÒ¼ö ÀÖ½À´Ï´Ù

5. °¡ÀÔÈÄ °ø°³°Ô½ÃÆÇ¿¡ ÀÚ½ÅÀÇ °³ÀÎÇÁ·ÎÇÊÀ» °­·ÂÇÏ°Ô ¾îÇÊÇØº¸¼¼¿ä
 

<font color = blue size = 18> ±×´ÙÀ½¿¡ ¹«½¼ÀÏÀÌ ÀÏ¾î³¯Áö Ã¥ÀÓ¸øÁ®</font>


** ÀÌÄÉÇÏ¸é ¿Àºü¾ß Ç°À¸·Î ¿­³ª¼½¾¯ ÂßÂß»§»§ ¾ð³ÄµéÀÌ È£¹Ú³ÕÄðÃ³·³ ¿ì·ç·è!! **

<h5>

º» ¸ÞÀÏÀº Á¤º¸Åë½Å¸Á ÀÌ¿ëÃËÁø¹ý ±ÔÁ¤¿¡ µû¶ó ±¤°í¸ÞÀÏÀÓÀ» Ç¥½ÃÇÏ¿´À¸¸ç ¼ö½Å°ÅºÎÀåÄ¡¸¦ ¸¶·ÃÇÏ°í ÀÖ½À´Ï´Ù. 
º» ¸ÞÀÏÁÖ¼Ò´Â ÀÎÅÍ³Ý»ó¿¡¼­ ¼öÁýÇÑ°ÍÀÌ¸ç ¸ÞÀÏÁÖ¼Ò¿Ü ¾î¶°ÇÑ °³ÀÎÁ¤º¸µµ °®°í ÀÖÁö ¾Ê½À´Ï´Ù
¿øÄ¡ ¾ÊÀº Á¤º¸¿´´Ù¸é Á¤ÁßÈ÷ »ç°ú µå¸®¸ç, ¼ö½Å°ÅºÎ¸¦ ÇØ ÁÖ½Ã¸é ´ÙÀ½ºÎÅÍ´Â ¸ÞÀÏÀÌ ¹ß¼ÛµÇÁö ¾ÊÀ» °ÍÀÔ´Ï´Ù
<a href="mailto:srej52@hanmail.net?subject=¼ö½Å°ÅºÎ">¼ö½Å°ÅºÎ</a>


  </BODY>
</HTML>

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Feb 18 17:22:59 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA20880
	for <tsvwg-archive@odin.ietf.org>; Mon, 18 Feb 2002 17:22:59 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA16003
	for tsvwg-archive@odin.ietf.org; Mon, 18 Feb 2002 17:23:02 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA14349;
	Mon, 18 Feb 2002 16:49:38 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA14320
	for <tsvwg@ns.ietf.org>; Mon, 18 Feb 2002 16:49:36 -0500 (EST)
Received: from c003.snv.cp.net (c003-h013.c003.snv.cp.net [209.228.32.227])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id QAA19750
	for <tsvwg@ietf.org>; Mon, 18 Feb 2002 16:49:28 -0500 (EST)
Received: (cpmta 24619 invoked from network); 18 Feb 2002 13:49:01 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.227) with SMTP; 18 Feb 2002 13:49:01 -0800
X-Sent: 18 Feb 2002 21:49:01 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <tsvwg@ietf.org>
Subject: RE: [Tsvwg] Status of TCP ULP Framing
Date: Mon, 18 Feb 2002 13:51:34 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJCEMICPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Importance: Normal
In-Reply-To: <3C708EBA.6040504@isi.edu>
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Joe,

Joe, I agree.  TUF, COWS, FIM, NTD, and COBS are all techniques to enable
ULP based placement within a layer beneath TCP.  The transport serves as the
interface to the application and with inclusion of this layer, additional
parameters must be exchanged.  Modifications of TCP to accommodate placement
steering must be included within drafts to allow review.  TCP specifications
are not limited to structures sent on the wire and placement to buffers
associated by PDU content does not resemble normal TCP behavior from the
perspective of the application.  Creating a generalized format for
synchronization and steering would reduce proliferation of TCP variants.
These drafts only address a desire to synchronize to ULP structures when
received out of sequence.  As a result, placement must accommodate each
application and the impact on TCP remains unrevealed.

Doug

> Stephen Bailey wrote:
>
> > Joe,
> >
> >
> >>A protocol, in this case TCP, is more than a wire specification;
> >>it defines the behavior of the endpoints, as well as the API.
> >>
> >>At least two of those three are changing.
> >>
> >
> > To the best of our understanding, the behavior described for a
> > framing-aware TCP supporting PDU alignment-mode is not a change from
> > what is allowed by the TCP protocol.  In fact, if the application
> > knew, or guessed at the EMSS, and tiled PDUs appropriately, some
> > current TCPs would be highly likely to behave exactly like a
> > framing-aware TCP.  In the cases where the two would differ, the
> > framing-aware TCP behavior is a change from current implementation
> > practice, but it is not a change from what is specified.
>
>
> There is absolutely no correlation between ADUs and PDUs in TCP. A given
> TCP implementation is free to 'tile' as it sees fit, including
> overlapping.
>
> What you really have is a way to embed a new protocol inside TCP, using
> TCP's retransmission/etc., but defining whatever new semantics you
> desire, including alignment. The protocol 'backs off' to TCP when the
> receiver verification fails. This reduces the overhead of connection
> establishment for your protocol when you succeed, at the expense of
> complicating an existing protocol for no other good reason.
>
> Any 'escape' mechanism that attempts to circumvent this characteristic
> is, at least IMO, a new protocol. Embedding it in TCP to accelerate its
> acceptance (vs., e.g., using a new protocol ID) does not make it a
> feature of TCP.
>
>
> Joe


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Feb 18 18:09:06 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA22435
	for <tsvwg-archive@odin.ietf.org>; Mon, 18 Feb 2002 18:09:06 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id SAA18696
	for tsvwg-archive@odin.ietf.org; Mon, 18 Feb 2002 18:09:10 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA17186;
	Mon, 18 Feb 2002 17:36:24 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA17158
	for <tsvwg@ns.ietf.org>; Mon, 18 Feb 2002 17:36:22 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-220.dallas.net [209.44.42.220])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA21255
	for <tsvwg@ietf.org>; Mon, 18 Feb 2002 17:36:15 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id QAA16855;
	Mon, 18 Feb 2002 16:36:05 -0600
Date: Mon, 18 Feb 2002 16:36:05 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Douglas Otis <dotis@sanlight.net>
Cc: tsvwg@ietf.org
Subject: Re: [Tsvwg] Status of TCP ULP Framing
Message-ID: <20020218163605.C15127@openss7.org>
Reply-To: bidulock@openss7.org
References: <3C708EBA.6040504@isi.edu> <NEBBJGDMMLHHCIKHGBEJCEMICPAA.dotis@sanlight.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <NEBBJGDMMLHHCIKHGBEJCEMICPAA.dotis@sanlight.net>; from dotis@sanlight.net on Mon, Feb 18, 2002 at 01:51:34PM -0800
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Douglas,

With the number of vendors claiming commercial deployments of SCTP
(and SIGTRAN UAs for that matter) I find it hard to believe that
any of these steps are required as intermediary steps to just using
SCTP.  Will TSV next try to shoe-horn multi-homing into TCP?

--brian

On Mon, 18 Feb 2002, Douglas Otis wrote:

> Joe,
> 
> Joe, I agree.  TUF, COWS, FIM, NTD, and COBS are all techniques to enable
> ULP based placement within a layer beneath TCP.  The transport serves as the
> interface to the application and with inclusion of this layer, additional
> parameters must be exchanged.  Modifications of TCP to accommodate placement
> steering must be included within drafts to allow review.  TCP specifications
> are not limited to structures sent on the wire and placement to buffers
> associated by PDU content does not resemble normal TCP behavior from the
> perspective of the application.  Creating a generalized format for
> synchronization and steering would reduce proliferation of TCP variants.
> These drafts only address a desire to synchronize to ULP structures when
> received out of sequence.  As a result, placement must accommodate each
> application and the impact on TCP remains unrevealed.
> 
> Doug
> 
> > Stephen Bailey wrote:
> >
> > > Joe,
> > >
> > >
> > >>A protocol, in this case TCP, is more than a wire specification;
> > >>it defines the behavior of the endpoints, as well as the API.
> > >>
> > >>At least two of those three are changing.
> > >>
> > >
> > > To the best of our understanding, the behavior described for a
> > > framing-aware TCP supporting PDU alignment-mode is not a change from
> > > what is allowed by the TCP protocol.  In fact, if the application
> > > knew, or guessed at the EMSS, and tiled PDUs appropriately, some
> > > current TCPs would be highly likely to behave exactly like a
> > > framing-aware TCP.  In the cases where the two would differ, the
> > > framing-aware TCP behavior is a change from current implementation
> > > practice, but it is not a change from what is specified.
> >
> >
> > There is absolutely no correlation between ADUs and PDUs in TCP. A given
> > TCP implementation is free to 'tile' as it sees fit, including
> > overlapping.
> >
> > What you really have is a way to embed a new protocol inside TCP, using
> > TCP's retransmission/etc., but defining whatever new semantics you
> > desire, including alignment. The protocol 'backs off' to TCP when the
> > receiver verification fails. This reduces the overhead of connection
> > establishment for your protocol when you succeed, at the expense of
> > complicating an existing protocol for no other good reason.
> >
> > Any 'escape' mechanism that attempts to circumvent this characteristic
> > is, at least IMO, a new protocol. Embedding it in TCP to accelerate its
> > acceptance (vs., e.g., using a new protocol ID) does not make it a
> > feature of TCP.
> >
> >
> > Joe
> 
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Feb 18 20:25:46 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA25920
	for <tsvwg-archive@odin.ietf.org>; Mon, 18 Feb 2002 20:25:46 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA24818
	for tsvwg-archive@odin.ietf.org; Mon, 18 Feb 2002 20:25:50 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA23468;
	Mon, 18 Feb 2002 19:53:09 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA23439
	for <tsvwg@ns.ietf.org>; Mon, 18 Feb 2002 19:53:07 -0500 (EST)
Received: from motgate3.mot.com (motgate3.mot.com [144.189.100.103])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA25213
	for <tsvwg@ietf.org>; Mon, 18 Feb 2002 19:53:02 -0500 (EST)
Received: [from pobox2.mot.com (pobox2.mot.com [136.182.15.8]) by motgate3.mot.com (motgate3 2.1) with ESMTP id RAA21871; Mon, 18 Feb 2002 17:42:18 -0700 (MST)]
Received: [from relay2.cig.mot.com (relay2.cig.mot.com [136.182.15.24]) by pobox2.mot.com (MOT-pobox2 2.0) with ESMTP id RAA17414; Mon, 18 Feb 2002 17:53:01 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay2.cig.mot.com (8.11.4/8.11.4) with ESMTP id g1J0qvr29380;
	Mon, 18 Feb 2002 18:52:57 -0600 (CST)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id SAA13403;
	Mon, 18 Feb 2002 18:57:52 -0600 (CST)
Message-ID: <3C71A29F.9A407A73@email.mot.com>
Date: Mon, 18 Feb 2002 18:55:59 -0600
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: TSV WG list <tsvwg@ietf.org>, bdp003@motorola.com,
        Senaka.Balasuriya@motorola.com, yoonie@verbaltek.com, smaes@us.ibm.com,
        hgarudad@qualcomm.com
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] DSR draft revised
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi, all,

The revised DSR over RTP draft has been submitted. The new version (v01)
incorporates all the feedback and comments made so far during the WG
last call.

There is _no_ technical change made in this revision. However, thanks to
the careful review and commenting by John Lazzaro, Priscilla Walther,
and others, quite some editorial changes are made and I hope these
changes will greatly improve the readability of the document.

Here is a list of some of the editorial changes made:

1. Added a list of acronyms to improve readability.

2. The payload diagram is made more clear by showing the RTP header.

3. More sub-section headers are added to make the discussion more
structured.

4. A new section is added to explain the usage of RTP header fields, per
John's suggestion.

5. Missing text for "Encoding considerations" is added in section 6.

6. References are grouped into normative and informative.

7. A reference to ROHC hc is added as informative.

The new version should appear on the IETF web-site soon.

regards,
-Qiaobing Xie
Motorola

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Mon Feb 18 20:32:03 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA26064
	for <tsvwg-archive@odin.ietf.org>; Mon, 18 Feb 2002 20:32:03 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA25396
	for tsvwg-archive@odin.ietf.org; Mon, 18 Feb 2002 20:32:07 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA23993;
	Mon, 18 Feb 2002 20:00:55 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA23956
	for <tsvwg@ns.ietf.org>; Mon, 18 Feb 2002 20:00:52 -0500 (EST)
Received: from motgate2.mot.com (motgate2.mot.com [136.182.1.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA25352
	for <tsvwg@ietf.org>; Mon, 18 Feb 2002 20:00:48 -0500 (EST)
Received: [from mothost.mot.com (mothost.mot.com [129.188.137.101]) by motgate2.mot.com (motgate2 2.1) with ESMTP id SAA15955 for <tsvwg@ietf.org>; Mon, 18 Feb 2002 18:00:51 -0700 (MST)]
Received: [from relay2.cig.mot.com (relay2.cig.mot.com [136.182.15.24]) by mothost.mot.com (MOT-mothost 2.0) with ESMTP id SAA27080 for <tsvwg@ietf.org>; Mon, 18 Feb 2002 18:00:50 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay2.cig.mot.com (8.11.4/8.11.4) with ESMTP id g1J10mr29880;
	Mon, 18 Feb 2002 19:00:48 -0600 (CST)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id TAA13415;
	Mon, 18 Feb 2002 19:05:43 -0600 (CST)
Message-ID: <3C71A477.49D4DE55@email.mot.com>
Date: Mon, 18 Feb 2002 19:03:51 -0600
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: TSV WG list <tsvwg@ietf.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] Recall: DSR draft revised
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Sorry, my previous email with the above subject line was sent to the
wrong list. My apology.

-Qiaobing

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 19 09:46:42 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA21385
	for <tsvwg-archive@odin.ietf.org>; Tue, 19 Feb 2002 09:46:42 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA09643
	for tsvwg-archive@odin.ietf.org; Tue, 19 Feb 2002 09:46:44 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA07417;
	Tue, 19 Feb 2002 09:20:21 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA07385
	for <tsvwg@optimus.ietf.org>; Tue, 19 Feb 2002 09:20:19 -0500 (EST)
Received: from localhost ([211.50.255.50])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id JAA20136
	for <tsvwg@ietf.org>; Tue, 19 Feb 2002 09:20:14 -0500 (EST)
Message-Id: <200202191420.JAA20136@ietf.org>
Reply-To: pricelist@hutmail.com
From: ÄÄÁ¤º¸<pricelist@hutmail.com>
To: tsvwg@ietf.org
Mime-Version: 1.0
Content-Type: text/html; charset="ks_c_5601-1987"
Date: Tue, 19 Feb 2002 23:20:08 +0900
Subject: [Tsvwg] {±¤ °í} ÄÄÇ»ÅÍ°¡°ÝÁ¤º¸ ¸ÞÀÏ¸µ È¸¿øÀÌ µÇ¼¼¿ä (^-^)~~
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

<html>
<head>
<title>ÄÄÇ»ÅÍ °¡°ÝÁ¤º¸ ¸ÞÀÏ¸µ ½ÅÃ» ¹Þ½À´Ï´Ù.</title>
</head>
<BODY text=black vLink=red aLink=red link=red bgColor=white>
<P align=center><FONT size=2>¸ÕÀú Çã¶ô¾øÀÌ ±ÛÀ» ¿Ã¸°°Í¿¡ ´ëÇØ Áø½ÉÀ¸·Î »ç°ú¸»¾¸ µå¸³´Ï´Ù.<BR>±ÍÇÏÀÇ ÀÌ¸ÞÀÏ ÁÖ¼Ò´Â
ÀÎÅÍ³Ý °Ô½ÃÆÇÀ» ÅëÇÏ¿© ÃëÇÕÇÏ¿´½À´Ï´Ù<BR></FONT></P>
<P align=center><FONT size=2><B>ÄÄÇ»ÅÍ °¡°ÝÁ¤º¸</B>¸¦ <BR>¸ÅÀÏ ¸ÅÀÏ È¸¿ø´ÔÀÇ ÀÌ¸ÞÀÏ·Î ¹ß¼ÛÇØ µå¸®´Â
¼­ºñ½ºÀÔ´Ï´Ù.<BR>Ç×»ó ÃÖ½ÅÀÇ °¡°ÝÁ¤º¸°¡ È¸¿ø´ÔÀÇ ÀÌ¸ÞÀÏ¼Ó¿¡ ´ã°ÜÁ® ÀÖ°Ô ÇØµå¸®°Ú½À´Ï´Ù.<BR><BR>ÇöÀç ÀÏ°£(È­,¸ñ,Åä¿äÀÏ ¹ßÇà) ¹×
ÁÖ°£(¸ñ¿äÀÏ ¹ßÇà) ÇüÅÂ·Î ¹ßÇàÁßÀÔ´Ï´Ù.<BR>ÀÏ¿äÀÏ ¹× °øÈÞÀÏÀº ¹ßÇàÀ» ÇÏÁö ¾Ê½À´Ï´Ù.<BR><BR>¹°·Ð º» ¼­ºñ½º´Â <FONT
color=red>¹«·á</FONT>ÀÔ´Ï´Ù.</P>
<P align=center><FONT size=2>±¸µ¶½ÅÃ»Àº ¸ÞÀÏÁÖ¼Ò¸¸ ½áÁÖ½Ã¸é ³¡ÀÔ´Ï´Ù.. °£´ÜÇÏÁÒ..^^</FONT></P><!-- form_begin -->
<TABLE height=80 cellSpacing=0 cellPadding=0 width=240 align=center
bgColor=orange border=0>
<FORM action=http://www.emag21.com/subscriber/skin_register.php method=post
target=_new><INPUT type=hidden value=20000501000426 name=magid> <INPUT
type=hidden value=insert name=mod>
<TBODY>
<TR>
<TD width=213 height=20>
<P align=center><B><FONT size=2>ÄÄÇ»ÅÍ °¡°ÝÁ¤º¸ [ÀÏ°£] </FONT></B></P></TD>
<TD vAlign=top align=right width=27 height=27 rowSpan=2><FONT size=2><IMG
src="http://www.infomail.co.kr/readform/img/f01_fold.gif"></FONT></TD></TR>
<TR>
<TD width=213 background=http://www.infomail.co.kr/readform/img/f01_bg.gif
height=7><FONT size=2><IMG height=1
src="http://www.infomail.co.kr/readform/img/dot.gif" width=1></FONT></TD></TR>
<TR>
<TD background=http://www.infomail.co.kr/readform/img/f01_bg.gif
colSpan=2><TABLE cellSpacing=2 cellPadding=0 width=220 align=center
border=0>
<TBODY>
<TR>
<TD><FONT size=2><IMG
src="http://www.infomail.co.kr/readform/img/t_email.gif"></FONT></TD>
<TD><FONT size=2><INPUT
style="BORDER-RIGHT: 1px dashed; BORDER-TOP: 1px dashed; BORDER-LEFT: 1px dashed; BORDER-BOTTOM: 1px dashed; HEIGHT: 13pt"
tabIndex=1 maxLength=30 size=14 value=tsvwg@ietf.org name=email>
</FONT></TD>
<TD vAlign=bottom><FONT size=2><INPUT type=image
src="http://www.orange1004.com/mag/0111/daily.gif"
valign="bottom"></FONT></TD></TR>
<TR>
<TD></TD>
<TD background=http://www.infomail.co.kr/readform/img/f01_bg.gif
colSpan=2></TD></TR></TBODY></TABLE></TD></TR>
<TR>
<TD colSpan=2 height=2><IMG height=1
src="http://www.infomail.co.kr/readform/img/dot.gif"
width=1></TD></TR></FORM></TBODY></TABLE><BR><BR>
<TABLE height=80 cellSpacing=0 cellPadding=0 width=240 align=center
bgColor=lightblue border=0>
<FORM action=http://www.emag21.com/subscriber/skin_register.php method=post
target=_new><INPUT type=hidden value=20000529001185 name=magid> <INPUT
type=hidden value=insert name=mod>
<TBODY>
<TR>
<TD width=213 height=20>
<P align=center><B><FONT size=2>ÄÄÇ»ÅÍ °¡°ÝÁ¤º¸ [ÁÖ°£] </FONT></B></P></TD>
<TD vAlign=top align=right width=27 height=27 rowSpan=2><FONT size=2><IMG
src="http://www.infomail.co.kr/readform/img/f01_fold.gif"></FONT></TD></TR>
<TR>
<TD width=213 background=http://www.infomail.co.kr/readform/img/f01_bg.gif
height=7><FONT size=2><IMG height=1
src="http://www.infomail.co.kr/readform/img/dot.gif" width=1></FONT></TD></TR>
<TR>
<TD background=http://www.infomail.co.kr/readform/img/f01_bg.gif
colSpan=2><TABLE cellSpacing=2 cellPadding=0 width=220 align=center
border=0>
<TBODY>
<TR>
<TD><FONT size=2><IMG
src="http://www.infomail.co.kr/readform/img/t_email.gif"></FONT></TD>
<TD><FONT size=2><INPUT
style="BORDER-RIGHT: 1px dashed; BORDER-TOP: 1px dashed; BORDER-LEFT: 1px dashed; BORDER-BOTTOM: 1px dashed; HEIGHT: 13pt"
tabIndex=1 maxLength=30 size=14 value=tsvwg@ietf.org name=email>
</FONT></TD>
<TD vAlign=bottom><FONT size=2><INPUT type=image
src="http://www.orange1004.com/mag/0111/weekly.gif"
valign="bottom"></FONT></TD></TR>
<TR>
<TD></TD>
<TD background=http://www.infomail.co.kr/readform/img/f01_bg.gif
colSpan=2></TD></TR></TBODY></TABLE></TD></TR>
<TR>
<TD colSpan=2 height=2><IMG height=1
src="http://www.infomail.co.kr/readform/img/dot.gif"
width=1></TD></TR></FORM></TBODY></TABLE><BR>
<TABLE height=80 cellSpacing=0 cellPadding=0 width=240 align=center
bgColor=#c4ea02 border=0>
<FORM action=http://www.infomail.co.kr/cgi-bin/fan_rreg method=post
target=fan_rreg><INPUT type=hidden value=reg name=mode> <INPUT type=hidden
value=20817 name=author_id>
<TBODY>
<TR>
<TD align=right height=27><FONT size=2>ÇÑ¸ÞÀÏ »ç¿ëÀÚ¿ë [ÀÏ°£]</FONT></TD>
<TD align=right><IMG
src="http://www.infomail.co.kr/readform/img/f01_fold.gif"></TD></TR>
<TR>
<TD colSpan=2>
<TABLE cellSpacing=2 cellPadding=0 width=220 align=center border=0>
<TBODY>
<TR>
<TD><IMG
src="http://www.infomail.co.kr/readform/img/t_email.gif"></TD>
<TD><INPUT
style="BORDER-RIGHT: 1px dashed; BORDER-TOP: 1px dashed; BORDER-LEFT: 1px dashed; BORDER-BOTTOM: 1px dashed; HEIGHT: 13pt"
tabIndex=1 maxLength=30 size=12 value=tsvwg@ietf.org name=email></TD>
<TD vAlign=bottom><INPUT type=image
src="http://www.infomail.co.kr/readform/img/f01_btn_sr.gif"></TD></TR>
<TR>
<TD>&nbsp;</TD>
<TD colSpan=2><IMG
src="http://www.infomail.co.kr/readform/img/f01_arrow.gif" border=0>
<FONT
size=2>ÄÄÇ»ÅÍ°¡°ÝÁ¤º¸</FONT></TD></TR></TBODY></TABLE></TD></TR></TBODY></TABLE><BR>
<TABLE height=80 cellSpacing=0 cellPadding=0 width=240 align=center
bgColor=#c4ea02 border=0><FORM method="post"
action="http://www.infomail.co.kr/cgi-bin/fan_rreg" target="fan_rreg"><INPUT
type=hidden value=reg name=mode> <INPUT type=hidden value=22413
name=author_id>
<TBODY>
<TR>
<TD align=right height=27><FONT size=2>ÇÑ¸ÞÀÏ »ç¿ëÀÚ¿ë [ÁÖ°£]</FONT></TD>
<TD align=right><IMG
src="http://www.infomail.co.kr/readform/img/f01_fold.gif"></TD></TR>
<TR>
<TD colSpan=2>
<TABLE cellSpacing=2 cellPadding=0 width=220 align=center border=0>
<TBODY>
<TR>
<TD><IMG
src="http://www.infomail.co.kr/readform/img/t_email.gif"></TD>
<TD><INPUT
style="BORDER-RIGHT: 1px dashed; BORDER-TOP: 1px dashed; BORDER-LEFT: 1px dashed; BORDER-BOTTOM: 1px dashed; HEIGHT: 13pt"
tabIndex=1 maxLength=30 size=12 value=tsvwg@ietf.org name=email></TD>
<TD vAlign=bottom><INPUT type=image
src="http://www.infomail.co.kr/readform/img/f01_btn_sr.gif"></TD></TR>
<TR>
<TD>&nbsp;</TD>
<TD colSpan=2><IMG
src="http://www.infomail.co.kr/readform/img/f01_arrow.gif" border=0>
<FONT size=2>ÄÄÇ»ÅÍ°¡°ÝÁ¤º¸</FONT></TD></TR></TBODY></TABLE></TD></TR></TBODY></TABLE>
<P align=center><B><FONT color=red><A
href="http://www.emag21.com/home/?url=./home.html&amp;magid=20000501000426&amp;site=A"
target=_blank>Áö³­¸ÞÀÏ¸µº¸±â</A> ¡Ú</FONT></B></P>
<P align=center><FONT color=gray size=2>º» ¸ÞÀÏ·Î ±âºÐ »óÇÏ¼Ì´Ù¸é ±íÀÌ»ç°ú µå¸³´Ï´Ù.!<BR><A href="mailto:comprice@lycos.co.kr?subject=¼ö½Å°ÅºÎ ¸ÞÀÏ&amp;body=°¡°ÝÁ¤º¸ ¸ÞÀÏ¸µ ±¤°í ¼ö½Å°ÅºÎ ÇÕ´Ï´Ù."><FONT
color=blue>¼ö½Å°ÅºÎ</FONT></A> ÇØÁÖ½Ã¸é ¹Ýº¹¹ß¼ÛÇÏÁö ¾Êµµ·Ï ÇÏ°Ú½À´Ï´Ù.<BR>º»¸ÞÀÏÀº <FONT
color=#ff8000>1È¸¼º</FONT> ¸ÞÀÏÀÔ´Ï´Ù.</FONT></P>
<P align=center><A href="mailto:comprice@lycos.co.kr?subject=¼ö½Å°ÅºÎ ¸ÞÀÏ&amp;body=°¡°ÝÁ¤º¸ ¸ÞÀÏ¸µ ±¤°í ¼ö½Å°ÅºÎ ÇÕ´Ï´Ù."><FONT
color=blue>¼ö½Å°ÅºÎ</FONT></A>&nbsp;&lt;== ¼ö½Å°ÅºÎ´Â ²À ¿ä±â¸¦ ´­·¯ÁÖ¼¼¿ä.. °¨»çÇÕ´Ï´Ù.</P><!-- form_end --></FORM></FONT>
</BODY>
</html>

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 19 18:15:02 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA08168
	for <tsvwg-archive@odin.ietf.org>; Tue, 19 Feb 2002 18:15:02 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id SAA09150
	for tsvwg-archive@odin.ietf.org; Tue, 19 Feb 2002 18:15:03 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA07577;
	Tue, 19 Feb 2002 17:31:46 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA07553
	for <tsvwg@optimus.ietf.org>; Tue, 19 Feb 2002 17:31:44 -0500 (EST)
Received: from localhost ([211.245.86.158])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id RAA07381
	for <tsvwg@ietf.org>; Tue, 19 Feb 2002 17:31:39 -0500 (EST)
Message-Id: <200202192231.RAA07381@ietf.org>
Reply-To: fihterbird@yahoo.co.kr
From: fighterbird<fighterbird@yahoo.co.kr>
To: tsvwg@ietf.org
Mime-Version: 1.0
Content-Type: text/html; charset="ks_c_5601-1987"
Date: Wed, 20 Feb 2002 07:31:11 +0900
Subject: [Tsvwg] ´ç½ÅÀÌ Áö±Ý Á¤¸» ¿øÇÏ´Â °ÍÀº ¹«¾ùÀÔ´Ï±î?
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

<html> 
<BODY style="FONT-SIZE: 10pt; FONT-FAMILY: µ¸¿ò" background=white> 
<P><IMG src="http://www.nutritionforlife.com/test/country_banner.gif"></P> 
<P><FONT color=black size=3><B>º» ¸ÞÀÏÀº Á¤º¸Åë½Å¸Á ÀÌ¿ëÃËÁø ¹× Á¤º¸º¸È£ µî¿¡ °üÇÑ ¹ý·ü Á¦ 50Á¶¿¡ ÀÇ°ÅÇÑ <FONT 
color=blue size=3>[±¤°í]</FONT> ¸ÞÀÏÀÔ´Ï´Ù. <BR>ÀüÀÚ¿ìÆí ÁÖ¼Ò ¿Ü ¾î¶°ÇÑ °³ÀÎ Á¤º¸µµ °¡Áö°í ÀÖÁö ¾ÊÀ¸¹Ç·Î ¾È½ÉÇÏ½Ã±â 
¹Ù¶ø´Ï´Ù.<BR>ÀÌ¸ÞÀÏ ¼ö½ÅÀ» ¿øÄ¡ ¾ÊÀ¸½Ã¸é ¾Æ·¡ÁÖ¼Ò·Î <FONT color=blue size=3>[¼ö½Å°ÅºÎ]</FONT></A> ¸¦ÇÏ¿© 
ÁÖ½Ê½Ã¿À. ÀúÈñ´Â ±ÍÇÏÀÇ ÀÇ»ç¸¦ Á¸ÁßÇÕ´Ï´Ù.<BR>¼ö½Å°ÅºÎµÈ ¸ÞÀÏÀº Àý´ë·Î ¸ÞÀÏÀ» Àç¹ß¼ÛÇÏÁö ¾Ê°Ú½À´Ï´Ù.<BR>¼ö½Å°ÅºÎ´Â <FONT 
color=red size=3><A href="mailto:l99922@hanmail.net">l99922@hanmail.net</A></FONT>À¸·Î 
ÇØÁÖ½Ê½Ã¿ä.<BR><BR><BR>»õ·Î¿î ºÎ¾÷À» ¿øÇÏ½Ê´Ï±î?<BR><BR>Á÷Àå»ýÈ°À» ÇÏ¸é¼­ °¡Á¤ÀÏÀ» ÇÏ¸é¼­ ´ÔÀÌ ¾ó¸¶µçÁö<BR><BR>»ýÈ°À» 
¿©À¯·Ó°Ô ¿µÀ§ÇÒ¼ö ÀÖ½À´Ï´Ù.<BR><BR>´ÔÀÌ ¸¹Àº °æÇèÀÌ ÀÖÀ» ÇÊ¿äµµ ¾ø½À´Ï´Ù.<BR><BR>¹°°ÇÀ» ÆÈ·Á°í ¾îµô ´Ù´ÒÇÊ¿äµµ 
¾ø½À´Ï´Ù.<BR><BR>¾ÆÄ§Àú³á ÃâÅð±Ù ÇÒÇÊ¿äµµ ¾ø½À´Ï´Ù.<BR><BR>»ýÈ°ÀÇ ¸¹Àº ½Ã°£À» ÇÒ¾ÖÇÒ ÇÊ¿äµµ ¾ø½À´Ï´Ù.<BR><BR>Á¶±Ý¸¸ °ü½ÉÀ» 
°¡Áö¸é ½ÃÀÛÇÒ¼öÀÖ½À´Ï´Ù.<BR><BR>´ÔÀÌ °ü½ÉÀÌ ÀÖ´Ù¸é Àú¿¡°Ô ¿¬¶ôÀ» ÁÖ½Ê½Ã¿ä.<BR><BR>´Ô¿¡°Ô È®½ÅÀ» ÁÖ±âÀü¿¡´Â Àý´ë ´Ô¿¡°Ô ±ÇÇÏÁö 
¾Ê°Ú½À´Ï´Ù.<BR><BR>³×Æ®¿÷È¸»çÀÇ »ç¾÷ÀÇ ¿ä±¸Á¶°ÇÀº ±×È¸»çÀÇ Á¦Ç°°ú ±× ¶óÀÎÇü¼º¿¡ ÀÖ´Â°ÍÀÔ´Ï´Ù. </B></FONT></P> 
<P><FONT color=black size=3><B>±×µÎ°¡Áö Á¶°Ç¿¡ ¸ðµÎ ºÎÇÕµÇ´Â°ÍÀÌ <FONT color=red 
size=3>NFLI</FONT>¶ó ÇÒ¼öÀÖ½À´Ï´Ù.</B></FONT></P> 
<P><FONT color=black size=3><B>±×ÀÌÀ¯´Â ¹Ì±¹¿¡¼­ 20³â°¡±îÀÌ ÅºÅºÇÑ ±â¹ÝÀ» ÀÌ¹Ì ±¸ÃàÇÏ¿´°í</B></FONT></P> 
<P><FONT color=black size=3><B>Á¦Ç°¶ÇÇÑ Àç±¸¸ÅÀ² 80%ÀÌ»óÀ¸·Î ¿ì¿ù¼ºÀ» ÀÔÁõ¹ÞÀº¹Ù ÀÖ½À´Ï´Ù.</B></FONT></P> 
<P><FONT color=black size=3><B>500¿©°¡ÁöÀÇ °¡Á¤¿ëÇ°¿¡¼­ °Ç°­½ÄÇ°,»ýÇÊÇ°µîÀÇ ¼ÒºñÀÚÀÇ</B></FONT></P> 
<P><FONT color=black size=3><B>±¸¸ÅÁ¶°Ç¿¡ ºÎÇÕÇÏ´Â ´Ù¾çÇÑ Á¦Ç°±ºÀ¸·Î ±¸¼ºµÇ¾î ÀÖ½À´Ï´Ù.</B></FONT></P> 
<P><FONT color=black size=3><B>±×¸®°í ´õ¿íÀÌ ³×Æ®¿÷»ç¾÷ÀÇ Ã¹Â° ¼º°øÁ¶°ÇÀÎ ÃÊ±â Çìµå¶óÀÎÀÇ 
À§Ä¡¸¦</B></FONT></P> 
<P><FONT color=black size=3><B>È®º¸ÇÒ¼ö ÀÖ´ÂÁ¡ ¶ÇÇÑ ±²ÀåÇÑ ¸Å·ÂÀÌ°ÚÁö¿ä..</B></FONT></P> 
<P><FONT color=black size=3><B>´ÔÀÌ ºÎ¾÷À» »ý°¢ÇÏ°í °è½Å´Ù¸é ÀÌ¸¸ÇÑ ¸Å·ÂÀûÀÎÀÏÀº ¾Æ¸¶µµ 
¾øÀ»°ÍÀÔ´Ï´Ù.</B></FONT></P> 
<P><FONT color=black size=3><B>º¸»óÇÃ·£¿¡¼­ »ìÆìº¸¸é 4¸íÀÇ ¸ÞÆ®¸¯½º ¹æ½ÄÀ¸·Î ÇÏÀ§¶óÀÎ¿¡</B></FONT></P> 
<P><FONT color=black size=3><B>ÀÚµ¿¹èÄ¡µÇ´Â ¹æ½ÄÀ» ÅÃÇÏ°í ÀÖ½À´Ï´Ù.</B></FONT></P> 
<P><FONT color=black size=3><B>´ÔÀÌ ¾î´ÀÁ¤µµÀÇ ¶óÀÎ¸¸ ±¸ÃàÇÒ ´É·ÂÀÌ µÇ½Å´Ù¸é ´ÔÀÇ Æò»ýµ¿¾ÈÀÇ</B></FONT></P> 
<P><FONT color=black size=3><B>¼öÀÔ¿øÀ¸·Î¼­ ´ÔÀÇ»ýÈ°À» À±ÅÃÇÏ°Ô ¹Ù²Ù¾î µå¸±°ÍÀÔ´Ï´Ù.</B></FONT></P> 
<P><FONT color=black size=3><B>Àú¸¦ ºñ·ÔÇÑ ÀúÈñ Å¬·´¿¡¼­´Â ´ÔÀÇ ¼º°øÀ» À§ÇØ ¸ðµç ³ë·ÂÀ» ¾Æ³¢Áö 
¾ÊÀ»°ÍÀÔ´Ï´Ù.</B></FONT></P> 
<P><FONT color=black size=3><B>´Ô¿¡°Ô ¿ä±¸µÇ´Â°ÍÀº ÀÌ»ç¾÷¿¡ÀÖ¾î ÃÖ¼ÒÇÑÀÇ °ü½É»ÓÀÔ´Ï´Ù.</B></FONT></P> 
<P><FONT color=black size=3><B>½ÃÀÛÇÏ½Ê½Ã¿ä Á¦°¡ ÃÖ¼±ÀÇ ³ë·ÂÀ¸·Î ´ÔÀ» Áö¿øÇØ µå¸®°Ú½À´Ï´Ù.</B></FONT></P> 
<P><FONT color=black size=3><B>´ÔÀÌ ¼º°øÇÒ¶§±îÁö ¿ä±¸µÇ´Â ¸ðµç ³ëÇÏ¿ì¸¦ Áö¿øÇØ µå¸®°Ú½À´Ï´Ù. 
<BR></B></FONT><FONT color=black size=3><B><BR><BR>°¡»ó¿Â¶óÀÎ µî·ÏÇÏ½Ã°í ´Ù¿î¶óÀÎ Çü¼ºµÇ´Â°Í 
Á÷Á¢<BR><BR>´Ù¿î¶óÀÎ Á¶È¸¿¡¼­ È®ÀÎÇÏ¼¼¿ä...<BR><BR>´ç½Å°ú ´ç½ÅÀÌ È«º¸ÇÏ¿© Çü¼ºµÇ´Â ÆÄÆ®³Ê±îÁö<BR><BR>¸ðµÎ È¨ÆäÀÌÁö¸¦ 
½Ç½Ã°£À¸·Î Áö¿øÇÕ´Ï´Ù.<BR><BR><FONT color=blue 
size=3><A target=_blank href="http://ebiz7.net/myteam/index.php?id=l99922">http://ebiz7.net/myteam/index.php?id=l99922</a></FONT><BR><BR>º»ÀÎÀÇ ¾ÆÀÌµð´Â ¾Æ·¡¿¡¼­ 
°¡»ó¸â¹ö µî·ÏÀÇ ¾ÆÀÌµð¸¦ ¸»ÇÕ´Ï´Ù.<BR><BR>º» Å¬·´¿¡¼­´Â Àý´ë ´ÔÀÇ ÀÇ»ç¸¦ Á¸ÁßÇÕ´Ï´Ù.<BR><BR>º» Å¬·´À» ¼º°øÀÇ ¹ßÆÇÀ¸·Î »ïÀ¸½Ã±æ 
¹Ù¶ø´Ï´Ù.<BR><BR>¾Æ·¡ÀÇ È¨ÆäÀÌÁö¸¦ Å¬¸¯ÇÏ¿© °¡ÀÔÇÏ½ÅÈÄ<BR><BR>¿¬¶ôÀ» ÁÖ½Ê½Ã¿ä.<BR><BR>±×·³ ´õÀÚ¼¼ÇÑ ¼³¸íÀ» µéÀ¸½Ã·Á¸é Àú¿¡°Ô 
<BR><BR>¹®ÀÇ¸¦ ÇÏ½Ê½Ã¿ä.<BR><BR><FONT color=red size=3><A href="mailto:l99922@hanmail.net">l99922@hanmail.net</A></FONT> 
<BR><BR><FONT color=red size=3>ÀüÈ­ : 
016-812-4190</FONT>&nbsp;<BR><BR><BR><BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
</B></FONT></P> 
<P><FONT color=black 
size=3><B>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
<A target=_blank href="http://ebiz7.net/myteam/index.php?id=l99922"><FONT 
face="Times New Roman" color=blue size=5>È¨ÆäÀÌÁö¹Ù·Î°¡±â </FONT></A> <br><br><br> 
</BODY> 
</html> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 19 20:37:18 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA10013
	for <tsvwg-archive@odin.ietf.org>; Tue, 19 Feb 2002 20:37:17 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA14453
	for tsvwg-archive@odin.ietf.org; Tue, 19 Feb 2002 20:37:18 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA13214;
	Tue, 19 Feb 2002 20:04:42 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA13186
	for <tsvwg@optimus.ietf.org>; Tue, 19 Feb 2002 20:04:40 -0500 (EST)
Received: from msgbas1.cos.agilent.com (msgbas1x.cos.agilent.com [192.25.240.36])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA09647
	for <tsvwg@ietf.org>; Tue, 19 Feb 2002 20:04:39 -0500 (EST)
From: pat_thaler@agilent.com
Received: from msgrel1.cos.agilent.com (msgrel1.cos.agilent.com [130.29.152.77])
	by msgbas1.cos.agilent.com (Postfix) with ESMTP
	id DD17819E0; Tue, 19 Feb 2002 18:04:39 -0700 (MST)
Received: from axcsbh6.cos.agilent.com (axcsbh6.cos.agilent.com [130.29.152.171])
	by msgrel1.cos.agilent.com (Postfix) with SMTP
	id CA1A010B; Tue, 19 Feb 2002 18:04:39 -0700 (MST)
Received: from 130.29.152.171 by axcsbh6.cos.agilent.com (InterScan E-Mail VirusWall NT); Tue, 19 Feb 2002 18:04:39 -0700
Received: by axcsbh6.cos.agilent.com with Internet Mail Service (5.5.2653.19)
	id <FBZRKLWY>; Tue, 19 Feb 2002 18:04:39 -0700
Message-ID: <1BEBA5E8600DD4119A50009027AF54A00A512968@axcs04.cos.agilent.com>
To: steph@cs.uchicago.edu, pat_thaler@agilent.com
Cc: tsvwg@ietf.org
Subject: RE: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch f
	ramer for TC P 
Date: Tue, 19 Feb 2002 18:04:30 -0700
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Steph,

I wouldn't consider this a leading candidate as the potential inefficiencies
are 
too large. Packing can improve the efficiency if there is a stream of short
PDUs.
However, it is very common to have a mix of long and short PDUs in the data
path.
When a small PDU is followed by a large PDU such that the combination of the
two 
doesn't fit in the fixed length segment, then one has no alternative but to
pad
the segment containing the short PDU and start a new segment with the large
PDU.

For bulk data protocols, the overwhelming majority of data in large PDUs as
you say
IF one is comparing bytes in small packets vs bytes in large packets.
However, 
by packet count, there are a lot of small packets. Usually more than half
the packets
are short (commands, status, R2T's, etc). The mixture of packet sizes
reduces 
efficiency with a fixed length scheme.

Bandwidth in the LAN isn't expensive, but it isn't free either. More
importantly,
bandwith for applications crossing the WAN is expensive and wasting it in a
fixed
length "aligning FIM" would not be acceptable.

Regards,
Pat

-----Original Message-----
From: Stephen Bailey [mailto:steph@cs.uchicago.edu]
Sent: Friday, February 15, 2002 11:31 AM
To: THALER,PAT (A-Roseville,ex1)
Cc: tsvwg@ietf.org
Subject: Re: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch
framer for TC P 


Pat,

> I don't understand what you are suggesting.

Actually, it seems you do understand what he's suggesting, but are not
buying it :^)

The way I'd define it, it's not that the ULP PDUs are fixed length,
but that the framing protocol may add padding within a finite
interval.  For best transport efficiency, you would need to do what
TUF calls `ULPDU packing'.  You may end up with more than one ULPDU
per finite interval.

It actually seems like it might be a pretty good scheme.  Your finite
interval would be EMSS, and the segmentation algorithm would be
similar to the current TUF except to get markers at the beginning of
segments.

This might be a new leading candidate.  It's no touch and completely
safe.  The cost is that you must either do ULPDU packing (the
unpacking is actually the burdensome part, but I personally never
thought it was too horrible), or live with inefficient transport of
small messages.  For bulk data protocols (e.g. storage) the
overwhelming majority of the data you carry is in large
(interval-sized) PDUs, so a small loss of transport efficiency for the
control PDUs won't hurt.

Alright, you've done it again Jim!  By muddying the waters we clarify
them :-?

Steph

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 20 12:21:12 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA07629
	for <tsvwg-archive@odin.ietf.org>; Wed, 20 Feb 2002 12:21:12 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id MAA08023
	for tsvwg-archive@odin.ietf.org; Wed, 20 Feb 2002 12:21:15 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA05136;
	Wed, 20 Feb 2002 11:58:41 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA05107
	for <tsvwg@optimus.ietf.org>; Wed, 20 Feb 2002 11:58:39 -0500 (EST)
Received: from palrel12.hp.com (palrel12.hp.com [156.153.255.237])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA06539
	for <tsvwg@ietf.org>; Wed, 20 Feb 2002 11:58:35 -0500 (EST)
Received: from xparelay2.corp.hp.com (xparelay2.corp.hp.com [15.58.137.112])
	by palrel12.hp.com (Postfix) with ESMTP
	id 1F274600818; Wed, 20 Feb 2002 08:58:07 -0800 (PST)
Received: from xpabh3.corp.hp.com (xpabh3.corp.hp.com [15.58.136.223])
	by xparelay2.corp.hp.com (Postfix) with ESMTP
	id 0557391; Wed, 20 Feb 2002 08:58:02 -0800 (PST)
Received: by xpabh3.corp.hp.com with Internet Mail Service (5.5.2653.19)
	id <F2PKDJNA>; Wed, 20 Feb 2002 08:58:01 -0800
Message-ID: <499DC368E25AD411B3F100902740AD650AFC7DBD@xrose03.rose.hp.com>
From: "WENDT,JIM (HP-Roseville,ex1)" <jim_wendt@hp.com>
To: "'pat_thaler@agilent.com'" <pat_thaler@agilent.com>, steph@cs.uchicago.edu
Cc: tsvwg@ietf.org
Subject: RE: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch f
	 ramer for TC P 
Date: Wed, 20 Feb 2002 08:58:00 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Pat & Steph,
There are two new variations under discussion here.
Maybe we can establish unique names for them now.

* "Padded FIM"
  Steph's version inserts pad bytes after one or more
  packed ULP PDUs to reach the next marker interval
  (which is the EMSS and forces the marker to
  occupy the first bytes in the next TCP segment 
  payload). Sort of a padded-FIM scheme.

* "Aligning FIM"
  My original scheme was to align ULP PDUs to the start
  of TCP segments, pack mutiple PDUs into a TCP segment
  if possible, and allow the markers to fall where they
  may in the data stream (and not necessarily in the
  same TCP segment as the ULP PDU they point to). This
  scheme doesn't insert pad bytes into the stream.
  When the ULP PDUs are limited to the EMSS, it does
  allow each ULP PDU to be fully contained in a single
  TCP segment and allows that segment to be immediately
  and fully processed and direct placed with no need
  to maintain residual state information. The impact on
  network bandwidth is that some shorter TCP segments
  may be emitted (i.e. more protocol header bytes
  vs data bytes) depending upon the characteristics
  of the ULP packet stream.

Jim


> -----Original Message-----
> From: pat_thaler@agilent.com [mailto:pat_thaler@agilent.com]
> Sent: Tuesday, February 19, 2002 5:05 PM
> To: steph@cs.uchicago.edu; pat_thaler@agilent.com
> Cc: tsvwg@ietf.org
> Subject: RE: [Tsvwg] RE: [rdma] a proposal for a different 
> No-Data-Touch
> f ramer for TC P 
> 
> 
> Steph,
> 
> I wouldn't consider this a leading candidate as the potential 
> inefficiencies
> are 
> too large. Packing can improve the efficiency if there is a 
> stream of short
> PDUs.
> However, it is very common to have a mix of long and short 
> PDUs in the data
> path.
> When a small PDU is followed by a large PDU such that the 
> combination of the
> two 
> doesn't fit in the fixed length segment, then one has no 
> alternative but to
> pad
> the segment containing the short PDU and start a new segment 
> with the large
> PDU.
> 
> For bulk data protocols, the overwhelming majority of data in 
> large PDUs as
> you say
> IF one is comparing bytes in small packets vs bytes in large packets.
> However, 
> by packet count, there are a lot of small packets. Usually 
> more than half
> the packets
> are short (commands, status, R2T's, etc). The mixture of packet sizes
> reduces 
> efficiency with a fixed length scheme.
> 
> Bandwidth in the LAN isn't expensive, but it isn't free either. More
> importantly,
> bandwith for applications crossing the WAN is expensive and 
> wasting it in a
> fixed
> length "aligning FIM" would not be acceptable.
> 
> Regards,
> Pat
> 
> -----Original Message-----
> From: Stephen Bailey [mailto:steph@cs.uchicago.edu]
> Sent: Friday, February 15, 2002 11:31 AM
> To: THALER,PAT (A-Roseville,ex1)
> Cc: tsvwg@ietf.org
> Subject: Re: [Tsvwg] RE: [rdma] a proposal for a different 
> No-Data-Touch
> framer for TC P 
> 
> 
> Pat,
> 
> > I don't understand what you are suggesting.
> 
> Actually, it seems you do understand what he's suggesting, but are not
> buying it :^)
> 
> The way I'd define it, it's not that the ULP PDUs are fixed length,
> but that the framing protocol may add padding within a finite
> interval.  For best transport efficiency, you would need to do what
> TUF calls `ULPDU packing'.  You may end up with more than one ULPDU
> per finite interval.
> 
> It actually seems like it might be a pretty good scheme.  Your finite
> interval would be EMSS, and the segmentation algorithm would be
> similar to the current TUF except to get markers at the beginning of
> segments.
> 
> This might be a new leading candidate.  It's no touch and completely
> safe.  The cost is that you must either do ULPDU packing (the
> unpacking is actually the burdensome part, but I personally never
> thought it was too horrible), or live with inefficient transport of
> small messages.  For bulk data protocols (e.g. storage) the
> overwhelming majority of the data you carry is in large
> (interval-sized) PDUs, so a small loss of transport efficiency for the
> control PDUs won't hurt.
> 
> Alright, you've done it again Jim!  By muddying the waters we clarify
> them :-?
> 
> Steph
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 20 16:48:07 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA21360
	for <tsvwg-archive@odin.ietf.org>; Wed, 20 Feb 2002 16:48:07 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA23034
	for tsvwg-archive@odin.ietf.org; Wed, 20 Feb 2002 16:48:11 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA22634;
	Wed, 20 Feb 2002 16:34:02 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA22605
	for <tsvwg@optimus.ietf.org>; Wed, 20 Feb 2002 16:34:00 -0500 (EST)
Received: from c003.snv.cp.net (c003-h015.c003.snv.cp.net [209.228.32.229])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id QAA21057
	for <tsvwg@ietf.org>; Wed, 20 Feb 2002 16:33:55 -0500 (EST)
Received: (cpmta 23495 invoked from network); 20 Feb 2002 13:33:13 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.229) with SMTP; 20 Feb 2002 13:33:13 -0800
X-Sent: 20 Feb 2002 21:33:13 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: <tsvwg@ietf.org>
Subject: RE: [Tsvwg] RE: [rdma] a proposal for a different No-Data-Touch f ramer for TC P 
Date: Wed, 20 Feb 2002 13:35:52 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJMENBCPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
In-Reply-To: <499DC368E25AD411B3F100902740AD650AFC7DBD@xrose03.rose.hp.com>
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Importance: Normal
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Jim,

There could be a third variant.  FIM with TLV bundling by creating a common
header confirmed by a checking scheme followed by standardized encapsulation
of ULP data units called Chunks.  This TLV bundling would reduce the affects
of alignment attempts.

Doug

> Pat & Steph,
> There are two new variations under discussion here.
> Maybe we can establish unique names for them now.
>
> * "Padded FIM"
>   Steph's version inserts pad bytes after one or more
>   packed ULP PDUs to reach the next marker interval
>   (which is the EMSS and forces the marker to
>   occupy the first bytes in the next TCP segment
>   payload). Sort of a padded-FIM scheme.
>
> * "Aligning FIM"
>   My original scheme was to align ULP PDUs to the start
>   of TCP segments, pack mutiple PDUs into a TCP segment
>   if possible, and allow the markers to fall where they
>   may in the data stream (and not necessarily in the
>   same TCP segment as the ULP PDU they point to). This
>   scheme doesn't insert pad bytes into the stream.
>   When the ULP PDUs are limited to the EMSS, it does
>   allow each ULP PDU to be fully contained in a single
>   TCP segment and allows that segment to be immediately
>   and fully processed and direct placed with no need
>   to maintain residual state information. The impact on
>   network bandwidth is that some shorter TCP segments
>   may be emitted (i.e. more protocol header bytes
>   vs data bytes) depending upon the characteristics
>   of the ULP packet stream.
>
> Jim
>
>
> > -----Original Message-----
> > From: pat_thaler@agilent.com [mailto:pat_thaler@agilent.com]
> > Sent: Tuesday, February 19, 2002 5:05 PM
> > To: steph@cs.uchicago.edu; pat_thaler@agilent.com
> > Cc: tsvwg@ietf.org
> > Subject: RE: [Tsvwg] RE: [rdma] a proposal for a different
> > No-Data-Touch
> > f ramer for TC P
> >
> >
> > Steph,
> >
> > I wouldn't consider this a leading candidate as the potential
> > inefficiencies
> > are
> > too large. Packing can improve the efficiency if there is a
> > stream of short
> > PDUs.
> > However, it is very common to have a mix of long and short
> > PDUs in the data
> > path.
> > When a small PDU is followed by a large PDU such that the
> > combination of the
> > two
> > doesn't fit in the fixed length segment, then one has no
> > alternative but to
> > pad
> > the segment containing the short PDU and start a new segment
> > with the large
> > PDU.
> >
> > For bulk data protocols, the overwhelming majority of data in
> > large PDUs as
> > you say
> > IF one is comparing bytes in small packets vs bytes in large packets.
> > However,
> > by packet count, there are a lot of small packets. Usually
> > more than half
> > the packets
> > are short (commands, status, R2T's, etc). The mixture of packet sizes
> > reduces
> > efficiency with a fixed length scheme.
> >
> > Bandwidth in the LAN isn't expensive, but it isn't free either. More
> > importantly,
> > bandwith for applications crossing the WAN is expensive and
> > wasting it in a
> > fixed
> > length "aligning FIM" would not be acceptable.
> >
> > Regards,
> > Pat
> >
> > -----Original Message-----
> > From: Stephen Bailey [mailto:steph@cs.uchicago.edu]
> > Sent: Friday, February 15, 2002 11:31 AM
> > To: THALER,PAT (A-Roseville,ex1)
> > Cc: tsvwg@ietf.org
> > Subject: Re: [Tsvwg] RE: [rdma] a proposal for a different
> > No-Data-Touch
> > framer for TC P
> >
> >
> > Pat,
> >
> > > I don't understand what you are suggesting.
> >
> > Actually, it seems you do understand what he's suggesting, but are not
> > buying it :^)
> >
> > The way I'd define it, it's not that the ULP PDUs are fixed length,
> > but that the framing protocol may add padding within a finite
> > interval.  For best transport efficiency, you would need to do what
> > TUF calls `ULPDU packing'.  You may end up with more than one ULPDU
> > per finite interval.
> >
> > It actually seems like it might be a pretty good scheme.  Your finite
> > interval would be EMSS, and the segmentation algorithm would be
> > similar to the current TUF except to get markers at the beginning of
> > segments.
> >
> > This might be a new leading candidate.  It's no touch and completely
> > safe.  The cost is that you must either do ULPDU packing (the
> > unpacking is actually the burdensome part, but I personally never
> > thought it was too horrible), or live with inefficient transport of
> > small messages.  For bulk data protocols (e.g. storage) the
> > overwhelming majority of the data you carry is in large
> > (interval-sized) PDUs, so a small loss of transport efficiency for the
> > control PDUs won't hurt.
> >
> > Alright, you've done it again Jim!  By muddying the waters we clarify
> > them :-?
> >
> > Steph
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> >
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
>


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 21 08:41:57 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA16716
	for <tsvwg-archive@odin.ietf.org>; Thu, 21 Feb 2002 08:41:57 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA21118
	for tsvwg-archive@odin.ietf.org; Thu, 21 Feb 2002 08:42:00 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA20065;
	Thu, 21 Feb 2002 08:25:54 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA20034
	for <tsvwg@optimus.ietf.org>; Thu, 21 Feb 2002 08:25:51 -0500 (EST)
Received: from gorilla.mchh.siemens.de (gorilla.mchh.siemens.de [194.138.158.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA16266
	for <tsvwg@ietf.org>; Thu, 21 Feb 2002 08:25:47 -0500 (EST)
Received: from blues.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.227])
	by gorilla.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id OAA21871;
	Thu, 21 Feb 2002 14:25:44 +0100 (MET)
Received: from mchh273e.demchh201e.icn.siemens.de ([139.21.200.83])
	by blues.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id OAA15245;
	Thu, 21 Feb 2002 14:25:35 +0100 (MET)
Received: by MCHH273E with Internet Mail Service (5.5.2653.19)
	id <1JWY1NCD>; Thu, 21 Feb 2002 14:25:42 +0100
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF788@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'bidulock@openss7.org'" <bidulock@openss7.org>
Cc: "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "'discussion@sctp.de'"
	 <discussion@sctp.de>
Date: Thu, 21 Feb 2002 14:25:26 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="ISO-8859-1"
Subject: [Tsvwg] RE: [SCTP-Discussion] Multihoming
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Brian,

I know this. But I do not see why accepting the INIT-ACK
coming from an unknown address, but the known one is
in the list of addresses, is a serious problem, at least
from the security point of view.
As I mentioned, the assoc will not gain the expected 
network fault tolerance as it could be expected by only
looking at the number of paths, but does that satisfy
the non establishment of the assoc. It is sure that
the network is not configured as it should, but ...

Best regards
Michael


Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> Sent:	Thursday, February 21, 2002 2:05 PM
> To:	Tuexen Michael
> Cc:	'discussion@sctp.de'
> Subject:	Re: [SCTP-Discussion] Multihoming
> 
> Michael,
> 
> Not even TCP will accept such an arrangement.  Say TCP launches
> a SYN from B to A_1.  If A responds with a SYN from A_2 to B the
> connection will not succeed.
> 
> --brian
> 
> On Thu, 21 Feb 2002, Tuexen Michael wrote:
> 
> > Dear Roland,
> > 
> > considering the case where the SCTP end point A only announces
> > A2, but A1 is used as a source IP address is definitly a
> > misconfiguration of the host!
> > If a SCTP end point only announces a proper subset of its IP-addresses
> > it must make sure that only addresses of this subset are used as 
> > source IP addresses. The most simple way to make this sure is to
> > include all addresses. This is the reason why you can read the 
> > Implemention note in RFC 2960. The sctplib also supports using this
> > subset but then your forwaring on the host must be in tune with this
> > configuration as stated in the documentation.
> > 
> > Considering the other case: B sends traffic from B1 to A2. A replies
> > then from A1 to B1. This also means that your forwaring setup is
> > not in a proper way configured so that you can really make use of the
> > multihoming support of SCTP. The problems are described in 
> > http://www.ietf.org/internet-drafts/draft-coene-sctp-multihome-03.txt
> > Nevertheless, when an SCTP endpoint receives an INIT-ACK it should
> > check if it can find a TCB based on the source IP address and the
> > IP addresses listed in the INIT-ACK. This seems not to be implemented
> > in pre17 and will be added in a future release. This would result
> > in being able to establish the association, but you do not have the
> > redundancy you think based on the number of paths.
> > 
> > Best regards
> > Michael
> >      
> > 
> > Michael Tuexen    Tel.:   +49 89 722 47210
> > Siemens AG        Fax:    +49 89 722 48212
> > ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
> > 
> > > -----Original Message-----
> > > From:	Roland Gloeckler [SMTP:gloeckle@ind.uni-stuttgart.de]
> > > Sent:	Thursday, February 21, 2002 1:13 PM
> > > To:	bidulock@openss7.org
> > > Cc:	discussion@sctp.de
> > > Subject:	Re: [SCTP-Discussion] Multihoming
> > > 
> > > Dear Mr. Bidulock, dear all:
> > > 
> > > My question is not completely answered by your response.
> > > 
> > > My problem (a little bit more detailed) is the following:
> > > 
> > > A has two interfaces with distinct IP addresses (A_1 and A_2).
> > > B sends an INIT to one of them, e.g., A_2. The other interface A_1 is not 
> > > known to B yet.
> > > 
> > > In the sctp implementation A sends the INIT_ACK via the interface which is 
> > > the default entry in the IP routing table, e.g., A_1.
> > > 
> > > Now there are two error scenarios:
> > > 
> > > A has initialised a multihomed instance with both IP addresses.
> > > For sending the INIT ACK it uses A_1 which is not known to B before it reads 
> > > the address list (with A_1 and A_2) in this INIT ACK chunk. 
> > > B responds with an ABORT.
> > > 
> > > or
> > > 
> > > A has initialised a singlehomed instance with the address A_2, which receives > 
> > > the INIT.
> > > For sending the INIT ACK it uses A_1 which is not even initialised in the 
> > > SCTP instance. In the cookie it uses A_2.
> > > So it is obvious that SCTP has no control over the IP source address.
> > > An association cannot be initialized. 
> > > 
> > > To resume:
> > > 
> > > If the INIT_ACK chunk of an instance A uses in its IP header a source IP 
> > > address that is different from the destination address of the INIT chunk from 
> > > B, B refuses to accept it.
> > > 
> > > 
> > > Section 5.1.2.c of the RFC says: 
> > > 
> > > "IMPLEMENTATION NOTE: In some cases (e.g., when the implementation doesn't 
> > > control the source IP address that is used for transmitting), an endpoint 
> > > might need to include in its INIT or INIT ACK all possible IP addresses from 
> > > which packets to the peer could be transmitted."
> > > 
> > > B would have to read the list of IP addresses in the INIT ACK before it 
> > > searches in its association list to know the chunk belongs to A.
> > > 
> > > Or should the implementation force its IP layer to use the INIT destination 
> > > address as the source address for the INIT ACK? The RFC tries to cover cases 
> > > where this is not the case.
> > > 
> > > Regards,
> > > 
> > > Roland
> > > 
> > > 
> > > 
> > > On Friday 01 February 2002 20:38, Brian F. G. Bidulock wrote:
> > > > Roland,
> > > >
> > > > The effect your describe was discussed on TSVWG.  The
> > > > defficiency of SCTP was characterized as RFC2960's inability> 
> > > > to provide for reliable initialization of an association.
> > > >
> > > > The effect you describe is required by RFC 2960 section 5.1
> > > > paragraph (B): "... The destination IP address of the INIT ACK
> > > > MUST be set to the source IP address of the INIT to which this
> > > > INIT ACK is responding."
> > > >
> > > > The behavior exhibited by SCTPlib is correct.  The
> > > > implementation should reply with an ABORT in response to the
> > > > INIT ACK sent to the wrong address.
> > > >
> > > > --brian
> > > >
> > > > On Fri, 01 Feb 2002, Roland Gloeckler wrote:
> > > > > Dear all,
> > > > >
> > > > > when I was working with SCTPlib 1 pre14 I found an
> > > > > interesting behaviour.
> > > > >
> > > > > I had one PC (let's call it "A") with two network cards
> > > > > (let's call them A_1 and A_2) connected to another one
> > > > > (let's call it "B") with only one (B_1).
> > > > >
> > > > > The configuration was: Interfaces A_1 and B_1 were connected
> > > > > to the same LAN and A_2 was connected to the LAN via a
> > > > > router.
> > > > >
> > > > > After some minor changes to testengine.c I was able to use
> > > > > test_tool to initialise an SCTP multihoming instance on A
> > > > > for the addresses A_1 and A_2.
> > > > >
> > > > > In the script for B I tried to start an association with A,
> > > > > but only included one IP address, e.g., A_2.
> > > > >
> > > > > B successfully sent an INIT message to A_2, which responded
> > > > > with a valid INIT_ACK.
> > > > >
> > > > > If A responded via A_2 all went well.
> > > > >
> > > > > BUT: 	When A used the other interface A_1 to send the
> > > > > INIT_ACK (which included the list of its own addresses A_1
> > > > > and A_2 belonging to that instance) to B, B was not able to
> > > > > find the address A_1 in its association list, since it
> > > > > didn't know it yet, and answered with an ABORT message to A.
> > > > >
> > > > > If I change the destination address in B's script to A_1 and
> > > > > change the routing table in A to send packets to B via A_2
> > > > > the result is the same.
> > > > >
> > > > > To summarize:
> > > > >
> > > > > If the INIT_ACK chunk of a multihomed instance A uses a
> > > > > source IP address that is different from the one B sent the
> > > > > INIT chunk to, B refuses to accept it.
> > > > >
> > > > > When I was looking into the RFC to search for more details,
> > > > > I only found out that the behaviour is according to the> 
> > > > > rules that no INIT_ACK is to be accepted which uses an
> > > > > unknown source transport address.
> > > > >
> > > > > Now I wonder if this is a general problem not answered by
> > > > > the RFC.
> > > > >
> > > > > I can think of several ways to come around the problem for
> > > > > my purposes, but they are not adequate for general purposes.
> > > > >
> > > > > Has anyone heard of this situation before? Please tell me.
> > > > >
> > > > > Regards,
> > > > >
> > > > > Roland
> > > > > _______________________________________________
> > > > > discussion mailing list
> > > > > discussion@sctp.de
> > > > > http://www.sctp.de/mailman/listinfo/discussion
> > > _______________________________________________
> > > discussion mailing list
> > > discussion@sctp.de
> > > http://www.sctp.de/mailman/listinfo/discussion
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 21 10:43:17 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA21146
	for <tsvwg-archive@odin.ietf.org>; Thu, 21 Feb 2002 10:43:17 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA29797
	for tsvwg-archive@odin.ietf.org; Thu, 21 Feb 2002 10:43:19 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA28042;
	Thu, 21 Feb 2002 10:21:19 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA28014
	for <tsvwg@optimus.ietf.org>; Thu, 21 Feb 2002 10:21:15 -0500 (EST)
Received: from stewartlin.chicago.il.us (host-64-47-24-166.masergy.com [64.47.24.166] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA20094
	for <tsvwg@ietf.org>; Thu, 21 Feb 2002 10:21:10 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewartlin.chicago.il.us (8.11.0/8.11.0) with ESMTP id g1LFMer12977;
	Thu, 21 Feb 2002 09:22:40 -0600
Message-ID: <3C75104E.53BC132F@stewart.chicago.il.us>
Date: Thu, 21 Feb 2002 09:20:46 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
CC: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "'discussion@sctp.de'" <discussion@sctp.de>
Subject: Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
References: <F92978223B01D311ACFF0008C71EE19D015DF788@MCHH225E>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Michael:

From digging through this thread is this the scenario:


Endpoint 1                                           Endpoint 2
IP-A                                                    IP-C
IP-B                                                    IP-D
------------Frm:IP-A:INIT(List-IP-A)--------->To:IP-C
    <---------------To:IP-A---Frm:IP-D:INIT-ACK(List-IP-C,IP-D)----

Now the above case IS legal and there is specific wording in
the RFC that covers this. The idea is that Endpoint-1 MUST
lookup not only the FROM address in the IP header but also
MUST lookup in the list of addresses (IP-C/IP-D in this case)
if it can't find its TCB.

R

Tuexen Michael wrote:
> 
> Brian,
> 
> I know this. But I do not see why accepting the INIT-ACK
> coming from an unknown address, but the known one is
> in the list of addresses, is a serious problem, at least
> from the security point of view.
> As I mentioned, the assoc will not gain the expected
> network fault tolerance as it could be expected by only
> looking at the number of paths, but does that satisfy
> the non establishment of the assoc. It is sure that
> the network is not configured as it should, but ...
> 
> Best regards
> Michael
> 
> Michael Tuexen    Tel.:   +49 89 722 47210
> Siemens AG        Fax:    +49 89 722 48212
> ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
> 
> > -----Original Message-----
> > From: Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > Sent: Thursday, February 21, 2002 2:05 PM
> > To:   Tuexen Michael
> > Cc:   'discussion@sctp.de'
> > Subject:      Re: [SCTP-Discussion] Multihoming
> >
> > Michael,
> >
> > Not even TCP will accept such an arrangement.  Say TCP launches
> > a SYN from B to A_1.  If A responds with a SYN from A_2 to B the
> > connection will not succeed.
> >
> > --brian
> >
> > On Thu, 21 Feb 2002, Tuexen Michael wrote:
> >
> > > Dear Roland,
> > >
> > > considering the case where the SCTP end point A only announces
> > > A2, but A1 is used as a source IP address is definitly a
> > > misconfiguration of the host!
> > > If a SCTP end point only announces a proper subset of its IP-addresses
> > > it must make sure that only addresses of this subset are used as
> > > source IP addresses. The most simple way to make this sure is to
> > > include all addresses. This is the reason why you can read the
> > > Implemention note in RFC 2960. The sctplib also supports using this
> > > subset but then your forwaring on the host must be in tune with this
> > > configuration as stated in the documentation.
> > >
> > > Considering the other case: B sends traffic from B1 to A2. A replies
> > > then from A1 to B1. This also means that your forwaring setup is
> > > not in a proper way configured so that you can really make use of the
> > > multihoming support of SCTP. The problems are described in
> > > http://www.ietf.org/internet-drafts/draft-coene-sctp-multihome-03.txt
> > > Nevertheless, when an SCTP endpoint receives an INIT-ACK it should
> > > check if it can find a TCB based on the source IP address and the
> > > IP addresses listed in the INIT-ACK. This seems not to be implemented
> > > in pre17 and will be added in a future release. This would result
> > > in being able to establish the association, but you do not have the
> > > redundancy you think based on the number of paths.
> > >
> > > Best regards
> > > Michael
> > >
> > >
> > > Michael Tuexen    Tel.:   +49 89 722 47210
> > > Siemens AG        Fax:    +49 89 722 48212
> > > ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
> > >
> > > > -----Original Message-----
> > > > From:     Roland Gloeckler [SMTP:gloeckle@ind.uni-stuttgart.de]
> > > > Sent:     Thursday, February 21, 2002 1:13 PM
> > > > To:       bidulock@openss7.org
> > > > Cc:       discussion@sctp.de
> > > > Subject:  Re: [SCTP-Discussion] Multihoming
> > > >
> > > > Dear Mr. Bidulock, dear all:
> > > >
> > > > My question is not completely answered by your response.
> > > >
> > > > My problem (a little bit more detailed) is the following:
> > > >
> > > > A has two interfaces with distinct IP addresses (A_1 and A_2).
> > > > B sends an INIT to one of them, e.g., A_2. The other interface A_1 is not
> > > > known to B yet.
> > > >
> > > > In the sctp implementation A sends the INIT_ACK via the interface which is
> > > > the default entry in the IP routing table, e.g., A_1.
> > > >
> > > > Now there are two error scenarios:
> > > >
> > > > A has initialised a multihomed instance with both IP addresses.
> > > > For sending the INIT ACK it uses A_1 which is not known to B before it reads
> > > > the address list (with A_1 and A_2) in this INIT ACK chunk.
> > > > B responds with an ABORT.
> > > >
> > > > or
> > > >
> > > > A has initialised a singlehomed instance with the address A_2, which receives >
> > > > the INIT.
> > > > For sending the INIT ACK it uses A_1 which is not even initialised in the
> > > > SCTP instance. In the cookie it uses A_2.
> > > > So it is obvious that SCTP has no control over the IP source address.
> > > > An association cannot be initialized.
> > > >
> > > > To resume:
> > > >
> > > > If the INIT_ACK chunk of an instance A uses in its IP header a source IP
> > > > address that is different from the destination address of the INIT chunk from
> > > > B, B refuses to accept it.
> > > >
> > > >
> > > > Section 5.1.2.c of the RFC says:
> > > >
> > > > "IMPLEMENTATION NOTE: In some cases (e.g., when the implementation doesn't
> > > > control the source IP address that is used for transmitting), an endpoint
> > > > might need to include in its INIT or INIT ACK all possible IP addresses from
> > > > which packets to the peer could be transmitted."
> > > >
> > > > B would have to read the list of IP addresses in the INIT ACK before it
> > > > searches in its association list to know the chunk belongs to A.
> > > >
> > > > Or should the implementation force its IP layer to use the INIT destination
> > > > address as the source address for the INIT ACK? The RFC tries to cover cases
> > > > where this is not the case.
> > > >
> > > > Regards,
> > > >
> > > > Roland
> > > >
> > > >
> > > >
> > > > On Friday 01 February 2002 20:38, Brian F. G. Bidulock wrote:
> > > > > Roland,
> > > > >
> > > > > The effect your describe was discussed on TSVWG.  The
> > > > > defficiency of SCTP was characterized as RFC2960's inability>
> > > > > to provide for reliable initialization of an association.
> > > > >
> > > > > The effect you describe is required by RFC 2960 section 5.1
> > > > > paragraph (B): "... The destination IP address of the INIT ACK
> > > > > MUST be set to the source IP address of the INIT to which this
> > > > > INIT ACK is responding."
> > > > >
> > > > > The behavior exhibited by SCTPlib is correct.  The
> > > > > implementation should reply with an ABORT in response to the
> > > > > INIT ACK sent to the wrong address.
> > > > >
> > > > > --brian
> > > > >
> > > > > On Fri, 01 Feb 2002, Roland Gloeckler wrote:
> > > > > > Dear all,
> > > > > >
> > > > > > when I was working with SCTPlib 1 pre14 I found an
> > > > > > interesting behaviour.
> > > > > >
> > > > > > I had one PC (let's call it "A") with two network cards
> > > > > > (let's call them A_1 and A_2) connected to another one
> > > > > > (let's call it "B") with only one (B_1).
> > > > > >
> > > > > > The configuration was: Interfaces A_1 and B_1 were connected
> > > > > > to the same LAN and A_2 was connected to the LAN via a
> > > > > > router.
> > > > > >
> > > > > > After some minor changes to testengine.c I was able to use
> > > > > > test_tool to initialise an SCTP multihoming instance on A
> > > > > > for the addresses A_1 and A_2.
> > > > > >
> > > > > > In the script for B I tried to start an association with A,
> > > > > > but only included one IP address, e.g., A_2.
> > > > > >
> > > > > > B successfully sent an INIT message to A_2, which responded
> > > > > > with a valid INIT_ACK.
> > > > > >
> > > > > > If A responded via A_2 all went well.
> > > > > >
> > > > > > BUT:  When A used the other interface A_1 to send the
> > > > > > INIT_ACK (which included the list of its own addresses A_1
> > > > > > and A_2 belonging to that instance) to B, B was not able to
> > > > > > find the address A_1 in its association list, since it
> > > > > > didn't know it yet, and answered with an ABORT message to A.
> > > > > >
> > > > > > If I change the destination address in B's script to A_1 and
> > > > > > change the routing table in A to send packets to B via A_2
> > > > > > the result is the same.
> > > > > >
> > > > > > To summarize:
> > > > > >
> > > > > > If the INIT_ACK chunk of a multihomed instance A uses a
> > > > > > source IP address that is different from the one B sent the
> > > > > > INIT chunk to, B refuses to accept it.
> > > > > >
> > > > > > When I was looking into the RFC to search for more details,
> > > > > > I only found out that the behaviour is according to the>
> > > > > > rules that no INIT_ACK is to be accepted which uses an
> > > > > > unknown source transport address.
> > > > > >
> > > > > > Now I wonder if this is a general problem not answered by
> > > > > > the RFC.
> > > > > >
> > > > > > I can think of several ways to come around the problem for
> > > > > > my purposes, but they are not adequate for general purposes.
> > > > > >
> > > > > > Has anyone heard of this situation before? Please tell me.
> > > > > >
> > > > > > Regards,
> > > > > >
> > > > > > Roland
> > > > > > _______________________________________________
> > > > > > discussion mailing list
> > > > > > discussion@sctp.de
> > > > > > http://www.sctp.de/mailman/listinfo/discussion
> > > > _______________________________________________
> > > > discussion mailing list
> > > > discussion@sctp.de
> > > > http://www.sctp.de/mailman/listinfo/discussion
> >
> > --
> > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 21 11:01:25 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA22057
	for <tsvwg-archive@odin.ietf.org>; Thu, 21 Feb 2002 11:01:25 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA01195
	for tsvwg-archive@odin.ietf.org; Thu, 21 Feb 2002 11:01:26 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA29519;
	Thu, 21 Feb 2002 10:39:39 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA29444
	for <tsvwg@optimus.ietf.org>; Thu, 21 Feb 2002 10:39:30 -0500 (EST)
Received: from gorilla.mchh.siemens.de (gorilla.mchh.siemens.de [194.138.158.18])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA20758
	for <tsvwg@ietf.org>; Thu, 21 Feb 2002 10:39:22 -0500 (EST)
Received: from moody.mchh.siemens.de (mail2.mchh.siemens.de [194.138.158.226])
	by gorilla.mchh.siemens.de (8.9.3/8.9.3) with ESMTP id QAA05757;
	Thu, 21 Feb 2002 16:39:14 +0100 (MET)
Received: from mchh246e.demchh201e.icn.siemens.de ([139.21.200.56])
	by moody.mchh.siemens.de (8.9.1/8.9.1) with ESMTP id QAA11581;
	Thu, 21 Feb 2002 16:39:10 +0100 (MET)
Received: by MCHH246E with Internet Mail Service (5.5.2653.19)
	id <FGJXF3L4>; Thu, 21 Feb 2002 16:39:16 +0100
Message-ID: <F92978223B01D311ACFF0008C71EE19D015DF78B@MCHH225E>
From: Tuexen Michael <Michael.Tuexen@icn.siemens.de>
To: "'Randall Stewart'" <randall@stewart.chicago.il.us>
Cc: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        "'tsvwg@ietf.org'"
	 <tsvwg@ietf.org>,
        "'discussion@sctp.de'" <discussion@sctp.de>,
        Roland Gloeckler <gloeckle@ind.uni-stuttgart.de>
Subject: RE: [Tsvwg] RE: [SCTP-Discussion] Multihoming
Date: Thu, 21 Feb 2002 16:39:07 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Dear all,

so my impression was OK, we will make sure that the
sctplib will behave like this in the next prerelease.

Best regards
Michael

Michael Tuexen    Tel.:   +49 89 722 47210
Siemens AG        Fax:    +49 89 722 48212
ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de

> -----Original Message-----
> From:	Randall Stewart [SMTP:randall@stewart.chicago.il.us]
> Sent:	Thursday, February 21, 2002 4:21 PM
> To:	Tuexen Michael
> Cc:	'bidulock@openss7.org'; 'tsvwg@ietf.org'; 'discussion@sctp.de'
> Subject:	Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
> 
> Michael:
> 
> From digging through this thread is this the scenario:
> 
> 
> Endpoint 1                                           Endpoint 2
> IP-A                                                    IP-C
> IP-B                                                    IP-D
> ------------Frm:IP-A:INIT(List-IP-A)--------->To:IP-C
>     <---------------To:IP-A---Frm:IP-D:INIT-ACK(List-IP-C,IP-D)----
> 
> Now the above case IS legal and there is specific wording in
> the RFC that covers this. The idea is that Endpoint-1 MUST
> lookup not only the FROM address in the IP header but also
> MUST lookup in the list of addresses (IP-C/IP-D in this case)
> if it can't find its TCB.
> 
> R
> 
> Tuexen Michael wrote:
> > 
> > Brian,
> > 
> > I know this. But I do not see why accepting the INIT-ACK
> > coming from an unknown address, but the known one is
> > in the list of addresses, is a serious problem, at least
> > from the security point of view.
> > As I mentioned, the assoc will not gain the expected
> > network fault tolerance as it could be expected by only
> > looking at the number of paths, but does that satisfy
> > the non establishment of the assoc. It is sure that
> > the network is not configured as it should, but ...
> > 
> > Best regards
> > Michael
> > 
> > Michael Tuexen    Tel.:   +49 89 722 47210
> > Siemens AG        Fax:    +49 89 722 48212
> > ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
> > 
> > > -----Original Message-----
> > > From: Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > > Sent: Thursday, February 21, 2002 2:05 PM
> > > To:   Tuexen Michael
> > > Cc:   'discussion@sctp.de'
> > > Subject:      Re: [SCTP-Discussion] Multihoming
> > >
> > > Michael,
> > >
> > > Not even TCP will accept such an arrangement.  Say TCP launches
> > > a SYN from B to A_1.  If A responds with a SYN from A_2 to B the
> > > connection will not succeed.
> > >
> > > --brian
> > >
> > > On Thu, 21 Feb 2002, Tuexen Michael wrote:
> > >
> > > > Dear Roland,
> > > >
> > > > considering the case where the SCTP end point A only announces
> > > > A2, but A1 is used as a source IP address is definitly a
> > > > misconfiguration of the host!
> > > > If a SCTP end point only announces a proper subset of its IP-addresses
> > > > it must make sure that only addresses of this subset are used as
> > > > source IP addresses. The most simple way to make this sure is to
> > > > include all addresses. This is the reason why you can read the
> > > > Implemention note in RFC 2960. The sctplib also supports using this
> > > > subset but then your forwaring on the host must be in tune with this
> > > > configuration as stated in the documentation.
> > > >
> > > > Considering the other case: B sends traffic from B1 to A2. A replies
> > > > then from A1 to B1. This also means that your forwaring setup is
> > > > not in a proper way configured so that you can really make use of the
> > > > multihoming support of SCTP. The problems are described in
> > > > http://www.ietf.org/internet-drafts/draft-coene-sctp-multihome-03.txt
> > > > Nevertheless, when an SCTP endpoint receives an INIT-ACK it should
> > > > check if it can find a TCB based on the source IP address and the
> > > > IP addresses listed in the INIT-ACK. This seems not to be implemented
> > > > in pre17 and will be added in a future release. This would result> 
> > > > in being able to establish the association, but you do not have the
> > > > redundancy you think based on the number of paths.
> > > >
> > > > Best regards
> > > > Michael
> > > >
> > > >
> > > > Michael Tuexen    Tel.:   +49 89 722 47210
> > > > Siemens AG        Fax:    +49 89 722 48212
> > > > ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
> > > >
> > > > > -----Original Message-----
> > > > > From:     Roland Gloeckler [SMTP:gloeckle@ind.uni-stuttgart.de]
> > > > > Sent:     Thursday, February 21, 2002 1:13 PM
> > > > > To:       bidulock@openss7.org
> > > > > Cc:       discussion@sctp.de
> > > > > Subject:  Re: [SCTP-Discussion] Multihoming
> > > > >
> > > > > Dear Mr. Bidulock, dear all:
> > > > >
> > > > > My question is not completely answered by your response.
> > > > >
> > > > > My problem (a little bit more detailed) is the following:
> > > > >
> > > > > A has two interfaces with distinct IP addresses (A_1 and A_2).
> > > > > B sends an INIT to one of them, e.g., A_2. The other interface A_1 is not
> > > > > known to B yet.
> > > > >
> > > > > In the sctp implementation A sends the INIT_ACK via the interface which is
> > > > > the default entry in the IP routing table, e.g., A_1.
> > > > >
> > > > > Now there are two error scenarios:
> > > > >
> > > > > A has initialised a multihomed instance with both IP addresses.
> > > > > For sending the INIT ACK it uses A_1 which is not known to B before it reads
> > > > > the address list (with A_1 and A_2) in this INIT ACK chunk.
> > > > > B responds with an ABORT.
> > > > >
> > > > > or
> > > > >
> > > > > A has initialised a singlehomed instance with the address A_2, which receives >
> > > > > the INIT.
> > > > > For sending the INIT ACK it uses A_1 which is not even initialised in the
> > > > > SCTP instance. In the cookie it uses A_2.
> > > > > So it is obvious that SCTP has no control over the IP source address.
> > > > > An association cannot be initialized.
> > > > >
> > > > > To resume:
> > > > >
> > > > > If the INIT_ACK chunk of an instance A uses in its IP header a source IP
> > > > > address that is different from the destination address of the INIT chunk from
> > > > > B, B refuses to accept it.
> > > > >
> > > > >
> > > > > Section 5.1.2.c of the RFC says:
> > > > >
> > > > > "IMPLEMENTATION NOTE: In some cases (e.g., when the implementation doesn't
> > > > > control the source IP address that is used for transmitting), an endpoint
> > > > > might need to include in its INIT or INIT ACK all possible IP addresses from
> > > > > which packets to the peer could be transmitted."
> > > > >
> > > > > B would have to read the list of IP addresses in the INIT ACK before it
> > > > > searches in its association list to know the chunk belongs to A.
> > > > >
> > > > > Or should the implementation force its IP layer to use the INIT destination
> > > > > address as the source address for the INIT ACK? The RFC tries to cover cases
> > > > > where this is not the case.
> > > > >
> > > > > Regards,
> > > > >
> > > > > Roland
> > > > >
> > > > >
> > > > >
> > > > > On Friday 01 February 2002 20:38, Brian F. G. Bidulock wrote:
> > > > > > Roland,
> > > > > >
> > > > > > The effect your describe was discussed on TSVWG.  The
> > > > > > defficiency of SCTP was characterized as RFC2960's inability>
> > > > > > to provide for reliable initialization of an association.
> > > > > >
> > > > > > The effect you describe is required by RFC 2960 section 5.1
> > > > > > paragraph (B): "... The destination IP address of the INIT ACK
> > > > > > MUST be set to the source IP address of the INIT to which this
> > > > > > INIT ACK is responding."
> > > > > >
> > > > > > The behavior exhibited by SCTPlib is correct.  The
> > > > > > implementation should reply with an ABORT in response to the
> > > > > > INIT ACK sent to the wrong address.
> > > > > >
> > > > > > --brian
> > > > > >
> > > > > > On Fri, 01 Feb 2002, Roland Gloeckler wrote:> 
> > > > > > > Dear all,
> > > > > > >
> > > > > > > when I was working with SCTPlib 1 pre14 I found an
> > > > > > > interesting behaviour.
> > > > > > >
> > > > > > > I had one PC (let's call it "A") with two network cards
> > > > > > > (let's call them A_1 and A_2) connected to another one
> > > > > > > (let's call it "B") with only one (B_1).
> > > > > > >
> > > > > > > The configuration was: Interfaces A_1 and B_1 were connected
> > > > > > > to the same LAN and A_2 was connected to the LAN via a
> > > > > > > router.
> > > > > > >
> > > > > > > After some minor changes to testengine.c I was able to use
> > > > > > > test_tool to initialise an SCTP multihoming instance on A
> > > > > > > for the addresses A_1 and A_2.
> > > > > > >
> > > > > > > In the script for B I tried to start an association with A,
> > > > > > > but only included one IP address, e.g., A_2.
> > > > > > >
> > > > > > > B successfully sent an INIT message to A_2, which responded
> > > > > > > with a valid INIT_ACK.
> > > > > > >
> > > > > > > If A responded via A_2 all went well.
> > > > > > >
> > > > > > > BUT:  When A used the other interface A_1 to send the
> > > > > > > INIT_ACK (which included the list of its own addresses A_1
> > > > > > > and A_2 belonging to that instance) to B, B was not able to
> > > > > > > find the address A_1 in its association list, since it
> > > > > > > didn't know it yet, and answered with an ABORT message to A.
> > > > > > >
> > > > > > > If I change the destination address in B's script to A_1 and
> > > > > > > change the routing table in A to send packets to B via A_2
> > > > > > > the result is the same.
> > > > > > >
> > > > > > > To summarize:
> > > > > > >
> > > > > > > If the INIT_ACK chunk of a multihomed instance A uses a
> > > > > > > source IP address that is different from the one B sent the
> > > > > > > INIT chunk to, B refuses to accept it.
> > > > > > >
> > > > > > > When I was looking into the RFC to search for more details,
> > > > > > > I only found out that the behaviour is according to the>
> > > > > > > rules that no INIT_ACK is to be accepted which uses an
> > > > > > > unknown source transport address.
> > > > > > >
> > > > > > > Now I wonder if this is a general problem not answered by
> > > > > > > the RFC.
> > > > > > >
> > > > > > > I can think of several ways to come around the problem for
> > > > > > > my purposes, but they are not adequate for general purposes.
> > > > > > >
> > > > > > > Has anyone heard of this situation before? Please tell me.
> > > > > > >
> > > > > > > Regards,
> > > > > > >
> > > > > > > Roland
> > > > > > > _______________________________________________
> > > > > > > discussion mailing list
> > > > > > > discussion@sctp.de
> > > > > > > http://www.sctp.de/mailman/listinfo/discussion
> > > > > _______________________________________________
> > > > > discussion mailing list
> > > > > discussion@sctp.de
> > > > > http://www.sctp.de/mailman/listinfo/discussion
> > >
> > > --
> > > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> > >                         ¦ Therefore  all  progress  depends on the ¦
> > >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> > 
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 21 13:55:07 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA29394
	for <tsvwg-archive@odin.ietf.org>; Thu, 21 Feb 2002 13:55:07 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA19632
	for tsvwg-archive@odin.ietf.org; Thu, 21 Feb 2002 13:55:09 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA18734;
	Thu, 21 Feb 2002 13:38:12 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA18705
	for <tsvwg@optimus.ietf.org>; Thu, 21 Feb 2002 13:38:09 -0500 (EST)
Received: from calliope1.fm.intel.com (fmfdns01.fm.intel.com [132.233.247.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA28632
	for <tsvwg@ietf.org>; Thu, 21 Feb 2002 13:38:06 -0500 (EST)
Received: from fmsmsxvs043.fm.intel.com (fmsmsxv043-1.fm.intel.com [132.233.48.128])
	by calliope1.fm.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.51 2002/02/19 21:12:32 root Exp $) with SMTP id SAA13873
	for <tsvwg@ietf.org>; Thu, 21 Feb 2002 18:38:07 GMT
Received: from fmsmsx019.fm.intel.com ([132.233.42.130])
 by fmsmsxvs043.fm.intel.com (NAVGW 2.5.1.16) with SMTP id M2002022110371401706
 ; Thu, 21 Feb 2002 10:37:14 -0800
Received: by fmsmsx019.fm.intel.com with Internet Mail Service (5.5.2653.19)
	id <19DN89GN>; Thu, 21 Feb 2002 10:38:06 -0800
Message-ID: <E9B3905226DDD411BC0F0090276D21C20C64A1AC@FMSMSX30>
From: "Guo, Xingang" <xingang.guo@intel.com>
To: "'Randall Stewart'" <randall@stewart.chicago.il.us>,
        Tuexen Michael <Michael.Tuexen@icn.siemens.de>
Cc: "'bidulock@openss7.org'" <bidulock@openss7.org>,
        "'tsvwg@ietf.org'"<tsvwg@ietf.org>,
        "'discussion@sctp.de'" <discussion@sctp.de>
Subject: RE: [Tsvwg] RE: [SCTP-Discussion] Multihoming
Date: Thu, 21 Feb 2002 10:38:01 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Randall,

As far as I know, this is exactly what lksctp does.

www.sourceforge.net/projects/lksctp


--Xingang


> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Thursday, February 21, 2002 7:21 AM
> To: Tuexen Michael
> Cc: 'bidulock@openss7.org'; 'tsvwg@ietf.org'; 'discussion@sctp.de'
> Subject: Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
> 
> 
> Michael:
> 
> From digging through this thread is this the scenario:
> 
> 
> Endpoint 1                                           Endpoint 2
> IP-A                                                    IP-C
> IP-B                                                    IP-D
> ------------Frm:IP-A:INIT(List-IP-A)--------->To:IP-C
>     <---------------To:IP-A---Frm:IP-D:INIT-ACK(List-IP-C,IP-D)----
> 
> Now the above case IS legal and there is specific wording in
> the RFC that covers this. The idea is that Endpoint-1 MUST
> lookup not only the FROM address in the IP header but also
> MUST lookup in the list of addresses (IP-C/IP-D in this case)
> if it can't find its TCB.
> 
> R
> 
> Tuexen Michael wrote:
> > 
> > Brian,
> > 
> > I know this. But I do not see why accepting the INIT-ACK
> > coming from an unknown address, but the known one is
> > in the list of addresses, is a serious problem, at least
> > from the security point of view.
> > As I mentioned, the assoc will not gain the expected
> > network fault tolerance as it could be expected by only
> > looking at the number of paths, but does that satisfy
> > the non establishment of the assoc. It is sure that
> > the network is not configured as it should, but ...
> > 
> > Best regards
> > Michael
> > 
> > Michael Tuexen    Tel.:   +49 89 722 47210
> > Siemens AG        Fax:    +49 89 722 48212
> > ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
> > 
> > > -----Original Message-----
> > > From: Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > > Sent: Thursday, February 21, 2002 2:05 PM
> > > To:   Tuexen Michael
> > > Cc:   'discussion@sctp.de'
> > > Subject:      Re: [SCTP-Discussion] Multihoming
> > >
> > > Michael,
> > >
> > > Not even TCP will accept such an arrangement.  Say TCP launches
> > > a SYN from B to A_1.  If A responds with a SYN from A_2 to B the
> > > connection will not succeed.
> > >
> > > --brian
> > >
> > > On Thu, 21 Feb 2002, Tuexen Michael wrote:
> > >
> > > > Dear Roland,
> > > >
> > > > considering the case where the SCTP end point A only announces
> > > > A2, but A1 is used as a source IP address is definitly a
> > > > misconfiguration of the host!
> > > > If a SCTP end point only announces a proper subset of 
> its IP-addresses
> > > > it must make sure that only addresses of this subset are used as
> > > > source IP addresses. The most simple way to make this sure is to
> > > > include all addresses. This is the reason why you can read the
> > > > Implemention note in RFC 2960. The sctplib also 
> supports using this
> > > > subset but then your forwaring on the host must be in 
> tune with this
> > > > configuration as stated in the documentation.
> > > >
> > > > Considering the other case: B sends traffic from B1 to 
> A2. A replies
> > > > then from A1 to B1. This also means that your forwaring setup is
> > > > not in a proper way configured so that you can really 
> make use of the
> > > > multihoming support of SCTP. The problems are described in
> > > > 
> http://www.ietf.org/internet-drafts/draft-coene-sctp-multihome-03.txt
> > > > Nevertheless, when an SCTP endpoint receives an 
> INIT-ACK it should
> > > > check if it can find a TCB based on the source IP 
> address and the
> > > > IP addresses listed in the INIT-ACK. This seems not to 
> be implemented
> > > > in pre17 and will be added in a future release. This 
> would result
> > > > in being able to establish the association, but you do 
> not have the
> > > > redundancy you think based on the number of paths.
> > > >
> > > > Best regards
> > > > Michael
> > > >
> > > >
> > > > Michael Tuexen    Tel.:   +49 89 722 47210
> > > > Siemens AG        Fax:    +49 89 722 48212
> > > > ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
> > > >
> > > > > -----Original Message-----
> > > > > From:     Roland Gloeckler 
> [SMTP:gloeckle@ind.uni-stuttgart.de]
> > > > > Sent:     Thursday, February 21, 2002 1:13 PM
> > > > > To:       bidulock@openss7.org
> > > > > Cc:       discussion@sctp.de
> > > > > Subject:  Re: [SCTP-Discussion] Multihoming
> > > > >
> > > > > Dear Mr. Bidulock, dear all:
> > > > >
> > > > > My question is not completely answered by your response.
> > > > >
> > > > > My problem (a little bit more detailed) is the following:
> > > > >
> > > > > A has two interfaces with distinct IP addresses (A_1 and A_2).
> > > > > B sends an INIT to one of them, e.g., A_2. The other 
> interface A_1 is not
> > > > > known to B yet.
> > > > >
> > > > > In the sctp implementation A sends the INIT_ACK via 
> the interface which is
> > > > > the default entry in the IP routing table, e.g., A_1.
> > > > >
> > > > > Now there are two error scenarios:
> > > > >
> > > > > A has initialised a multihomed instance with both IP 
> addresses.
> > > > > For sending the INIT ACK it uses A_1 which is not 
> known to B before it reads
> > > > > the address list (with A_1 and A_2) in this INIT ACK chunk.
> > > > > B responds with an ABORT.
> > > > >
> > > > > or
> > > > >
> > > > > A has initialised a singlehomed instance with the 
> address A_2, which receives >
> > > > > the INIT.
> > > > > For sending the INIT ACK it uses A_1 which is not 
> even initialised in the
> > > > > SCTP instance. In the cookie it uses A_2.
> > > > > So it is obvious that SCTP has no control over the IP 
> source address.
> > > > > An association cannot be initialized.
> > > > >
> > > > > To resume:
> > > > >
> > > > > If the INIT_ACK chunk of an instance A uses in its IP 
> header a source IP
> > > > > address that is different from the destination 
> address of the INIT chunk from
> > > > > B, B refuses to accept it.
> > > > >
> > > > >
> > > > > Section 5.1.2.c of the RFC says:
> > > > >
> > > > > "IMPLEMENTATION NOTE: In some cases (e.g., when the 
> implementation doesn't
> > > > > control the source IP address that is used for 
> transmitting), an endpoint
> > > > > might need to include in its INIT or INIT ACK all 
> possible IP addresses from
> > > > > which packets to the peer could be transmitted."
> > > > >
> > > > > B would have to read the list of IP addresses in the 
> INIT ACK before it
> > > > > searches in its association list to know the chunk 
> belongs to A.
> > > > >
> > > > > Or should the implementation force its IP layer to 
> use the INIT destination
> > > > > address as the source address for the INIT ACK? The 
> RFC tries to cover cases
> > > > > where this is not the case.
> > > > >
> > > > > Regards,
> > > > >
> > > > > Roland
> > > > >
> > > > >
> > > > >
> > > > > On Friday 01 February 2002 20:38, Brian F. G. Bidulock wrote:
> > > > > > Roland,
> > > > > >
> > > > > > The effect your describe was discussed on TSVWG.  The
> > > > > > defficiency of SCTP was characterized as RFC2960's 
> inability>
> > > > > > to provide for reliable initialization of an association.
> > > > > >
> > > > > > The effect you describe is required by RFC 2960 section 5.1
> > > > > > paragraph (B): "... The destination IP address of 
> the INIT ACK
> > > > > > MUST be set to the source IP address of the INIT to 
> which this
> > > > > > INIT ACK is responding."
> > > > > >
> > > > > > The behavior exhibited by SCTPlib is correct.  The
> > > > > > implementation should reply with an ABORT in response to the
> > > > > > INIT ACK sent to the wrong address.
> > > > > >
> > > > > > --brian
> > > > > >
> > > > > > On Fri, 01 Feb 2002, Roland Gloeckler wrote:
> > > > > > > Dear all,
> > > > > > >
> > > > > > > when I was working with SCTPlib 1 pre14 I found an
> > > > > > > interesting behaviour.
> > > > > > >
> > > > > > > I had one PC (let's call it "A") with two network cards
> > > > > > > (let's call them A_1 and A_2) connected to another one
> > > > > > > (let's call it "B") with only one (B_1).
> > > > > > >
> > > > > > > The configuration was: Interfaces A_1 and B_1 
> were connected
> > > > > > > to the same LAN and A_2 was connected to the LAN via a
> > > > > > > router.
> > > > > > >
> > > > > > > After some minor changes to testengine.c I was able to use
> > > > > > > test_tool to initialise an SCTP multihoming instance on A
> > > > > > > for the addresses A_1 and A_2.
> > > > > > >
> > > > > > > In the script for B I tried to start an 
> association with A,
> > > > > > > but only included one IP address, e.g., A_2.
> > > > > > >
> > > > > > > B successfully sent an INIT message to A_2, which 
> responded
> > > > > > > with a valid INIT_ACK.
> > > > > > >
> > > > > > > If A responded via A_2 all went well.
> > > > > > >
> > > > > > > BUT:  When A used the other interface A_1 to send the
> > > > > > > INIT_ACK (which included the list of its own addresses A_1
> > > > > > > and A_2 belonging to that instance) to B, B was 
> not able to
> > > > > > > find the address A_1 in its association list, since it
> > > > > > > didn't know it yet, and answered with an ABORT 
> message to A.
> > > > > > >
> > > > > > > If I change the destination address in B's script 
> to A_1 and
> > > > > > > change the routing table in A to send packets to B via A_2
> > > > > > > the result is the same.
> > > > > > >
> > > > > > > To summarize:
> > > > > > >
> > > > > > > If the INIT_ACK chunk of a multihomed instance A uses a
> > > > > > > source IP address that is different from the one 
> B sent the
> > > > > > > INIT chunk to, B refuses to accept it.
> > > > > > >
> > > > > > > When I was looking into the RFC to search for 
> more details,
> > > > > > > I only found out that the behaviour is according to the>
> > > > > > > rules that no INIT_ACK is to be accepted which uses an
> > > > > > > unknown source transport address.
> > > > > > >
> > > > > > > Now I wonder if this is a general problem not answered by
> > > > > > > the RFC.
> > > > > > >
> > > > > > > I can think of several ways to come around the problem for
> > > > > > > my purposes, but they are not adequate for 
> general purposes.
> > > > > > >
> > > > > > > Has anyone heard of this situation before? Please tell me.
> > > > > > >
> > > > > > > Regards,
> > > > > > >
> > > > > > > Roland
> > > > > > > _______________________________________________
> > > > > > > discussion mailing list
> > > > > > > discussion@sctp.de
> > > > > > > http://www.sctp.de/mailman/listinfo/discussion
> > > > > _______________________________________________
> > > > > discussion mailing list
> > > > > discussion@sctp.de
> > > > > http://www.sctp.de/mailman/listinfo/discussion
> > >
> > > --
> > > Brian F. G. Bidulock    ¦ The reasonable man adapts 
> himself to the ¦
> > > bidulock@openss7.org    ¦ world; the unreasonable one 
> persists in  ¦
> > > http://www.openss7.org/ ¦ trying  to adapt the  world  to 
> himself. ¦
> > >                         ¦ Therefore  all  progress  
> depends on the ¦
> > >                         ¦ unreasonable man. -- George 
> Bernard Shaw ¦
> > 
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 21 16:25:38 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA04811
	for <tsvwg-archive@odin.ietf.org>; Thu, 21 Feb 2002 16:25:38 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA27246
	for tsvwg-archive@odin.ietf.org; Thu, 21 Feb 2002 16:25:42 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA26329;
	Thu, 21 Feb 2002 16:05:07 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA26299
	for <tsvwg@optimus.ietf.org>; Thu, 21 Feb 2002 16:05:00 -0500 (EST)
Received: from stewartlin.chicago.il.us (host-64-47-24-166.masergy.com [64.47.24.166] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA04171
	for <tsvwg@ietf.org>; Thu, 21 Feb 2002 16:04:55 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewartlin.chicago.il.us (8.11.0/8.11.0) with ESMTP id g1LL6er13672;
	Thu, 21 Feb 2002 15:06:40 -0600
Message-ID: <3C7560F0.5ADDF8F2@stewart.chicago.il.us>
Date: Thu, 21 Feb 2002 15:04:48 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: "Guo, Xingang" <xingang.guo@intel.com>
CC: Tuexen Michael <Michael.Tuexen@icn.siemens.de>,
        "'bidulock@openss7.org'" <bidulock@openss7.org>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "'discussion@sctp.de'" <discussion@sctp.de>
Subject: Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
References: <E9B3905226DDD411BC0F0090276D21C20C64A1AC@FMSMSX30>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

\
I hope it is.. La Monte and I talked about this
case MANY times :>

R

"Guo, Xingang" wrote:
> 
> Randall,
> 
> As far as I know, this is exactly what lksctp does.
> 
> www.sourceforge.net/projects/lksctp
> 
> --Xingang
> 
> > -----Original Message-----
> > From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> > Sent: Thursday, February 21, 2002 7:21 AM
> > To: Tuexen Michael
> > Cc: 'bidulock@openss7.org'; 'tsvwg@ietf.org'; 'discussion@sctp.de'
> > Subject: Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
> >
> >
> > Michael:
> >
> > From digging through this thread is this the scenario:
> >
> >
> > Endpoint 1                                           Endpoint 2
> > IP-A                                                    IP-C
> > IP-B                                                    IP-D
> > ------------Frm:IP-A:INIT(List-IP-A)--------->To:IP-C
> >     <---------------To:IP-A---Frm:IP-D:INIT-ACK(List-IP-C,IP-D)----
> >
> > Now the above case IS legal and there is specific wording in
> > the RFC that covers this. The idea is that Endpoint-1 MUST
> > lookup not only the FROM address in the IP header but also
> > MUST lookup in the list of addresses (IP-C/IP-D in this case)
> > if it can't find its TCB.
> >
> > R
> >
> > Tuexen Michael wrote:
> > >
> > > Brian,
> > >
> > > I know this. But I do not see why accepting the INIT-ACK
> > > coming from an unknown address, but the known one is
> > > in the list of addresses, is a serious problem, at least
> > > from the security point of view.
> > > As I mentioned, the assoc will not gain the expected
> > > network fault tolerance as it could be expected by only
> > > looking at the number of paths, but does that satisfy
> > > the non establishment of the assoc. It is sure that
> > > the network is not configured as it should, but ...
> > >
> > > Best regards
> > > Michael
> > >
> > > Michael Tuexen    Tel.:   +49 89 722 47210
> > > Siemens AG        Fax:    +49 89 722 48212
> > > ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
> > >
> > > > -----Original Message-----
> > > > From: Brian F. G. Bidulock [SMTP:bidulock@openss7.org]
> > > > Sent: Thursday, February 21, 2002 2:05 PM
> > > > To:   Tuexen Michael
> > > > Cc:   'discussion@sctp.de'
> > > > Subject:      Re: [SCTP-Discussion] Multihoming
> > > >
> > > > Michael,
> > > >
> > > > Not even TCP will accept such an arrangement.  Say TCP launches
> > > > a SYN from B to A_1.  If A responds with a SYN from A_2 to B the
> > > > connection will not succeed.
> > > >
> > > > --brian
> > > >
> > > > On Thu, 21 Feb 2002, Tuexen Michael wrote:
> > > >
> > > > > Dear Roland,
> > > > >
> > > > > considering the case where the SCTP end point A only announces
> > > > > A2, but A1 is used as a source IP address is definitly a
> > > > > misconfiguration of the host!
> > > > > If a SCTP end point only announces a proper subset of
> > its IP-addresses
> > > > > it must make sure that only addresses of this subset are used as
> > > > > source IP addresses. The most simple way to make this sure is to
> > > > > include all addresses. This is the reason why you can read the
> > > > > Implemention note in RFC 2960. The sctplib also
> > supports using this
> > > > > subset but then your forwaring on the host must be in
> > tune with this
> > > > > configuration as stated in the documentation.
> > > > >
> > > > > Considering the other case: B sends traffic from B1 to
> > A2. A replies
> > > > > then from A1 to B1. This also means that your forwaring setup is
> > > > > not in a proper way configured so that you can really
> > make use of the
> > > > > multihoming support of SCTP. The problems are described in
> > > > >
> > http://www.ietf.org/internet-drafts/draft-coene-sctp-multihome-03.txt
> > > > > Nevertheless, when an SCTP endpoint receives an
> > INIT-ACK it should
> > > > > check if it can find a TCB based on the source IP
> > address and the
> > > > > IP addresses listed in the INIT-ACK. This seems not to
> > be implemented
> > > > > in pre17 and will be added in a future release. This
> > would result
> > > > > in being able to establish the association, but you do
> > not have the
> > > > > redundancy you think based on the number of paths.
> > > > >
> > > > > Best regards
> > > > > Michael
> > > > >
> > > > >
> > > > > Michael Tuexen    Tel.:   +49 89 722 47210
> > > > > Siemens AG        Fax:    +49 89 722 48212
> > > > > ICN WN CS SE 5    E-mail: Michael.Tuexen@icn.siemens.de
> > > > >
> > > > > > -----Original Message-----
> > > > > > From:     Roland Gloeckler
> > [SMTP:gloeckle@ind.uni-stuttgart.de]
> > > > > > Sent:     Thursday, February 21, 2002 1:13 PM
> > > > > > To:       bidulock@openss7.org
> > > > > > Cc:       discussion@sctp.de
> > > > > > Subject:  Re: [SCTP-Discussion] Multihoming
> > > > > >
> > > > > > Dear Mr. Bidulock, dear all:
> > > > > >
> > > > > > My question is not completely answered by your response.
> > > > > >
> > > > > > My problem (a little bit more detailed) is the following:
> > > > > >
> > > > > > A has two interfaces with distinct IP addresses (A_1 and A_2).
> > > > > > B sends an INIT to one of them, e.g., A_2. The other
> > interface A_1 is not
> > > > > > known to B yet.
> > > > > >
> > > > > > In the sctp implementation A sends the INIT_ACK via
> > the interface which is
> > > > > > the default entry in the IP routing table, e.g., A_1.
> > > > > >
> > > > > > Now there are two error scenarios:
> > > > > >
> > > > > > A has initialised a multihomed instance with both IP
> > addresses.
> > > > > > For sending the INIT ACK it uses A_1 which is not
> > known to B before it reads
> > > > > > the address list (with A_1 and A_2) in this INIT ACK chunk.
> > > > > > B responds with an ABORT.
> > > > > >
> > > > > > or
> > > > > >
> > > > > > A has initialised a singlehomed instance with the
> > address A_2, which receives >
> > > > > > the INIT.
> > > > > > For sending the INIT ACK it uses A_1 which is not
> > even initialised in the
> > > > > > SCTP instance. In the cookie it uses A_2.
> > > > > > So it is obvious that SCTP has no control over the IP
> > source address.
> > > > > > An association cannot be initialized.
> > > > > >
> > > > > > To resume:
> > > > > >
> > > > > > If the INIT_ACK chunk of an instance A uses in its IP
> > header a source IP
> > > > > > address that is different from the destination
> > address of the INIT chunk from
> > > > > > B, B refuses to accept it.
> > > > > >
> > > > > >
> > > > > > Section 5.1.2.c of the RFC says:
> > > > > >
> > > > > > "IMPLEMENTATION NOTE: In some cases (e.g., when the
> > implementation doesn't
> > > > > > control the source IP address that is used for
> > transmitting), an endpoint
> > > > > > might need to include in its INIT or INIT ACK all
> > possible IP addresses from
> > > > > > which packets to the peer could be transmitted."
> > > > > >
> > > > > > B would have to read the list of IP addresses in the
> > INIT ACK before it
> > > > > > searches in its association list to know the chunk
> > belongs to A.
> > > > > >
> > > > > > Or should the implementation force its IP layer to
> > use the INIT destination
> > > > > > address as the source address for the INIT ACK? The
> > RFC tries to cover cases
> > > > > > where this is not the case.
> > > > > >
> > > > > > Regards,
> > > > > >
> > > > > > Roland
> > > > > >
> > > > > >
> > > > > >
> > > > > > On Friday 01 February 2002 20:38, Brian F. G. Bidulock wrote:
> > > > > > > Roland,
> > > > > > >
> > > > > > > The effect your describe was discussed on TSVWG.  The
> > > > > > > defficiency of SCTP was characterized as RFC2960's
> > inability>
> > > > > > > to provide for reliable initialization of an association.
> > > > > > >
> > > > > > > The effect you describe is required by RFC 2960 section 5.1
> > > > > > > paragraph (B): "... The destination IP address of
> > the INIT ACK
> > > > > > > MUST be set to the source IP address of the INIT to
> > which this
> > > > > > > INIT ACK is responding."
> > > > > > >
> > > > > > > The behavior exhibited by SCTPlib is correct.  The
> > > > > > > implementation should reply with an ABORT in response to the
> > > > > > > INIT ACK sent to the wrong address.
> > > > > > >
> > > > > > > --brian
> > > > > > >
> > > > > > > On Fri, 01 Feb 2002, Roland Gloeckler wrote:
> > > > > > > > Dear all,
> > > > > > > >
> > > > > > > > when I was working with SCTPlib 1 pre14 I found an
> > > > > > > > interesting behaviour.
> > > > > > > >
> > > > > > > > I had one PC (let's call it "A") with two network cards
> > > > > > > > (let's call them A_1 and A_2) connected to another one
> > > > > > > > (let's call it "B") with only one (B_1).
> > > > > > > >
> > > > > > > > The configuration was: Interfaces A_1 and B_1
> > were connected
> > > > > > > > to the same LAN and A_2 was connected to the LAN via a
> > > > > > > > router.
> > > > > > > >
> > > > > > > > After some minor changes to testengine.c I was able to use
> > > > > > > > test_tool to initialise an SCTP multihoming instance on A
> > > > > > > > for the addresses A_1 and A_2.
> > > > > > > >
> > > > > > > > In the script for B I tried to start an
> > association with A,
> > > > > > > > but only included one IP address, e.g., A_2.
> > > > > > > >
> > > > > > > > B successfully sent an INIT message to A_2, which
> > responded
> > > > > > > > with a valid INIT_ACK.
> > > > > > > >
> > > > > > > > If A responded via A_2 all went well.
> > > > > > > >
> > > > > > > > BUT:  When A used the other interface A_1 to send the
> > > > > > > > INIT_ACK (which included the list of its own addresses A_1
> > > > > > > > and A_2 belonging to that instance) to B, B was
> > not able to
> > > > > > > > find the address A_1 in its association list, since it
> > > > > > > > didn't know it yet, and answered with an ABORT
> > message to A.
> > > > > > > >
> > > > > > > > If I change the destination address in B's script
> > to A_1 and
> > > > > > > > change the routing table in A to send packets to B via A_2
> > > > > > > > the result is the same.
> > > > > > > >
> > > > > > > > To summarize:
> > > > > > > >
> > > > > > > > If the INIT_ACK chunk of a multihomed instance A uses a
> > > > > > > > source IP address that is different from the one
> > B sent the
> > > > > > > > INIT chunk to, B refuses to accept it.
> > > > > > > >
> > > > > > > > When I was looking into the RFC to search for
> > more details,
> > > > > > > > I only found out that the behaviour is according to the>
> > > > > > > > rules that no INIT_ACK is to be accepted which uses an
> > > > > > > > unknown source transport address.
> > > > > > > >
> > > > > > > > Now I wonder if this is a general problem not answered by
> > > > > > > > the RFC.
> > > > > > > >
> > > > > > > > I can think of several ways to come around the problem for
> > > > > > > > my purposes, but they are not adequate for
> > general purposes.
> > > > > > > >
> > > > > > > > Has anyone heard of this situation before? Please tell me.
> > > > > > > >
> > > > > > > > Regards,
> > > > > > > >
> > > > > > > > Roland
> > > > > > > > _______________________________________________
> > > > > > > > discussion mailing list
> > > > > > > > discussion@sctp.de
> > > > > > > > http://www.sctp.de/mailman/listinfo/discussion
> > > > > > _______________________________________________
> > > > > > discussion mailing list
> > > > > > discussion@sctp.de
> > > > > > http://www.sctp.de/mailman/listinfo/discussion
> > > >
> > > > --
> > > > Brian F. G. Bidulock    ¦ The reasonable man adapts
> > himself to the ¦
> > > > bidulock@openss7.org    ¦ world; the unreasonable one
> > persists in  ¦
> > > > http://www.openss7.org/ ¦ trying  to adapt the  world  to
> > himself. ¦
> > > >                         ¦ Therefore  all  progress
> > depends on the ¦
> > > >                         ¦ unreasonable man. -- George
> > Bernard Shaw ¦
> > >
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> >
> > --
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> >

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 21 16:39:31 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA05112
	for <tsvwg-archive@odin.ietf.org>; Thu, 21 Feb 2002 16:39:30 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA28152
	for tsvwg-archive@odin.ietf.org; Thu, 21 Feb 2002 16:39:34 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA26968;
	Thu, 21 Feb 2002 16:21:17 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA26891
	for <tsvwg@optimus.ietf.org>; Thu, 21 Feb 2002 16:21:12 -0500 (EST)
Received: from localhost ([211.226.155.235])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id QAA04705
	for <tsvwg@ietf.org>; Thu, 21 Feb 2002 16:21:02 -0500 (EST)
Message-Id: <200202212121.QAA04705@ietf.org>
Reply-To: gtech01@gtech21.com
From: (ÁÖ)ÁöÅØ<gtech01@gtech21.com>
To: tsvwg@ietf.org
Mime-Version: 1.0
Content-Type: text/html; charset="ks_c_5601-1987"
Date: Fri, 22 Feb 2002 06:20:25 +0900
Subject: [Tsvwg] È¹±âÀûÀÎ ÈÞ´ëÆù¿ë ÀÌ¾îÆù !!!      [±¤°í¸ÞÀÏ]
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

<!-- saved from url=(0022)http://internet.e-mail -->
<HTML>
<HEAD>
<TITLE>mail02-1</TITLE>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=euc-kr">
<script language="JavaScript">
<!--
function na_open_window(name, url, left, top, width, height, toolbar, menubar, statusbar, scrollbar, resizable)
{
toolbar_str = toolbar ? 'yes' : 'no';
menubar_str = menubar ? 'yes' : 'no';
statusbar_str = statusbar ? 'yes' : 'no';
scrollbar_str = scrollbar ? 'yes' : 'no';
resizable_str = resizable ? 'yes' : 'no';
window.open(url, name, 'left='+left+',top='+top+',width='+width+',height='+height+',toolbar='+toolbar_str+',menubar='+menubar_str+',status='+statusbar_str+',scrollbars='+scrollbar_str+',resizable='+resizable_str);
}
// -->
</script>
<script language='javascript'>
<!-------
function news1(url){
window.open(url,'newwindow','toolbar=no,location=no,directories=no,status=no,menubar=no,scrollbars=yes,resizable=no,width=580,height=600');
};
----------->
</script>
<style>
<!--
input{ font-size: 9pt;border:0px solid;border-color:"balck"}
//-->
</style>
<style>
<!--
td { font-size : 9pt;   }
A:link  { font : 9pt;  	color : black; 	text-decoration : none;	font-family : ±¼¸²;	font-size : 9pt;  }
A:visited  {   text-decoration : none; color : black;	font-size : 9pt;  }
A:hover  { 	text-decoration : none; color : B787B5 ; font-size : 9pt;  }
-->
</style>
</HEAD>
<BODY BGCOLOR=#FFFFFF>
<TABLE WIDTH=800 BORDER=0 CELLPADDING=0 CELLSPACING=0 >
<TR>
<TD width=57>
</TD>
<TD bgcolor="EEEEEF" WIDTH=578 HEIGHT=71 style=padding-left:4px>
<font style=font-size:9pt color=black>
¢Á Çã¶ô¾øÀÌ ¸ÞÀÏÀ» º¸³»¼­ ÁË¼ÛÇÕ´Ï´Ù.<br>
¢Á º»¸ÞÀÏÀº Á¤º¸Åë½Å¸Á ÀÌ¿ëÃËÁø ¹× Á¤º¸º¸È£ µî¿¡ °üÇÑ ¹ý·ü Á¦ 50Á¶¿¡ ÀÇ°ÅÇÑ [±¤°í] ¸ÞÀÏÀÔ´Ï´Ù.<br>
¢Á e-mail ÁÖ¼Ò´Â ÀÎÅÍ³Ý»ó¿¡¼­ ÃëµæÇÏ¿´À¸¸ç, ÁÖ¼Ò¿Ü ¾î¶°ÇÑ °³ÀÎ Á¤º¸µµ °¡Áö°í ÀÖÁö ¾Ê½À´Ï´Ù.<br>
¢Á ¼ö½ÅÀ» ¿øÄ¡ ¾ÊÀ¸½Ã¸é '¼ö½Å°ÅºÎ'¸¦ ´­·¯ÁÖ¼¼¿ä.
<A href="mailto:gtech@gtech21.com?subject=tsvwg@ietf.orgÀÇ »ç¿ëÀÚ·Î½á ¼ö½ÅÀ» °ÅºÎÇÕ´Ï´Ù. &amp;body=tsvwg@ietf.orgÀ» ±ÍÇÏÀÇ ¸®½ºÆ®¿¡¼­ »èÁ¦¿ä¸Á!">[¼ö½Å°ÅºÎ]</A>
</font>
</TD>
<TD WIDTH=164 >
</TD>
</TR>
</TABLE>
<!-- ImageReady Slices (mail02-1.gif) -->
<TABLE WIDTH=800 BORDER=0 CELLPADDING=0 CELLSPACING=0>
<TR>
<TD WIDTH=800 HEIGHT=136>
<TABLE WIDTH=800 BORDER=0 CELLPADDING=0 CELLSPACING=0>
<TR>
<TD WIDTH=57 HEIGHT=136>&nbsp;</TD>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_01_02.gif" WIDTH=394 HEIGHT=136></TD>
<TD WIDTH=178 HEIGHT=136>
<!-- ImageReady Slices (mail02-1_01_03.gif) -->
<TABLE WIDTH=178 BORDER=0 CELLPADDING=0 CELLSPACING=0>
<TR>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_01_03_01.gif" WIDTH=178 HEIGHT=65></TD>
</TR>
<TR>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_01_03_02.gif" WIDTH=178 HEIGHT=44></TD> <!--Moving gif-->
</TR>
<TR>
<TD>
<A HREF="http://www.gtech21.com"><IMG SRC="http://www.gtech21.com/order/images/mail02-1_01_03_03.gif" WIDTH=178 HEIGHT=27 border=0></A></TD>
</TR>
</TABLE>
<!-- End ImageReady Slices -->
</TD>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_01_04.gif" WIDTH=9 HEIGHT=136></TD>
<TD WIDTH=162 HEIGHT=136>&nbsp;</TD>
</TR>
</TABLE>
</TD>
</TR>
<TR>
<TD WIDTH=800 HEIGHT=536>
<!-- ImageReady Slices (mail02-1_02.gif) -->
<TABLE WIDTH=800 BORDER=0 CELLPADDING=0 CELLSPACING=0>
<TR>
<TD WIDTH=57 HEIGHT=536>&nbsp;</TD>
<TD WIDTH=400 HEIGHT=536>
<!-- ImageReady Slices (mail02-1_02_02.gif) -->
<TABLE WIDTH=400 BORDER=0 CELLPADDING=0 CELLSPACING=0>
<TR>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_02_01.gif" WIDTH=400 HEIGHT=370></TD>
</TR>
<TR>
<TD WIDTH=400 HEIGHT=117>
<!-- ImageReady Slices (mail02-1_02_02_02.gif) -->
<TABLE WIDTH=400 BORDER=0 CELLPADDING=0 CELLSPACING=0>
<TR>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_02_02_01.gif" WIDTH=47 HEIGHT=117></TD>
<TD>
<A HREF="http://www.gtech21.com/template.php?Module=product&SubPage=2"><IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_02_02_02.gif" WIDTH=147 HEIGHT=117 border=0></A></TD>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_02_02_03.gif" WIDTH=30 HEIGHT=117></TD>
<TD>
<A HREF="http://www.gtech21.com/template.php?Module=product&SubPage=2"><IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_02_02_04.gif" WIDTH=129 HEIGHT=117 border=0></A></TD>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_02_02_05.gif" WIDTH=47 HEIGHT=117></TD>
</TR>
</TABLE>
<!-- End ImageReady Slices(mail02-1_02_02_02.gif) -->
</TD>
</TR>
<TR>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_02_03.gif" WIDTH=400 HEIGHT=49></TD>
</TR>
</TABLE>
<!-- End ImageReady Slices --(mail02-1_02_02.gif)-->
</TD>
<TD WIDTH=151 HEIGHT=536>
<!-- ImageReady Slices (mail02-1_02_03.gif) -->
<TABLE WIDTH=151 BORDER=0 CELLPADDING=0 CELLSPACING=0>
<TR>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_03_01.gif" WIDTH=151 HEIGHT=10></TD>
</TR>
<TR>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_03_02.gif" WIDTH=151 HEIGHT=59 border=0></TD>
</TR>
<TR>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_03_03.gif" WIDTH=151 HEIGHT=5></TD>
</TR>
<TR>
<TD>           <!--µ¿¿µ»óº¸±â ¸µÅ©°Å´Â°÷-->
<a href="http://www.gtech21.com/order/gtech3/moving.php"><IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_03_04.gif" WIDTH=151 HEIGHT=383 border=0></a></TD>
</TR>
<TR>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_03_05.gif" WIDTH=151 HEIGHT=14></TD>
</TR>
<TR>
<TD WIDTH=151 HEIGHT=27>
<!-- ImageReady Slices (mail02-1_02_03_06.gif) --> <!--µ¿¿µ»óº¸±â ¹öÆ° ¸µÅ©°Å´Â°÷-->
<TABLE WIDTH=151 BORDER=0 CELLPADDING=0 CELLSPACING=0>
<TR>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_03_06_01.gif" WIDTH=9 HEIGHT=27></TD>
<TD>
<a href="http://www.gtech21.com/order/gtech3/moving.php">
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_03_06_02.gif" WIDTH=135 HEIGHT=27 border=0></a></TD>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_03_06_03.gif" WIDTH=7 HEIGHT=27></TD>
</TR>
</TABLE>
<!-- End ImageReady Slices (mail02-1_02_03_06.gif) -->
</TD>
</TR>
<TR>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_03_07.gif" WIDTH=151 HEIGHT=4></TD>
</TR>
<TR>
<TD WIDTH=151 HEIGHT=27>
<!-- ImageReady Slices (mail02-1_02_03_08.gif) --> <!--±¸¸ÅÇÏ±â ¹öÆ° ¸µÅ©°Å´Â°÷-->
<TABLE WIDTH=151 BORDER=0 CELLPADDING=0 CELLSPACING=0>
<TR>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_03_08_01.gif" WIDTH=9 HEIGHT=27></TD>
<TD>
<A HREF="http://www.gtech21.com/order/orderframe.php3"><IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_03_08_02.gif" WIDTH=135 HEIGHT=27 border=0></A></TD>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_03_08_03.gif" WIDTH=7 HEIGHT=27></TD>
</TR>
</TABLE>
<!-- End ImageReady Slices (mail02-1_02_03_08.gif) -->
</TD>
</TR>
<TR>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_03_09.gif" WIDTH=151 HEIGHT=7></TD>
</TR>
</TABLE>
<!-- End ImageReady Slices (mail02-1_02_03.gif)  -->
</TD>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_02_04.gif" WIDTH=30 HEIGHT=536></TD>
<TD WIDTH=162 HEIGHT=536>&nbsp;</TD>
</TR>
</TABLE>
<!-- End ImageReady Slices (mail02-1_02.gif) -->
</TD>
</TR>
<TR>
<TD WIDTH=800 HEIGHT=41>
<!-- ImageReady Slices (mail02-1_04.gif) -->
<TABLE WIDTH=800 BORDER=0 CELLPADDING=0 CELLSPACING=0>
<TR>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_04_01.gif" WIDTH=60 HEIGHT=41></TD>
<TD WIDTH=575 HEIGHT=41 align=center bgcolor="">
<font style=font-size:9pt color=black>
<b>(ÁÖ)ÁöÅØ</b> ÀÎÃµ±¤¿ª½Ã °è¾ç±¸ ÀÛÀüµ¿ 852-73 Tel:(032)552-7222<br>
E-mail : <A HREF="mailto:gtech@gtech21.com">gtech@gtech21.com</A>&nbsp;&nbsp;&nbsp; Homepage : <A HREF="http://www.gtech21.com">http://www.gtech21.com</A>
</font>
</TD>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_04_03.gif" WIDTH=165 HEIGHT=41></TD>
</TR>
</TABLE>
<!-- End ImageReady Slices (mail02-1_04.gif) -->
</TD>
</TR>
<TR>
<TD WIDTH=800 HEIGHT=3>
<!-- ImageReady Slices (mail02-1_05.gif) -->
<TABLE WIDTH=800 BORDER=0 CELLPADDING=0 CELLSPACING=0>
<TR>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_05_01.gif" WIDTH=57 HEIGHT=3></TD>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_05_02.gif" WIDTH=581 HEIGHT=3></TD>
<TD>
<IMG SRC="http://www.gtech21.com/order/images/mail02-1_05_03.gif" WIDTH=162 HEIGHT=3></TD>
</TR>
</TABLE>
<!-- End ImageReady Slices (mail02-1_05.gif) -->
</TD>
</TR>
<TR>
<TD WIDTH=800 HEIGHT=13>
&nbsp;
</TD>
</TR>
</TABLE>
<!-- End ImageReady Slices -->
</BODY>
</HTML>

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Thu Feb 21 22:50:29 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA12494
	for <tsvwg-archive@odin.ietf.org>; Thu, 21 Feb 2002 22:50:25 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id WAA14079
	for tsvwg-archive@odin.ietf.org; Thu, 21 Feb 2002 22:50:29 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA13820;
	Thu, 21 Feb 2002 22:39:37 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA13791
	for <tsvwg@ns.ietf.org>; Thu, 21 Feb 2002 22:39:34 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-216.dallas.net [209.44.42.216])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA12173
	for <tsvwg@ietf.org>; Thu, 21 Feb 2002 22:39:29 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id VAA09628;
	Thu, 21 Feb 2002 21:39:18 -0600
Date: Thu, 21 Feb 2002 21:39:18 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <randall@stewart.chicago.il.us>
Cc: Tuexen Michael <Michael.Tuexen@icn.siemens.de>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>,
        "'discussion@sctp.de'" <discussion@sctp.de>
Subject: Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
Message-ID: <20020221213918.A9608@openss7.org>
Reply-To: bidulock@openss7.org
References: <F92978223B01D311ACFF0008C71EE19D015DF788@MCHH225E> <3C75104E.53BC132F@stewart.chicago.il.us>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <3C75104E.53BC132F@stewart.chicago.il.us>; from randall@stewart.chicago.il.us on Thu, Feb 21, 2002 at 09:20:46AM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Randall,

On Thu, 21 Feb 2002, Randall Stewart wrote:

> Now the above case IS legal and there is specific wording in
> the RFC that covers this.

Could you point to the section of the RFC that requires
that this scenario be permitted?

--brian



-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 22 06:04:38 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA26645
	for <tsvwg-archive@odin.ietf.org>; Fri, 22 Feb 2002 06:04:38 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA13779
	for tsvwg-archive@odin.ietf.org; Fri, 22 Feb 2002 06:04:41 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA13002;
	Fri, 22 Feb 2002 05:54:01 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA12974
	for <tsvwg@ns.ietf.org>; Fri, 22 Feb 2002 05:53:58 -0500 (EST)
Received: from mgw-x1.nokia.com (mgw-x1.nokia.com [131.228.20.21])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA26494
	for <tsvwg@ietf.org>; Fri, 22 Feb 2002 05:53:54 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir05nok.ntc.nokia.com (esvir05nokt.ntc.nokia.com [172.21.143.37])
	by mgw-x1.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1MArQ321099
	for <tsvwg@ietf.org>; Fri, 22 Feb 2002 12:53:26 +0200 (EET)
Received: from esebh002.NOE.Nokia.com (unverified) by esvir05nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T593a504f8dac158f25144a@esvir05nok.ntc.nokia.com>;
 Fri, 22 Feb 2002 12:53:51 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh002.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Fri, 22 Feb 2002 12:53:50 +0200
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] RE: [SCTP-Discussion] Multihoming
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
Date: Fri, 22 Feb 2002 12:53:50 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF26308EECA@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] RE: [SCTP-Discussion] Multihoming
Thread-Index: AcG7Up2AdBXgLb/dQ+eaA0IOpkHw/AAO6GXw
To: <bidulock@openss7.org>, <randall@stewart.chicago.il.us>
Cc: <Michael.Tuexen@icn.siemens.de>, <tsvwg@ietf.org>, <discussion@sctp.de>
X-OriginalArrivalTime: 22 Feb 2002 10:53:50.0894 (UTC) FILETIME=[364EDCE0:01C1BB8F]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id FAA12976
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi all!

	I also consider that situation to be legal. When I receive an INIT ACK I always look inside the parameters to see if it belongs to any open association (or in the COOKIE WAIT state). Moreover, I also do it with INIT chunks, to realize about establishment collisions or restart...

	If you don't allow this, I think it would be quite hard to detect INIT collisions in multihomed hosts... :-/

	I must admit that I forgot about this possibility until after having tested my implementation in a "brakeoff". But, could you Brian point to the section of the RFC that forbids this possibility? In my understanding it is a logical thing to do...

	Thanks!

	BR Iván Arias Rodríguez

> -----Original Message-----
> From: ext Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> Sent: 22. February 2002 5:39
> To: Randall Stewart
> Cc: Tuexen Michael; 'tsvwg@ietf.org'; 'discussion@sctp.de'
> Subject: Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
> 
> 
> Randall,
> 
> On Thu, 21 Feb 2002, Randall Stewart wrote:
> 
> > Now the above case IS legal and there is specific wording in
> > the RFC that covers this.
> 
> Could you point to the section of the RFC that requires
> that this scenario be permitted?
> 
> --brian
> 
> 
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 22 07:29:35 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA29554
	for <tsvwg-archive@odin.ietf.org>; Fri, 22 Feb 2002 07:29:34 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA18713
	for tsvwg-archive@odin.ietf.org; Fri, 22 Feb 2002 07:29:36 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA17594;
	Fri, 22 Feb 2002 07:14:53 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA17566
	for <tsvwg@optimus.ietf.org>; Fri, 22 Feb 2002 07:14:51 -0500 (EST)
Received: from stewartlin.chicago.il.us (host-64-47-24-166.masergy.com [64.47.24.166] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA28939
	for <tsvwg@ietf.org>; Fri, 22 Feb 2002 07:14:48 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewartlin.chicago.il.us (8.11.0/8.11.0) with ESMTP id g1MCGSr15245;
	Fri, 22 Feb 2002 06:16:28 -0600
Message-ID: <3C763630.97B78B39@stewart.chicago.il.us>
Date: Fri, 22 Feb 2002 06:14:40 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: bidulock@openss7.org, Michael.Tuexen@icn.siemens.de, tsvwg@ietf.org,
        discussion@sctp.de
Subject: Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
References: <58AEDE27D48F884285007ABF827FF26308EECA@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Ivan:

Well you hit the nail on the head as ususal... without
doing this setup between multi-homed hosts does not
work very well... consider

Host-A                                   Host-Z
IP-1                                     IP-5
IP-2                                     IP-6

---INIT(IP-1,IP-2)---\ /--INIT(IP-5,IP-6)---
                      X
toIP-1<-frm-IP-5-----/\------frm-IP-2------>toIP-6

Now unless inspection is done of the INIT/INIT-ACK the
two will NOT realize they are both setting up a association 
at the same time.

Another consideration is Address filtering. Often times
a host MUST use the IP address of the interface that you
are sending from.. or get the packet filtered so in the above

----Frm-IP-1:INIT(IP-1,IP-2)------to(IP-5)----------->
                                               (route to IP-1 points out
intf
                                                of IP6.. note this may
be
                                                from either only one
allowed
                                                default route.. or some
other
                                                reason).
 <----------to(IP-1)--INIT-ACK(IP-5,IP-6)--------From-IP-6

And we have the same secenario..

As far as looking for this text in the RFC .. sadly the full text
was chopped in one of the IESG passes (due to my lack of communication
skills as to why it is there)... and now all that is left is the
trival little statement of:

"
      IMPLEMENTATION NOTE: In some cases (e.g., when the implementation
      doesn't control the source IP address that is used for
      transmitting), an endpoint might need to include in its INIT or
      INIT ACK all possible IP addresses from which packets to the peer
      could be transmitted.
"

I think we will need to add a few lines to the implementors guide...
Also
there is NOTHING in the spec that makes this illegal as you point out.
I remember we discussed this amongst the many emails and I addressed
just
these points.

Well more grist for the implementors guide :>

R

Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Hi all!
> 
>         I also consider that situation to be legal. When I receive an INIT ACK I always look inside the parameters to see if it belongs to any open association (or in the COOKIE WAIT state). Moreover, I also do it with INIT chunks, to realize about establishment collisions or restart...
> 
>         If you don't allow this, I think it would be quite hard to detect INIT collisions in multihomed hosts... :-/
> 
>         I must admit that I forgot about this possibility until after having tested my implementation in a "brakeoff". But, could you Brian point to the section of the RFC that forbids this possibility? In my understanding it is a logical thing to do...
> 
>         Thanks!
> 
>         BR Iván Arias Rodríguez
> 
> > -----Original Message-----
> > From: ext Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > Sent: 22. February 2002 5:39
> > To: Randall Stewart
> > Cc: Tuexen Michael; 'tsvwg@ietf.org'; 'discussion@sctp.de'
> > Subject: Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
> >
> >
> > Randall,
> >
> > On Thu, 21 Feb 2002, Randall Stewart wrote:
> >
> > > Now the above case IS legal and there is specific wording in
> > > the RFC that covers this.
> >
> > Could you point to the section of the RFC that requires
> > that this scenario be permitted?
> >
> > --brian
> >
> >
> >
> > --
> > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> >
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> >

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 22 09:21:51 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA03082
	for <tsvwg-archive@odin.ietf.org>; Fri, 22 Feb 2002 09:21:51 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA24390
	for tsvwg-archive@odin.ietf.org; Fri, 22 Feb 2002 09:21:51 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA23907;
	Fri, 22 Feb 2002 09:07:56 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA23882
	for <tsvwg@optimus.ietf.org>; Fri, 22 Feb 2002 09:07:53 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-216.dallas.net [209.44.42.216])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA02485
	for <tsvwg@ietf.org>; Fri, 22 Feb 2002 09:07:47 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id IAA16850;
	Fri, 22 Feb 2002 08:07:29 -0600
Date: Fri, 22 Feb 2002 08:07:29 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <randall@stewart.chicago.il.us>
Cc: Ivan.Arias-Rodriguez@nokia.com, Michael.Tuexen@icn.siemens.de,
        tsvwg@ietf.org, discussion@sctp.de
Subject: Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
Message-ID: <20020222080729.A16506@openss7.org>
Reply-To: bidulock@openss7.org
References: <58AEDE27D48F884285007ABF827FF26308EECA@esebe017.NOE.Nokia.com> <3C763630.97B78B39@stewart.chicago.il.us>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
User-Agent: Mutt/1.2i
In-Reply-To: <3C763630.97B78B39@stewart.chicago.il.us>; from randall@stewart.chicago.il.us on Fri, Feb 22, 2002 at 06:14:40AM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id JAA23883
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Randall,

This is a different situation altogether that you are
talking about.  What you describe is of course already
handled in most implementations I believe.

The RFC says the INIT ACK MUST be sent to the source
address of the INIT.

We are questioning whether it is required to send the
INIT ACK *from* the destination address of the INIT,
or whether it is permitted to send them from any
element in the list.

IMO the INIT process (INIT INIT-ACK exchange) is
restricted to pair of transport endpoints (the one
sending the INIT and the one replying with an INIT-ACK).

There is absolutely no need to inject other IP addresses
into the IP header (source or destination address).  Multi-
homed addresses from which the INIT or INIT-ACK are sent
should appear at most in the address list in the chunk.

If you want to permit the behavior that the sender of
the INIT-ACK can use any old source address in the list,
there would be a conflict with the RFC restriction that
the INIT-ACK "MUST" be sent to the sender (source address)
of the INIT.

However, we discussed this before and it was the questioned
whether opening up the INIT/INIT-ACK exchange to the entire
address lists would not somehow weaken the secuity of the
INIT process.

If the INIT-ACK can come from any source address in the
list, where is the COOKIE-ECHO then sent?

--brian

On Fri, 22 Feb 2002, Randall Stewart wrote:

> Ivan:
> 
> Well you hit the nail on the head as ususal... without
> doing this setup between multi-homed hosts does not
> work very well... consider
> 
> Host-A                                   Host-Z
> IP-1                                     IP-5
> IP-2                                     IP-6
> 
> ---INIT(IP-1,IP-2)---\ /--INIT(IP-5,IP-6)---
>                       X
> toIP-1<-frm-IP-5-----/\------frm-IP-2------>toIP-6
> 
> Now unless inspection is done of the INIT/INIT-ACK the
> two will NOT realize they are both setting up a association 
> at the same time.
> 
> Another consideration is Address filtering. Often times
> a host MUST use the IP address of the interface that you
> are sending from.. or get the packet filtered so in the above
> 
> ----Frm-IP-1:INIT(IP-1,IP-2)------to(IP-5)----------->
>                                                (route to IP-1 points out
> intf
>                                                 of IP6.. note this may
> be
>                                                 from either only one
> allowed
>                                                 default route.. or some
> other
>                                                 reason).
>  <----------to(IP-1)--INIT-ACK(IP-5,IP-6)--------From-IP-6
> 
> And we have the same secenario..
> 
> As far as looking for this text in the RFC .. sadly the full text
> was chopped in one of the IESG passes (due to my lack of communication
> skills as to why it is there)... and now all that is left is the
> trival little statement of:
> 
> "
>       IMPLEMENTATION NOTE: In some cases (e.g., when the implementation
>       doesn't control the source IP address that is used for
>       transmitting), an endpoint might need to include in its INIT or
>       INIT ACK all possible IP addresses from which packets to the peer
>       could be transmitted.
> "
> 
> I think we will need to add a few lines to the implementors guide...
> Also
> there is NOTHING in the spec that makes this illegal as you point out.
> I remember we discussed this amongst the many emails and I addressed
> just
> these points.
> 
> Well more grist for the implementors guide :>
> 
> R
> 
> Ivan.Arias-Rodriguez@nokia.com wrote:
> > 
> >         Hi all!
> > 
> >         I also consider that situation to be legal. When I receive an INIT
> >         ACK I always look inside the parameters to see if it belongs to
> >         any open association (or in the COOKIE WAIT state). Moreover, I
> >         also do it with INIT chunks, to realize about establishment
> >         collisions or restart...
> > 
> >         If you don't allow this, I think it would be quite hard to detect
> >         INIT collisions in multihomed hosts... :-/
> > 
> >         I must admit that I forgot about this possibility until after
> >         having tested my implementation in a "brakeoff". But, could you
> >         Brian point to the section of the RFC that forbids this
> >         possibility? In my understanding it is a logical thing to do...
> > 
> >         Thanks!
> > 
> >         BR Iván Arias Rodríguez
> > 
> > > -----Original Message-----
> > > From: ext Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > > Sent: 22. February 2002 5:39
> > > To: Randall Stewart
> > > Cc: Tuexen Michael; 'tsvwg@ietf.org'; 'discussion@sctp.de'
> > > Subject: Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
> > >
> > >
> > > Randall,
> > >
> > > On Thu, 21 Feb 2002, Randall Stewart wrote:
> > >
> > > > Now the above case IS legal and there is specific wording in
> > > > the RFC that covers this.
> > >
> > > Could you point to the section of the RFC that requires
> > > that this scenario be permitted?
> > >
> > > --brian
> > >
> > >
> > >
> > > --
> > > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> > >                         ¦ Therefore  all  progress  depends on the ¦
> > >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> > >
> > > _______________________________________________
> > > tsvwg mailing list
> > > tsvwg@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > >
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 22 09:40:35 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA03771
	for <tsvwg-archive@odin.ietf.org>; Fri, 22 Feb 2002 09:40:34 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA25859
	for tsvwg-archive@odin.ietf.org; Fri, 22 Feb 2002 09:40:38 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA24372;
	Fri, 22 Feb 2002 09:21:37 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA24343
	for <tsvwg@optimus.ietf.org>; Fri, 22 Feb 2002 09:21:35 -0500 (EST)
Received: from mgw-x3.nokia.com (mgw-x3.nokia.com [131.228.20.26])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA03044
	for <tsvwg@ietf.org>; Fri, 22 Feb 2002 09:21:30 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir01nok.ntc.nokia.com (esvir01nokt.ntc.nokia.com [172.21.143.33])
	by mgw-x3.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1MEM2i21507
	for <tsvwg@ietf.org>; Fri, 22 Feb 2002 16:22:02 +0200 (EET)
Received: from esebh001.NOE.Nokia.com (unverified) by esvir01nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T593b0e6fa8ac158f2116f@esvir01nok.ntc.nokia.com>;
 Fri, 22 Feb 2002 16:21:31 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh001.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Fri, 22 Feb 2002 16:21:31 +0200
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] RE: [SCTP-Discussion] Multihoming
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
Date: Fri, 22 Feb 2002 16:21:30 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF26308EECE@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] RE: [SCTP-Discussion] Multihoming
Thread-Index: AcG7qkw9/YqidENuRdG7PEmoZjo+AgAACeSw
To: <bidulock@openss7.org>, <randall@stewart.chicago.il.us>
Cc: <Michael.Tuexen@icn.siemens.de>, <tsvwg@ietf.org>, <discussion@sctp.de>
X-OriginalArrivalTime: 22 Feb 2002 14:21:31.0089 (UTC) FILETIME=[3929BC10:01C1BBAC]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id JAA24344
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi all!

> Randall,
> 
> This is a different situation altogether that you are
> talking about.  What you describe is of course already
> handled in most implementations I believe.

	I think we have been all the time speaking about if it is legal or not sending the INIT ACK from a different source address than the destination address of the INIT...

	Which other situation do you speak about? :-/

> The RFC says the INIT ACK MUST be sent to the source
> address of the INIT.

	Yes, that's true. This is to check that the source address in the INIT is the real one. Basically the whole cookie mechanism has been designed to be able to make this check before reserving any resource...

> We are questioning whether it is required to send the
> INIT ACK *from* the destination address of the INIT,
> or whether it is permitted to send them from any
> element in the list.

	Exactly, this is what we are speaking about.

> IMO the INIT process (INIT INIT-ACK exchange) is
> restricted to pair of transport endpoints (the one
> sending the INIT and the one replying with an INIT-ACK).

	I don't think so... When you receive the INIT ACK you already know the addresses of your peer... Why limiting then?

> There is absolutely no need to inject other IP addresses
> into the IP header (source or destination address).  Multi-
> homed addresses from which the INIT or INIT-ACK are sent
> should appear at most in the address list in the chunk.

	Well... really there isn't any need at all... but... if you can manage... Maybe the INIT ACK sender is clumsy and cannot really specify its source address (but it is using all of them in any case)...

	How was that sentence? "be careful about what you send, but accept as much as you can...", something like that...

> If you want to permit the behavior that the sender of
> the INIT-ACK can use any old source address in the list,

	Old source? They are not old? They are in use, and they will be used in the association... Why old? If you are not using an address any more definitely you shouldn't include it at all nor use it as your source address (in fact, if it is old, it is because you don't use it any more... :-))

> there would be a conflict with the RFC restriction that
> the INIT-ACK "MUST" be sent to the sender (source address)
> of the INIT.

	No, they are two different things. The destination is compulsorily that of the source of the INIT chunk. But the source can be whichever (whichever you use, of course).

> However, we discussed this before and it was the questioned
> whether opening up the INIT/INIT-ACK exchange to the entire
> address lists would not somehow weaken the secuity of the
> INIT process.

	How? Sending the COOKIE ECHO chunk you are proving that you received the INIT ACK chunk, i.e. that you are who you are stating you are... What is the difference if it comes from one or another address?

> If the INIT-ACK can come from any source address in the
> list, where is the COOKIE-ECHO then sent?

	To any of the valid addresses of the association...

	Thanks!

	BR Iván Arias Rodríguez

> --brian
> 
> On Fri, 22 Feb 2002, Randall Stewart wrote:
> 
> > Ivan:
> > 
> > Well you hit the nail on the head as ususal... without
> > doing this setup between multi-homed hosts does not
> > work very well... consider
> > 
> > Host-A                                   Host-Z
> > IP-1                                     IP-5
> > IP-2                                     IP-6
> > 
> > ---INIT(IP-1,IP-2)---\ /--INIT(IP-5,IP-6)---
> >                       X
> > toIP-1<-frm-IP-5-----/\------frm-IP-2------>toIP-6
> > 
> > Now unless inspection is done of the INIT/INIT-ACK the
> > two will NOT realize they are both setting up a association 
> > at the same time.
> > 
> > Another consideration is Address filtering. Often times
> > a host MUST use the IP address of the interface that you
> > are sending from.. or get the packet filtered so in the above
> > 
> > ----Frm-IP-1:INIT(IP-1,IP-2)------to(IP-5)----------->
> >                                                (route to 
> IP-1 points out
> > intf
> >                                                 of IP6.. 
> note this may
> > be
> >                                                 from either only one
> > allowed
> >                                                 default 
> route.. or some
> > other
> >                                                 reason).
> >  <----------to(IP-1)--INIT-ACK(IP-5,IP-6)--------From-IP-6
> > 
> > And we have the same secenario..
> > 
> > As far as looking for this text in the RFC .. sadly the full text
> > was chopped in one of the IESG passes (due to my lack of 
> communication
> > skills as to why it is there)... and now all that is left is the
> > trival little statement of:
> > 
> > "
> >       IMPLEMENTATION NOTE: In some cases (e.g., when the 
> implementation
> >       doesn't control the source IP address that is used for
> >       transmitting), an endpoint might need to include in 
> its INIT or
> >       INIT ACK all possible IP addresses from which packets 
> to the peer
> >       could be transmitted.
> > "
> > 
> > I think we will need to add a few lines to the implementors guide...
> > Also
> > there is NOTHING in the spec that makes this illegal as you 
> point out.
> > I remember we discussed this amongst the many emails and I addressed
> > just
> > these points.
> > 
> > Well more grist for the implementors guide :>
> > 
> > R
> > 
> > Ivan.Arias-Rodriguez@nokia.com wrote:
> > > 
> > >         Hi all!
> > > 
> > >         I also consider that situation to be legal. When 
> I receive an INIT
> > >         ACK I always look inside the parameters to see if 
> it belongs to
> > >         any open association (or in the COOKIE WAIT 
> state). Moreover, I
> > >         also do it with INIT chunks, to realize about 
> establishment
> > >         collisions or restart...
> > > 
> > >         If you don't allow this, I think it would be 
> quite hard to detect
> > >         INIT collisions in multihomed hosts... :-/
> > > 
> > >         I must admit that I forgot about this possibility 
> until after
> > >         having tested my implementation in a "brakeoff". 
> But, could you
> > >         Brian point to the section of the RFC that forbids this
> > >         possibility? In my understanding it is a logical 
> thing to do...
> > > 
> > >         Thanks!
> > > 
> > >         BR Iván Arias Rodríguez
> > > 
> > > > -----Original Message-----
> > > > From: ext Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > > > Sent: 22. February 2002 5:39
> > > > To: Randall Stewart
> > > > Cc: Tuexen Michael; 'tsvwg@ietf.org'; 'discussion@sctp.de'
> > > > Subject: Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
> > > >
> > > >
> > > > Randall,
> > > >
> > > > On Thu, 21 Feb 2002, Randall Stewart wrote:
> > > >
> > > > > Now the above case IS legal and there is specific wording in
> > > > > the RFC that covers this.
> > > >
> > > > Could you point to the section of the RFC that requires
> > > > that this scenario be permitted?
> > > >
> > > > --brian
> > > >
> > > >
> > > >
> > > > --
> > > > Brian F. G. Bidulock    ¦ The reasonable man adapts 
> himself to the ¦
> > > > bidulock@openss7.org    ¦ world; the unreasonable one 
> persists in  ¦
> > > > http://www.openss7.org/ ¦ trying  to adapt the  world  
> to himself. ¦
> > > >                         ¦ Therefore  all  progress  
> depends on the ¦
> > > >                         ¦ unreasonable man. -- George 
> Bernard Shaw ¦
> > > >
> > > > _______________________________________________
> > > > tsvwg mailing list
> > > > tsvwg@ietf.org
> > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > >
> > 
> > -- 
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 
> -- 
> Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
>                         ¦ Therefore  all  progress  depends on the ¦
>                         ¦ unreasonable man. -- George Bernard Shaw ¦
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Fri Feb 22 10:16:01 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA05246
	for <tsvwg-archive@odin.ietf.org>; Fri, 22 Feb 2002 10:16:01 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA28549
	for tsvwg-archive@odin.ietf.org; Fri, 22 Feb 2002 10:16:05 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA27293;
	Fri, 22 Feb 2002 10:00:11 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA27184
	for <tsvwg@optimus.ietf.org>; Fri, 22 Feb 2002 10:00:06 -0500 (EST)
Received: from stewartlin.chicago.il.us (host-64-47-24-166.masergy.com [64.47.24.166] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA04601
	for <tsvwg@ietf.org>; Fri, 22 Feb 2002 10:00:01 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewartlin.chicago.il.us (8.11.0/8.11.0) with ESMTP id g1MF1gr15528;
	Fri, 22 Feb 2002 09:01:42 -0600
Message-ID: <3C765CEC.B1D12905@stewart.chicago.il.us>
Date: Fri, 22 Feb 2002 08:59:56 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: bidulock@openss7.org, Michael.Tuexen@icn.siemens.de, tsvwg@ietf.org,
        discussion@sctp.de
Subject: Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
References: <58AEDE27D48F884285007ABF827FF26308EECE@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Ivan:


I could not agree more with yor response :>


one comments.. below..

Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Hi all!
> 
> > Randall,
> >
> > This is a different situation altogether that you are
> > talking about.  What you describe is of course already
> > handled in most implementations I believe.
> 
>         I think we have been all the time speaking about if it is legal or not sending the INIT ACK from a different source address than the destination address of the INIT...
> 
>         Which other situation do you speak about? :-/
> 
> > The RFC says the INIT ACK MUST be sent to the source
> > address of the INIT.
> 
>         Yes, that's true. This is to check that the source address in the INIT is the real one. Basically the whole cookie mechanism has been designed to be able to make this check before reserving any resource...
> 
> > We are questioning whether it is required to send the
> > INIT ACK *from* the destination address of the INIT,
> > or whether it is permitted to send them from any
> > element in the list.
> 
>         Exactly, this is what we are speaking about.
> 
> > IMO the INIT process (INIT INIT-ACK exchange) is
> > restricted to pair of transport endpoints (the one
> > sending the INIT and the one replying with an INIT-ACK).
> 
>         I don't think so... When you receive the INIT ACK you already know the addresses of your peer... Why limiting then?
> 
> > There is absolutely no need to inject other IP addresses
> > into the IP header (source or destination address).  Multi-
> > homed addresses from which the INIT or INIT-ACK are sent
> > should appear at most in the address list in the chunk.
> 
>         Well... really there isn't any need at all... but... if you can manage... Maybe the INIT ACK sender is clumsy and cannot really specify its source address (but it is using all of them in any case)...
> 
>         How was that sentence? "be careful about what you send, but accept as much as you can...", something like that...
> 
> > If you want to permit the behavior that the sender of
> > the INIT-ACK can use any old source address in the list,
> 
>         Old source? They are not old? They are in use, and they will be used in the association... Why old? If you are not using an address any more definitely you shouldn't include it at all nor use it as your source address (in fact, if it is old, it is because you don't use it any more... :-))
> 
> > there would be a conflict with the RFC restriction that
> > the INIT-ACK "MUST" be sent to the sender (source address)
> > of the INIT.
> 
>         No, they are two different things. The destination is compulsorily that of the source of the INIT chunk. But the source can be whichever (whichever you use, of course).
> 
> > However, we discussed this before and it was the questioned
> > whether opening up the INIT/INIT-ACK exchange to the entire
> > address lists would not somehow weaken the secuity of the
> > INIT process.
> 
>         How? Sending the COOKIE ECHO chunk you are proving that you received the INIT ACK chunk, i.e. that you are who you are stating you are... What is the difference if it comes from one or another address?
> 
> > If the INIT-ACK can come from any source address in the
> > list, where is the COOKIE-ECHO then sent?
> 
>         To any of the valid addresses of the association...
> 

Or you can respond to the address that sent it to you.. in any case
you have a LIST of valid addresses at the point of sending
the cookie-echo ...

Note this IS legal and really any GOOD implementation will
do it. We probably should add this to the implementation guide
as a SHOULD and let the poor implementations that don't do
this just suffer the consequences of there bad implementation.. i.e.
the won't always be able to build associations.

R




> 
> > --brian
> >
> > On Fri, 22 Feb 2002, Randall Stewart wrote:
> >
> > > Ivan:
> > >
> > > Well you hit the nail on the head as ususal... without
> > > doing this setup between multi-homed hosts does not
> > > work very well... consider
> > >
> > > Host-A                                   Host-Z
> > > IP-1                                     IP-5
> > > IP-2                                     IP-6
> > >
> > > ---INIT(IP-1,IP-2)---\ /--INIT(IP-5,IP-6)---
> > >                       X
> > > toIP-1<-frm-IP-5-----/\------frm-IP-2------>toIP-6
> > >
> > > Now unless inspection is done of the INIT/INIT-ACK the
> > > two will NOT realize they are both setting up a association
> > > at the same time.
> > >
> > > Another consideration is Address filtering. Often times
> > > a host MUST use the IP address of the interface that you
> > > are sending from.. or get the packet filtered so in the above
> > >
> > > ----Frm-IP-1:INIT(IP-1,IP-2)------to(IP-5)----------->
> > >                                                (route to
> > IP-1 points out
> > > intf
> > >                                                 of IP6..
> > note this may
> > > be
> > >                                                 from either only one
> > > allowed
> > >                                                 default
> > route.. or some
> > > other
> > >                                                 reason).
> > >  <----------to(IP-1)--INIT-ACK(IP-5,IP-6)--------From-IP-6
> > >
> > > And we have the same secenario..
> > >
> > > As far as looking for this text in the RFC .. sadly the full text
> > > was chopped in one of the IESG passes (due to my lack of
> > communication
> > > skills as to why it is there)... and now all that is left is the
> > > trival little statement of:
> > >
> > > "
> > >       IMPLEMENTATION NOTE: In some cases (e.g., when the
> > implementation
> > >       doesn't control the source IP address that is used for
> > >       transmitting), an endpoint might need to include in
> > its INIT or
> > >       INIT ACK all possible IP addresses from which packets
> > to the peer
> > >       could be transmitted.
> > > "
> > >
> > > I think we will need to add a few lines to the implementors guide...
> > > Also
> > > there is NOTHING in the spec that makes this illegal as you
> > point out.
> > > I remember we discussed this amongst the many emails and I addressed
> > > just
> > > these points.
> > >
> > > Well more grist for the implementors guide :>
> > >
> > > R
> > >
> > > Ivan.Arias-Rodriguez@nokia.com wrote:
> > > >
> > > >         Hi all!
> > > >
> > > >         I also consider that situation to be legal. When
> > I receive an INIT
> > > >         ACK I always look inside the parameters to see if
> > it belongs to
> > > >         any open association (or in the COOKIE WAIT
> > state). Moreover, I
> > > >         also do it with INIT chunks, to realize about
> > establishment
> > > >         collisions or restart...
> > > >
> > > >         If you don't allow this, I think it would be
> > quite hard to detect
> > > >         INIT collisions in multihomed hosts... :-/
> > > >
> > > >         I must admit that I forgot about this possibility
> > until after
> > > >         having tested my implementation in a "brakeoff".
> > But, could you
> > > >         Brian point to the section of the RFC that forbids this
> > > >         possibility? In my understanding it is a logical
> > thing to do...
> > > >
> > > >         Thanks!
> > > >
> > > >         BR Iván Arias Rodríguez
> > > >
> > > > > -----Original Message-----
> > > > > From: ext Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > > > > Sent: 22. February 2002 5:39
> > > > > To: Randall Stewart
> > > > > Cc: Tuexen Michael; 'tsvwg@ietf.org'; 'discussion@sctp.de'
> > > > > Subject: Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
> > > > >
> > > > >
> > > > > Randall,
> > > > >
> > > > > On Thu, 21 Feb 2002, Randall Stewart wrote:
> > > > >
> > > > > > Now the above case IS legal and there is specific wording in
> > > > > > the RFC that covers this.
> > > > >
> > > > > Could you point to the section of the RFC that requires
> > > > > that this scenario be permitted?
> > > > >
> > > > > --brian
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Brian F. G. Bidulock    ¦ The reasonable man adapts
> > himself to the ¦
> > > > > bidulock@openss7.org    ¦ world; the unreasonable one
> > persists in  ¦
> > > > > http://www.openss7.org/ ¦ trying  to adapt the  world
> > to himself. ¦
> > > > >                         ¦ Therefore  all  progress
> > depends on the ¦
> > > > >                         ¦ unreasonable man. -- George
> > Bernard Shaw ¦
> > > > >
> > > > > _______________________________________________
> > > > > tsvwg mailing list
> > > > > tsvwg@ietf.org
> > > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > > >
> > >
> > > --
> > > Randall R. Stewart
> > > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> >
> > --
> > Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
> > bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
> > http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
> >                         ¦ Therefore  all  progress  depends on the ¦
> >                         ¦ unreasonable man. -- George Bernard Shaw ¦
> >
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Feb 22 12:16:33 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA11812
	for <tsvwg-archive@odin.ietf.org>; Fri, 22 Feb 2002 12:16:33 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id MAA08645
	for tsvwg-archive@odin.ietf.org; Fri, 22 Feb 2002 12:16:36 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA07761;
	Fri, 22 Feb 2002 12:01:33 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA07727
	for <tsvwg@ns.ietf.org>; Fri, 22 Feb 2002 12:01:31 -0500 (EST)
Received: from mgw-x3.nokia.com (mgw-x3.nokia.com [131.228.20.26])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA10910
	for <tsvwg@ietf.org>; Fri, 22 Feb 2002 12:01:26 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir01nok.ntc.nokia.com (esvir01nokt.ntc.nokia.com [172.21.143.33])
	by mgw-x3.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1MH20i02048
	for <tsvwg@ietf.org>; Fri, 22 Feb 2002 19:02:00 +0200 (EET)
Received: from esebh003.NOE.Nokia.com (unverified) by esvir01nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T593ba0e166ac158f2116f@esvir01nok.ntc.nokia.com> for <tsvwg@ietf.org>;
 Fri, 22 Feb 2002 19:01:28 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh003.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Fri, 22 Feb 2002 19:01:28 +0200
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
Date: Fri, 22 Feb 2002 19:01:28 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF26308EED0@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] RE: [SCTP-Discussion] Multihoming
Thread-Index: AcG7sZqgJ4sGL3WeTGaFJoxM8zPeegADPRMA
To: <tsvwg@ietf.org>
X-OriginalArrivalTime: 22 Feb 2002 17:01:28.0577 (UTC) FILETIME=[91B62F10:01C1BBC2]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id MAA07728
Subject: [Tsvwg] Questions about the AddIP extension
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi all!

	Some questions...

	What happens if the peer tells us to set as primary one address that is not part of the association? There isn't any error cause to state this, so, either you report a success (which is not the case, but...), or you send an error cause (which one?)...

	The same applies to the deletion of one address... But in this case the situation is somehow worse... :-/

	Moreover... I'm thinking now about something... uhm... Not sure if it would be a problem at all. :-0

	Imagine that we (the server) send the INIT ACK including the same Initiate Tag as the INIT sender (client). Then, using AddIP we (the server) do so that we erase our address and include one of the addresses of the client... If we haven't acknowledged anything yet, the client would be sending DATA chunks to itself and acknowledging them... It would receive its own DATA and SACK chunks, believing that is the server the one that is sending them... Until the client ends sending data to itself it won't stop the association...

	As the server is the one doing this "attack" I don't see the point of doing it, but... Basically, what I was thinking about is what should be done if the peer tries to add one of "our addresses" as if it was one of its addresses... Again, if we report an error, which one?

	Wouldn't it be nice to have an error cause saying something like "invalid IP address to perform the operation"? It could be used for all these "strange" situations... and maybe some more I can not think about now...

	By the way the "Error Cause Indication" parameter can be used to "Return Info on Success", how? Is this a mistake or not?

	Thanks!

	BR Iván Arias Rodríguez

> -----Original Message-----
> From: ext Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: 22. February 2002 17:00
> To: Arias-Rodriguez Ivan (NRC/Helsinki)
> Cc: bidulock@openss7.org; Michael.Tuexen@icn.siemens.de; 
> tsvwg@ietf.org;
> discussion@sctp.de
> Subject: Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
> 
> 
> Ivan:
> 
> 
> I could not agree more with yor response :>
> 
> 
> one comments.. below..
> 
> Ivan.Arias-Rodriguez@nokia.com wrote:
> > 
> >         Hi all!
> > 
> > > Randall,
> > >
> > > This is a different situation altogether that you are
> > > talking about.  What you describe is of course already
> > > handled in most implementations I believe.
> > 
> >         I think we have been all the time speaking about if 
> it is legal or not sending the INIT ACK from a different 
> source address than the destination address of the INIT...
> > 
> >         Which other situation do you speak about? :-/
> > 
> > > The RFC says the INIT ACK MUST be sent to the source
> > > address of the INIT.
> > 
> >         Yes, that's true. This is to check that the source 
> address in the INIT is the real one. Basically the whole 
> cookie mechanism has been designed to be able to make this 
> check before reserving any resource...
> > 
> > > We are questioning whether it is required to send the
> > > INIT ACK *from* the destination address of the INIT,
> > > or whether it is permitted to send them from any
> > > element in the list.
> > 
> >         Exactly, this is what we are speaking about.
> > 
> > > IMO the INIT process (INIT INIT-ACK exchange) is
> > > restricted to pair of transport endpoints (the one
> > > sending the INIT and the one replying with an INIT-ACK).
> > 
> >         I don't think so... When you receive the INIT ACK 
> you already know the addresses of your peer... Why limiting then?
> > 
> > > There is absolutely no need to inject other IP addresses
> > > into the IP header (source or destination address).  Multi-
> > > homed addresses from which the INIT or INIT-ACK are sent
> > > should appear at most in the address list in the chunk.
> > 
> >         Well... really there isn't any need at all... 
> but... if you can manage... Maybe the INIT ACK sender is 
> clumsy and cannot really specify its source address (but it 
> is using all of them in any case)...
> > 
> >         How was that sentence? "be careful about what you 
> send, but accept as much as you can...", something like that...
> > 
> > > If you want to permit the behavior that the sender of
> > > the INIT-ACK can use any old source address in the list,
> > 
> >         Old source? They are not old? They are in use, and 
> they will be used in the association... Why old? If you are 
> not using an address any more definitely you shouldn't 
> include it at all nor use it as your source address (in fact, 
> if it is old, it is because you don't use it any more... :-))
> > 
> > > there would be a conflict with the RFC restriction that
> > > the INIT-ACK "MUST" be sent to the sender (source address)
> > > of the INIT.
> > 
> >         No, they are two different things. The destination 
> is compulsorily that of the source of the INIT chunk. But the 
> source can be whichever (whichever you use, of course).
> > 
> > > However, we discussed this before and it was the questioned
> > > whether opening up the INIT/INIT-ACK exchange to the entire
> > > address lists would not somehow weaken the secuity of the
> > > INIT process.
> > 
> >         How? Sending the COOKIE ECHO chunk you are proving 
> that you received the INIT ACK chunk, i.e. that you are who 
> you are stating you are... What is the difference if it comes 
> from one or another address?
> > 
> > > If the INIT-ACK can come from any source address in the
> > > list, where is the COOKIE-ECHO then sent?
> > 
> >         To any of the valid addresses of the association...
> > 
> 
> Or you can respond to the address that sent it to you.. in any case
> you have a LIST of valid addresses at the point of sending
> the cookie-echo ...
> 
> Note this IS legal and really any GOOD implementation will
> do it. We probably should add this to the implementation guide
> as a SHOULD and let the poor implementations that don't do
> this just suffer the consequences of there bad implementation.. i.e.
> the won't always be able to build associations.
> 
> R
> 
> 
> 
> 
> > 
> > > --brian
> > >
> > > On Fri, 22 Feb 2002, Randall Stewart wrote:
> > >
> > > > Ivan:
> > > >
> > > > Well you hit the nail on the head as ususal... without
> > > > doing this setup between multi-homed hosts does not
> > > > work very well... consider
> > > >
> > > > Host-A                                   Host-Z
> > > > IP-1                                     IP-5
> > > > IP-2                                     IP-6
> > > >
> > > > ---INIT(IP-1,IP-2)---\ /--INIT(IP-5,IP-6)---
> > > >                       X
> > > > toIP-1<-frm-IP-5-----/\------frm-IP-2------>toIP-6
> > > >
> > > > Now unless inspection is done of the INIT/INIT-ACK the
> > > > two will NOT realize they are both setting up a association
> > > > at the same time.
> > > >
> > > > Another consideration is Address filtering. Often times
> > > > a host MUST use the IP address of the interface that you
> > > > are sending from.. or get the packet filtered so in the above
> > > >
> > > > ----Frm-IP-1:INIT(IP-1,IP-2)------to(IP-5)----------->
> > > >                                                (route to
> > > IP-1 points out
> > > > intf
> > > >                                                 of IP6..
> > > note this may
> > > > be
> > > >                                                 from 
> either only one
> > > > allowed
> > > >                                                 default
> > > route.. or some
> > > > other
> > > >                                                 reason).
> > > >  <----------to(IP-1)--INIT-ACK(IP-5,IP-6)--------From-IP-6
> > > >
> > > > And we have the same secenario..
> > > >
> > > > As far as looking for this text in the RFC .. sadly the 
> full text
> > > > was chopped in one of the IESG passes (due to my lack of
> > > communication
> > > > skills as to why it is there)... and now all that is left is the
> > > > trival little statement of:
> > > >
> > > > "
> > > >       IMPLEMENTATION NOTE: In some cases (e.g., when the
> > > implementation
> > > >       doesn't control the source IP address that is used for
> > > >       transmitting), an endpoint might need to include in
> > > its INIT or
> > > >       INIT ACK all possible IP addresses from which packets
> > > to the peer
> > > >       could be transmitted.
> > > > "
> > > >
> > > > I think we will need to add a few lines to the 
> implementors guide...
> > > > Also
> > > > there is NOTHING in the spec that makes this illegal as you
> > > point out.
> > > > I remember we discussed this amongst the many emails 
> and I addressed
> > > > just
> > > > these points.
> > > >
> > > > Well more grist for the implementors guide :>
> > > >
> > > > R
> > > >
> > > > Ivan.Arias-Rodriguez@nokia.com wrote:
> > > > >
> > > > >         Hi all!
> > > > >
> > > > >         I also consider that situation to be legal. When
> > > I receive an INIT
> > > > >         ACK I always look inside the parameters to see if
> > > it belongs to
> > > > >         any open association (or in the COOKIE WAIT
> > > state). Moreover, I
> > > > >         also do it with INIT chunks, to realize about
> > > establishment
> > > > >         collisions or restart...
> > > > >
> > > > >         If you don't allow this, I think it would be
> > > quite hard to detect
> > > > >         INIT collisions in multihomed hosts... :-/
> > > > >
> > > > >         I must admit that I forgot about this possibility
> > > until after
> > > > >         having tested my implementation in a "brakeoff".
> > > But, could you
> > > > >         Brian point to the section of the RFC that 
> forbids this
> > > > >         possibility? In my understanding it is a logical
> > > thing to do...
> > > > >
> > > > >         Thanks!
> > > > >
> > > > >         BR Iván Arias Rodríguez
> > > > >
> > > > > > -----Original Message-----
> > > > > > From: ext Brian F. G. Bidulock [mailto:bidulock@openss7.org]
> > > > > > Sent: 22. February 2002 5:39
> > > > > > To: Randall Stewart
> > > > > > Cc: Tuexen Michael; 'tsvwg@ietf.org'; 'discussion@sctp.de'
> > > > > > Subject: Re: [Tsvwg] RE: [SCTP-Discussion] Multihoming
> > > > > >
> > > > > >
> > > > > > Randall,
> > > > > >
> > > > > > On Thu, 21 Feb 2002, Randall Stewart wrote:
> > > > > >
> > > > > > > Now the above case IS legal and there is specific 
> wording in
> > > > > > > the RFC that covers this.
> > > > > >
> > > > > > Could you point to the section of the RFC that requires
> > > > > > that this scenario be permitted?
> > > > > >
> > > > > > --brian
> > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Brian F. G. Bidulock    ¦ The reasonable man adapts
> > > himself to the ¦
> > > > > > bidulock@openss7.org    ¦ world; the unreasonable one
> > > persists in  ¦
> > > > > > http://www.openss7.org/ ¦ trying  to adapt the  world
> > > to himself. ¦
> > > > > >                         ¦ Therefore  all  progress
> > > depends on the ¦
> > > > > >                         ¦ unreasonable man. -- George
> > > Bernard Shaw ¦
> > > > > >
> > > > > > _______________________________________________
> > > > > > tsvwg mailing list
> > > > > > tsvwg@ietf.org
> > > > > > https://www1.ietf.org/mailman/listinfo/tsvwg
> > > > > >
> > > >
> > > > --
> > > > Randall R. Stewart
> > > > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> > >
> > > --
> > > Brian F. G. Bidulock    ¦ The reasonable man adapts 
> himself to the ¦
> > > bidulock@openss7.org    ¦ world; the unreasonable one 
> persists in  ¦
> > > http://www.openss7.org/ ¦ trying  to adapt the  world  to 
> himself. ¦
> > >                         ¦ Therefore  all  progress  
> depends on the ¦
> > >                         ¦ unreasonable man. -- George 
> Bernard Shaw ¦
> > >
> > 
> > _______________________________________________
> > tsvwg mailing list
> > tsvwg@ietf.org
> > https://www1.ietf.org/mailman/listinfo/tsvwg
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Feb 22 12:18:23 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA11953
	for <tsvwg-archive@odin.ietf.org>; Fri, 22 Feb 2002 12:18:23 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id MAA08746
	for tsvwg-archive@odin.ietf.org; Fri, 22 Feb 2002 12:18:26 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA08111;
	Fri, 22 Feb 2002 12:04:06 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA08080
	for <tsvwg@ns.ietf.org>; Fri, 22 Feb 2002 12:04:03 -0500 (EST)
Received: from mgw-x1.nokia.com (mgw-x1.nokia.com [131.228.20.21])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA11076
	for <tsvwg@ietf.org>; Fri, 22 Feb 2002 12:03:59 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir05nok.ntc.nokia.com (esvir05nokt.ntc.nokia.com [172.21.143.37])
	by mgw-x1.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1MH3V309480
	for <tsvwg@ietf.org>; Fri, 22 Feb 2002 19:03:31 +0200 (EET)
Received: from esebh002.NOE.Nokia.com (unverified) by esvir05nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T593ba335b2ac158f25144a@esvir05nok.ntc.nokia.com> for <tsvwg@ietf.org>;
 Fri, 22 Feb 2002 19:04:01 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh002.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Fri, 22 Feb 2002 19:04:01 +0200
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] Questions about the AddIP extension
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
Date: Fri, 22 Feb 2002 19:04:01 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF263257B3D@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] RE: [SCTP-Discussion] Multihoming
Thread-Index: AcG7sZqgJ4sGL3WeTGaFJoxM8zPeegADPRMAAAEIdkA=
To: <tsvwg@ietf.org>
X-OriginalArrivalTime: 22 Feb 2002 17:04:01.0296 (UTC) FILETIME=[ECBD4100:01C1BBC2]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id MAA08081
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Ups!

	I included Randall's mail at the end just because I used that mail to take the tsvwg address...

	Sorry, my mistake...

	BR Iván Arias Rodríguez

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Feb 22 13:34:47 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA15500
	for <tsvwg-archive@odin.ietf.org>; Fri, 22 Feb 2002 13:34:47 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA13817
	for tsvwg-archive@odin.ietf.org; Fri, 22 Feb 2002 13:34:49 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA11769;
	Fri, 22 Feb 2002 13:16:00 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA11740
	for <tsvwg@ns.ietf.org>; Fri, 22 Feb 2002 13:15:58 -0500 (EST)
Received: from rwcrmhc51.attbi.com (rwcrmhc51.attbi.com [204.127.198.38])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA14675
	for <tsvwg@ietf.org>; Fri, 22 Feb 2002 13:15:56 -0500 (EST)
Received: from ieee.org ([12.248.77.94]) by rwcrmhc51.attbi.com
          (InterMail vM.4.01.03.27 201-229-121-127-20010626) with ESMTP
          id <20020222181526.RWOA2626.rwcrmhc51.attbi.com@ieee.org>;
          Fri, 22 Feb 2002 18:15:26 +0000
Message-ID: <3C768AAD.235F4F60@ieee.org>
Date: Fri, 22 Feb 2002 12:15:09 -0600
From: Peter Lei <peter.lei@ieee.org>
X-Mailer: Mozilla 4.79 [en] (Windows NT 5.0; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] Questions about the AddIP extension
References: <58AEDE27D48F884285007ABF827FF26308EED0@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Ivan,

comments inline:

Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Hi all!
> 
>         Some questions...
> 
>         What happens if the peer tells us to set as primary one address that is not part of the association? There isn't any error cause to state this, so, either you report a success (which is not the case, but...), or you send an error cause (which one?)...

The SET_PRIMARY_IP_ADDRESS asconf is just a request for the receiver to mark
the address as primary... a receiver MAY choose to honor the request or not.
Our implementation currently sends back an "Unresolvable Address" (code 5)
error cause indication if it doesn't recognize the address given; it was the
closest to "Invalid Address"...

>         The same applies to the deletion of one address... But in this case the situation is somehow worse... :-/

Hmm... here our implementation currently simply sends a "success" ack, since the
peer is deleting the (deleted) address anyways; otherwise, if we send an error
back, an implementation might consider it still valid (since the delete failed)
and continue to use it and then eventually mark that address unreachable.  Yes,
in any case, it does indicate that something had gone wrong already...

>         Moreover... I'm thinking now about something... uhm... Not sure if it would be a problem at all. :-0
>         Imagine that we (the server) send the INIT ACK including the same Initiate Tag as the INIT sender (client). Then, using AddIP we (the server) do so that we erase our address and include one of the addresses of the client... If we haven't acknowledged anything yet, the client would be sending DATA chunks to itself and acknowledging them... It would receive its own DATA and SACK chunks, believing that is the server the one that is sending them... Until the client ends sending data to itself it won't stop the association...
>
>         As the server is the one doing this "attack" I don't see the point of doing it, but... Basically, what I was thinking about is what should be done if the peer tries to add one of "our addresses" as if it was one of its addresses... Again, if we report an error, which one?
>
>         Wouldn't it be nice to have an error cause saying something like "invalid IP address to perform the operation"? It could be used for all these "strange" situations... and maybe some more I can not think about now...
>
>         By the way the "Error Cause Indication" parameter can be used to "Return Info on Success", how? Is this a mistake or not?

I hope that it is a mistake... :)
 
>         Thanks!

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Fri Feb 22 15:02:04 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA19856
	for <tsvwg-archive@odin.ietf.org>; Fri, 22 Feb 2002 15:02:04 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA22272
	for tsvwg-archive@odin.ietf.org; Fri, 22 Feb 2002 15:02:07 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA20486;
	Fri, 22 Feb 2002 14:45:32 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA20443
	for <tsvwg@ns.ietf.org>; Fri, 22 Feb 2002 14:45:29 -0500 (EST)
Received: from stewartlin.chicago.il.us (host-64-47-24-166.masergy.com [64.47.24.166] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA18894
	for <tsvwg@ietf.org>; Fri, 22 Feb 2002 14:45:25 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewartlin.chicago.il.us (8.11.0/8.11.0) with ESMTP id g1MJl8r15948;
	Fri, 22 Feb 2002 13:47:08 -0600
Message-ID: <3C769FD3.27470586@stewart.chicago.il.us>
Date: Fri, 22 Feb 2002 13:45:23 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] Questions about the AddIP extension
References: <58AEDE27D48F884285007ABF827FF26308EED0@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Hi all!
> 
>         Some questions...
> 
>         What happens if the peer tells us to set as primary one address that is not part of the association? There isn't any error cause to state this, so, either you report a success (which is not the case, but...), or you send an error cause (which one?)...
> 

For this one we never put in an error since it is completely optional
by the receiver of the set-primary to honor the request... its a
suggestion.
I do not think it is worth responding either way...


>         The same applies to the deletion of one address... But in this case the situation is somehow worse... :-/
> 
>         Moreover... I'm thinking now about something... uhm... Not sure if it would be a problem at all. :-0

The peer deleteing a non existant address is probably an indication
of an internal error of your peer.. not that it harms anything... if
I tell you sure you can delete an address that is NOT part of
the association.. does it harm anything (anymore than whatever
messed up state you are already in)?

> 
>         Imagine that we (the server) send the INIT ACK including the same Initiate Tag as the INIT sender (client). Then, using AddIP we (the server) do so that we erase our address and include one of the addresses of the client... If we haven't acknowledged anything yet, the client would be sending DATA chunks to itself and acknowledging them... It would receive its own DATA and SACK chunks, believing that is the server the one that is sending them... Until the client ends sending data to itself it won't stop the association...
> 
>         As the server is the one doing this "attack" I don't see the point of doing it, but... Basically, what I was thinking about is what should be done if the peer tries to add one of "our addresses" as if it was one of its addresses... Again, if we report an error, which one?
> 
>         Wouldn't it be nice to have an error cause saying something like "invalid IP address to perform the operation"? It could be used for all these "strange" situations... and maybe some more I can not think about now...
> 
>         By the way the "Error Cause Indication" parameter can be used to "Return Info on Success", how? Is this a mistake or not?


Seems to me you more illustrate an attack here. No peer that server or
client
would do this unless they either were:

a) -malicious
or
b) -messed up internally..

In either case at that point we should probably just abort the
association.
We may want to go examine our BSD code and think about doing this :>

R


-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Feb 23 12:20:39 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA21840
	for <tsvwg-archive@odin.ietf.org>; Sat, 23 Feb 2002 12:20:39 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id MAA22060
	for tsvwg-archive@odin.ietf.org; Sat, 23 Feb 2002 12:20:39 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA21782;
	Sat, 23 Feb 2002 12:07:31 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA21750
	for <tsvwg@optimus.ietf.org>; Sat, 23 Feb 2002 12:07:29 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA21722
	for <tsvwg@ietf.org>; Sat, 23 Feb 2002 12:07:24 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir04nok.ntc.nokia.com (esvir04nokt.ntc.nokia.com [172.21.143.36])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1NH7YZ23839
	for <tsvwg@ietf.org>; Sat, 23 Feb 2002 19:07:35 +0200 (EET)
Received: from esebh003.NOE.Nokia.com (unverified) by esvir04nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5940ccb257ac158f242fb@esvir04nok.ntc.nokia.com> for <tsvwg@ietf.org>;
 Sat, 23 Feb 2002 19:07:26 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh003.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Sat, 23 Feb 2002 19:07:26 +0200
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
Date: Sat, 23 Feb 2002 19:07:25 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF26308EED1@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] Questions about the AddIP extension
Thread-Index: AcG72XqBeUORpvtNQjST1XEKfN3TbgAsEWHg
To: <tsvwg@ietf.org>
X-OriginalArrivalTime: 23 Feb 2002 17:07:26.0160 (UTC) FILETIME=[9142C500:01C1BC8C]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id MAA21751
Subject: [Tsvwg] More questions about the AddIP extension
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi again!

	Suddenly I realized about one thing. The AddIP extension is designed in a way that if the ASCONF ACK chunk is lost and the ASCONF chunk is resent, the ASCONF receiver should answer in the same way as if it would have never received the first ASCONF chunk, not modifying anything as it should have done it the first time. This is of course the right way of doing.

	But, there is an implementation note in page 15 that suggests keeping the ASCONF ACK chunk sent as an "optimization"... I tend to think that this is not an optimization but a compulsory thing to do... The reason is that when parsing the ASCONF chunk for the second time the answer might be different because:

	- If we agree that we treat as a success adding an address already in use or deleting an address that was not in use, then the answer to the second ASCONF (including additions and deletions of addresses) will be effectively the same. However, in this case I think that this fact (considering these situations as successes) should be stated in the draft.

	- I can think about another weird but still valid situation. If in the ASCONF the peer tells us to set our primary address to A, and in the same chunk it tells us in a subsequent parameter to delete that address. Upon the receipt of the second ASCONF, the "Set Primary Address" parameter would fail, since that address is not in use... The ASCONF sender is doing strange things, but for example Peter Lei's implementation would produce a different answer for the second ASCONF... :-/

	So, I think that either keeping the ASCONF ACK chunk is compulsory, or the behavior in these "corner case" situations (actually the first case is not a corner case at all) should be clearly stated in the draft...

	Comments?

	Thanks!

	BR Iván Arias Rodríguez

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sat Feb 23 13:56:52 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA22738
	for <tsvwg-archive@odin.ietf.org>; Sat, 23 Feb 2002 13:56:52 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA24969
	for tsvwg-archive@odin.ietf.org; Sat, 23 Feb 2002 13:56:54 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA24712;
	Sat, 23 Feb 2002 13:42:57 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA24683
	for <tsvwg@optimus.ietf.org>; Sat, 23 Feb 2002 13:42:55 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA22689
	for <tsvwg@ietf.org>; Sat, 23 Feb 2002 13:42:51 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir03nok.nokia.com (esvir03nokt.ntc.nokia.com [172.21.143.35])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1NIh0Z16803
	for <tsvwg@ietf.org>; Sat, 23 Feb 2002 20:43:01 +0200 (EET)
Received: from esebh003.NOE.Nokia.com (unverified) by esvir03nok.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5941240c90ac158f23694@esvir03nok.nokia.com> for <tsvwg@ietf.org>;
 Sat, 23 Feb 2002 20:42:50 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh003.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Sat, 23 Feb 2002 20:42:52 +0200
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
Date: Sat, 23 Feb 2002 20:42:51 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF263257B41@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] Questions about the AddIP extension
Thread-Index: AcG72XqBeUORpvtNQjST1XEKfN3TbgAtZWfQ
To: <tsvwg@ietf.org>
X-OriginalArrivalTime: 23 Feb 2002 18:42:52.0377 (UTC) FILETIME=[E65A2490:01C1BC99]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id NAA24684
Subject: [Tsvwg] Some other questions about the AddIP extension
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Debugging my implementation I'm thinking about strange situations...

	What will happen if the peer tries to add an address that is being used by another association within this same endpoint (of course, both associations should be using the same destination port as well)? Silently discard the chunk (and maybe the whole datagram)?

	Thinking about my previous mail... There is a really weird case when the second ASCONF ACK might be different. When the peer (an attacker?) tells us to add an address that is being used by another association, but that association deletes that address before we receive the second ASCONF chunk... And there might be more situations (ok, REALLY strange ones... :-)). So this makes me think harder that either the last ASCONF ACK chunk should be compulsorily kept in memory (in fact, I think that this is what all of us are going to do), or it is stated in the draft that simply re-parsing the ASCONF chunk might not be the right thing to do... :-0

	What should we do if the peer deletes all the active addresses (i.e., after the deletion we only have inactive addresses)? Should we close the association? :-/ The ASCONF chunk must have come from one of the inactive addresses, so the card and the return path are working... Maybe we should try with the inactive addresses until we reach the Max.Association.Retrans... I guess this might be an implementation specific issue...

	Any comment? Thanks!

	BR Iván Arias Rodríguez

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Sun Feb 24 12:54:40 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA14116
	for <tsvwg-archive@odin.ietf.org>; Sun, 24 Feb 2002 12:54:35 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id MAA20620
	for tsvwg-archive@odin.ietf.org; Sun, 24 Feb 2002 12:54:37 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA19990;
	Sun, 24 Feb 2002 12:38:54 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id MAA19961
	for <tsvwg@optimus.ietf.org>; Sun, 24 Feb 2002 12:38:52 -0500 (EST)
Received: from relay6.kornet.net ([211.48.62.166])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA13939
	for <tsvwg@ietf.org>; Sun, 24 Feb 2002 12:38:47 -0500 (EST)
Received: from vivi-41a7ime526 (211.227.92.91) by relay6.kornet.net; 25 Feb 2002 02:38:05 +0900
Message-ID: <3c7924fe3c97ba3b@relay6.kornet.net> (added by relay6.kornet.net)
From: =?ks_c_5601-1987?B?VklWSSBLT1JFQQ==?= <vivi@vivi.co.kr>
To: tsvwg@ietf.org
Date: Mon, 25 Feb 2002 02:37:56 +0900
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_NextPart_000_0027_01C0F37A.93A56C00"
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Subject: [Tsvwg] =?ks_c_5601-1987?B?W7GksO1dIHRzdndntNQgsKGxuLD4wOW/obytIMavurDH0sDOsKG3ziCw7bHeu/3IsLChsbi4piC6uLO7teW4s7TPtNku?=
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

This is a multi-part message in MIME format.

------=_NextPart_000_0027_01C0F37A.93A56C00
Content-Type: text/plain;
	charset="ks_c_5601-1987"
Content-Transfer-Encoding: base64
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------=_NextPart_000_0027_01C0F37A.93A56C00
Content-Type: text/html;
	charset="ks_c_5601-1987"
Content-Transfer-Encoding: base64
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------=_NextPart_000_0027_01C0F37A.93A56C00--


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 25 06:40:37 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA05923
	for <tsvwg-archive@odin.ietf.org>; Mon, 25 Feb 2002 06:40:37 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA13684
	for tsvwg-archive@odin.ietf.org; Mon, 25 Feb 2002 06:40:40 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA12193;
	Mon, 25 Feb 2002 06:24:42 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA12164
	for <tsvwg@optimus.ietf.org>; Mon, 25 Feb 2002 06:24:40 -0500 (EST)
Received: from CNRI.Reston.VA.US (localhost [127.0.0.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA05280;
	Mon, 25 Feb 2002 06:24:37 -0500 (EST)
Message-Id: <200202251124.GAA05280@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
Cc: tsvwg@ietf.org
From: Internet-Drafts@ietf.org
Reply-to: Internet-Drafts@ietf.org
Date: Mon, 25 Feb 2002 06:24:37 -0500
Subject: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-initwin-01.txt
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the Transport Area Working Group Working Group of the IETF.

	Title		: Increasing TCP's Initial Window
	Author(s)	: M. Allman, S. Floyd, C. Partridge
	Filename	: draft-ietf-tsvwg-initwin-01.txt
	Pages		: 12
	Date		: 22-Feb-02
	
This document specifies an increase in the permitted initial window
for TCP from one segment to roughly 4K bytes.  This document also
discusses the advantages and disadvantages of the change, outlining
experimental results that indicate the costs and benefits.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-initwin-01.txt

To remove yourself from the IETF Announcement list, send a message to 
ietf-announce-request with the word unsubscribe in the body of the message.

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-tsvwg-initwin-01.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-tsvwg-initwin-01.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<20020222110946.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-tsvwg-initwin-01.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-tsvwg-initwin-01.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<20020222110946.I-D@ietf.org>

--OtherAccess--

--NextPart--



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 25 06:44:05 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA05992
	for <tsvwg-archive@odin.ietf.org>; Mon, 25 Feb 2002 06:44:05 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA13856
	for tsvwg-archive@odin.ietf.org; Mon, 25 Feb 2002 06:44:07 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA12227;
	Mon, 25 Feb 2002 06:24:47 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA12201
	for <tsvwg@optimus.ietf.org>; Mon, 25 Feb 2002 06:24:45 -0500 (EST)
Received: from CNRI.Reston.VA.US (localhost [127.0.0.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA05298;
	Mon, 25 Feb 2002 06:24:42 -0500 (EST)
Message-Id: <200202251124.GAA05298@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
Cc: tsvwg@ietf.org
From: Internet-Drafts@ietf.org
Reply-to: Internet-Drafts@ietf.org
Date: Mon, 25 Feb 2002 06:24:41 -0500
Subject: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-tcp-eifel-alg-03.txt
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the Transport Area Working Group Working Group of the IETF.

	Title		: The Eifel Detection Algorithm for TCP
	Author(s)	: L. Reiner, M. Meyer
	Filename	: draft-ietf-tsvwg-tcp-eifel-alg-03.txt
	Pages		: 10
	Date		: 22-Feb-02
	
The Eifel detection algorithm allows the TCP sender to detect a
posteriori whether it has entered loss recovery unnecessarily. It
already determines this in the beginning of loss recovery when the
first new ACK arrives after the timeout-based or fast retransmit has
been sent. The algorithm requires that the TCP Timestamps option
defined in RFC1323 is enabled for a connection. The idea is to use
the timestamps echoed in returning ACKs to eliminate the
retransmission ambiguity in TCP. The Eifel detection algorithm
provides a basis for future TCP enhancements. This includes response
algorithms to back out of loss recovery by restoring a TCP sender's
congestion control state.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-tcp-eifel-alg-03.txt

To remove yourself from the IETF Announcement list, send a message to 
ietf-announce-request with the word unsubscribe in the body of the message.

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-ietf-tsvwg-tcp-eifel-alg-03.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-ietf-tsvwg-tcp-eifel-alg-03.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<20020222110959.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-tsvwg-tcp-eifel-alg-03.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-ietf-tsvwg-tcp-eifel-alg-03.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<20020222110959.I-D@ietf.org>

--OtherAccess--

--NextPart--



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 25 06:55:35 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA06297
	for <tsvwg-archive@odin.ietf.org>; Mon, 25 Feb 2002 06:55:35 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA14216
	for tsvwg-archive@odin.ietf.org; Mon, 25 Feb 2002 06:55:37 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA13650;
	Mon, 25 Feb 2002 06:39:56 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA13617
	for <tsvwg@optimus.ietf.org>; Mon, 25 Feb 2002 06:39:54 -0500 (EST)
Received: from stewartlin.chicago.il.us (host-64-47-24-166.masergy.com [64.47.24.166] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA05890
	for <tsvwg@ietf.org>; Mon, 25 Feb 2002 06:39:51 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewartlin.chicago.il.us (8.11.0/8.11.0) with ESMTP id g1PBa6r29550;
	Mon, 25 Feb 2002 05:36:06 -0600
Message-ID: <3C7A2151.8750FD60@stewart.chicago.il.us>
Date: Mon, 25 Feb 2002 05:34:41 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] More questions about the AddIP extension
References: <58AEDE27D48F884285007ABF827FF26308EED1@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Hi again!
> 
>         Suddenly I realized about one thing. The AddIP extension is designed in a way that if the ASCONF ACK chunk is lost and the ASCONF chunk is resent, the ASCONF receiver should answer in the same way as if it would have never received the first ASCONF chunk, not modifying anything as it should have done it the first time. This is of course the right way of doing.
> 

Yep...

>         But, there is an implementation note in page 15 that suggests keeping the ASCONF ACK chunk sent as an "optimization"... I tend to think that this is not an optimization but a compulsory thing to do... The reason is that when parsing the ASCONF chunk for the second time the answer might be different because:
> 
>         - If we agree that we treat as a success adding an address already in use or deleting an address that was not in use, then the answer to the second ASCONF (including additions and deletions of addresses) will be effectively the same. However, in this case I think that this fact (considering these situations as successes) should be stated in the draft.
> 
>         - I can think about another weird but still valid situation. If in the ASCONF the peer tells us to set our primary address to A, and in the same chunk it tells us in a subsequent parameter to delete that address. Upon the receipt of the second ASCONF, the "Set Primary Address" parameter would fail, since that address is not in use... The ASCONF sender is doing strange things, but for example Peter Lei's implementation would produce a different answer for the second ASCONF... :-/

No... you are required to give the same answer.. 

> 
>         So, I think that either keeping the ASCONF ACK chunk is compulsory, or the behavior in these "corner case" situations (actually the first case is not a corner case at all) should be clearly stated in the draft...
> 


I disagree here Ivan... there may be some better way than keeping the
asconf-ack around. Peter and I do this in our implementation but
let us not limit implementors. It is required they they always
answer the same but let's not dictate how you implement that.

I could also image an implementation that keeps a state history
of some kind and thus can reconstruct a state (ip addresses, primaries
and such) at any given time. So when an ASCONF-ACK arrives it
can send back the exact same answer. Remember you can only
have ONE ASCONF out at a time.. so in a state keeping implementation
you would only have to keep one set of state (if you were worried
about space)..


R

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 25 06:58:27 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA06327
	for <tsvwg-archive@odin.ietf.org>; Mon, 25 Feb 2002 06:58:27 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA14342
	for tsvwg-archive@odin.ietf.org; Mon, 25 Feb 2002 06:58:29 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA13868;
	Mon, 25 Feb 2002 06:44:08 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA13828
	for <tsvwg@optimus.ietf.org>; Mon, 25 Feb 2002 06:44:05 -0500 (EST)
Received: from ganymede.or.intel.com (jffdns01.or.intel.com [134.134.248.3])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA05986
	for <tsvwg@ietf.org>; Mon, 25 Feb 2002 06:44:02 -0500 (EST)
Received: from orsmsxvs040.jf.intel.com (orsmsxvs040.jf.intel.com [192.168.65.206])
	by ganymede.or.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.51 2002/02/19 21:12:32 root Exp $) with SMTP id LAA26415
	for <tsvwg@ietf.org>; Mon, 25 Feb 2002 11:43:34 GMT
Received: from orsmsx26.jf.intel.com ([192.168.65.26])
 by orsmsxvs040.jf.intel.com (NAVGW 2.5.1.16) with SMTP id M2002022503485816538
 for <tsvwg@ietf.org>; Mon, 25 Feb 2002 03:48:58 -0800
Received: by orsmsx26.jf.intel.com with Internet Mail Service (5.5.2653.19)
	id <1YPLMNZQ>; Mon, 25 Feb 2002 03:43:33 -0800
Message-ID: <A1171C2850D8D4119D7B0002A508C76401466831@bgsmsx91.iind.intel.com>
From: "Buraga, Partha S" <p_buraga@trillium.com>
To: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Date: Mon, 25 Feb 2002 03:41:33 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: [Tsvwg] Regarding Scope of IPv6 addresses
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi

If an Ipv6 address is Link Local address, would it mean, it has scope to
Site Local and Global ?
If an Ipv6 address is Site Local address, then the scope will be only for
Site or for both Site and Global ?
Pls clarify my doubts.

thanks
partha

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 25 06:59:11 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA06359
	for <tsvwg-archive@odin.ietf.org>; Mon, 25 Feb 2002 06:59:11 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA14358
	for tsvwg-archive@odin.ietf.org; Mon, 25 Feb 2002 06:59:14 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA13915;
	Mon, 25 Feb 2002 06:44:17 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA13886
	for <tsvwg@optimus.ietf.org>; Mon, 25 Feb 2002 06:44:15 -0500 (EST)
Received: from stewartlin.chicago.il.us (host-64-47-24-166.masergy.com [64.47.24.166] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA06006
	for <tsvwg@ietf.org>; Mon, 25 Feb 2002 06:44:11 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewartlin.chicago.il.us (8.11.0/8.11.0) with ESMTP id g1PBjWr29565;
	Mon, 25 Feb 2002 05:45:32 -0600
Message-ID: <3C7A2387.D137251E@stewart.chicago.il.us>
Date: Mon, 25 Feb 2002 05:44:08 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] Some other questions about the AddIP extension
References: <58AEDE27D48F884285007ABF827FF263257B41@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Debugging my implementation I'm thinking about strange situations...
> 
>         What will happen if the peer tries to add an address that is being used by another association within this same endpoint (of course, both associations should be using the same destination port as well)? Silently discard the chunk (and maybe the whole datagram)?


So are you saying


Host 1                                         Host-2

IP-A                                          IP-Z
IP-B                                          IP-Y
IP-C                                          IP-X

(IP-A,     <--------------asoc 1-------------------> (IP-Z,IP-Z,
port:5555)
 IP-B,
 port:4444)

(IP-C:port4444)<--------------asoc 2------------------->(IP-X:port:5555)

Then assoc 2 attempts to do
         addip(IP-A)------------------------------>


> 
>         Thinking about my previous mail... There is a really weird case when the second ASCONF ACK might be different. When the peer (an attacker?) tells us to add an address that is being used by another association, but that association deletes that address before we receive the second ASCONF chunk... And there might be more situations (ok, REALLY strange ones... :-)).

So you are also saying that the attacker has guessed both the
sequence number and the V-Tag on the association?


> So this makes me think harder that either the last ASCONF ACK chunk should be compulsorily kept in memory (in fact, I think that this is what all of us are going to do), or it is stated in the draft that simply re-parsing the ASCONF chunk might not be the right thing to do... :-0

But that is an implementation decision Ivan.. You are REQUIRED TO ANSWER
the SAME.. How you do it is up to you but you MUST answer the same.

Please don't confuse requirements with implemention decisions..

It is like the SACK Gap-Acks.. I can think of NO easy way to
list the Gap-Acks to my peer out of order.. but we do NOT make
a requirement that they be in order since some future person
may come up with something brilliant that makes this happen.

There is NO reason to dictate implementation in the spec only
require what must be done..

> 
>         What should we do if the peer deletes all the active addresses (i.e., after the deletion we only have inactive addresses)? Should we close the association? :-/ The ASCONF chunk must have come from one of the inactive addresses, so the card and the return path are working... Maybe we should try with the inactive addresses until we reach the Max.Association.Retrans... I guess this might be an implementation specific issue...
> 

Well this is an issue no matter if you have add-ip or not.. what does
your implementation do if all the addresses go in-active?

Ours continues to attempt to both send and heartbeat on the
in-active addresses until Max.association.retrans is reached...

Add-ip should not influence this...






-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 25 08:02:11 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA08269
	for <tsvwg-archive@odin.ietf.org>; Mon, 25 Feb 2002 08:02:11 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA17917
	for tsvwg-archive@odin.ietf.org; Mon, 25 Feb 2002 08:02:14 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA16789;
	Mon, 25 Feb 2002 07:48:09 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA16758
	for <tsvwg@optimus.ietf.org>; Mon, 25 Feb 2002 07:48:07 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA07597
	for <tsvwg@ietf.org>; Mon, 25 Feb 2002 07:48:02 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir03nok.nokia.com (esvir03nokt.ntc.nokia.com [172.21.143.35])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1PCmDZ26313
	for <tsvwg@ietf.org>; Mon, 25 Feb 2002 14:48:13 +0200 (EET)
Received: from esebh003.NOE.Nokia.com (unverified) by esvir03nok.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T594a2bf284ac158f23078@esvir03nok.nokia.com>;
 Mon, 25 Feb 2002 14:48:03 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh003.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Mon, 25 Feb 2002 14:48:03 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] Some other questions about the AddIP extension
Date: Mon, 25 Feb 2002 14:48:03 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF26308EED8@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] Some other questions about the AddIP extension
Thread-Index: AcG98b+Pzc2IQ4mEREaT1z6qiMSjMwAB2NCg
To: <randall@stewart.chicago.il.us>
Cc: <tsvwg@ietf.org>
X-OriginalArrivalTime: 25 Feb 2002 12:48:03.0375 (UTC) FILETIME=[A9F1BBF0:01C1BDFA]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id HAA16759
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

> Ivan.Arias-Rodriguez@nokia.com wrote:
> > 
> >         Debugging my implementation I'm thinking about 
> strange situations...
> > 
> >         What will happen if the peer tries to add an 
> address that is being used by another association within this 
> same endpoint (of course, both associations should be using 
> the same destination port as well)? Silently discard the 
> chunk (and maybe the whole datagram)?
> 
> 
> So are you saying
> 
> 
> Host 1                                         Host-2
> 
> IP-A                                          IP-Z
> IP-B                                          IP-Y
> IP-C                                          IP-X
> 
> (IP-A,     <--------------asoc 1-------------------> (IP-Z,IP-Z, port:5555)
>  IP-B,
>  port:4444)
> 
> (IP-C:port4444)<--------------asoc 2------------------->(IP-X:port:5555)
> 
> Then assoc 2 attempts to do
>          addip(IP-A)------------------------------>
> 

	Well, not exactly. If asoc 1 has "IP-X, IP-Z" then the answer is yes. Because (IP-A, 4444) <-> (IP-X, 5555) would be a valid address/port pair in both associations...

> > 
> >         Thinking about my previous mail... There is a 
> really weird case when the second ASCONF ACK might be 
> different. When the peer (an attacker?) tells us to add an 
> address that is being used by another association, but that 
> association deletes that address before we receive the second 
> ASCONF chunk... And there might be more situations (ok, 
> REALLY strange ones... :-)).
> 
> So you are also saying that the attacker has guessed both the
> sequence number and the V-Tag on the association?

	No. I was thinking about another case where simply re-parsing the ASCONF chunk is not enough to assure that the resulting ASCONF ACK would be the same. It would be like being in the situation shown in the graph above, and then the addip(IP-A) is resent but meanwhile, asoc 1 sent delete(IP-A). Thus, initially the addip(IP-A) was illegal (and my question continues, what should be done?), but upon receipt of the retransmitted ASCONF it would be a valid situation...

> > So this makes me think harder that either the last ASCONF 
> ACK chunk should be compulsorily kept in memory (in fact, I 
> think that this is what all of us are going to do), or it is 
> stated in the draft that simply re-parsing the ASCONF chunk 
> might not be the right thing to do... :-0
> 
> But that is an implementation decision Ivan.. You are 
> REQUIRED TO ANSWER
> the SAME.. How you do it is up to you but you MUST answer the same.
> 
> Please don't confuse requirements with implemention decisions..

	Ok, but then I think that having a note somewhere in C3 saying that simply reparsing the ASCONF chunk might not be enough to assure that we are producing the same ASCONF ACK as we originally did...

> It is like the SACK Gap-Acks.. I can think of NO easy way to
> list the Gap-Acks to my peer out of order.. but we do NOT make
> a requirement that they be in order since some future person
> may come up with something brilliant that makes this happen.
> 
> There is NO reason to dictate implementation in the spec only
> require what must be done..
> 
> > 
> >         What should we do if the peer deletes all the 
> active addresses (i.e., after the deletion we only have 
> inactive addresses)? Should we close the association? :-/ The 
> ASCONF chunk must have come from one of the inactive 
> addresses, so the card and the return path are working... 
> Maybe we should try with the inactive addresses until we 
> reach the Max.Association.Retrans... I guess this might be an 
> implementation specific issue...
> > 
> 
> Well this is an issue no matter if you have add-ip or not.. what does
> your implementation do if all the addresses go in-active?
> 
> Ours continues to attempt to both send and heartbeat on the
> in-active addresses until Max.association.retrans is reached...
> 
> Add-ip should not influence this...

	Ok!

	Thanks for your answer!

	BR Iván Arias Rodríguez

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 25 08:17:12 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA08746
	for <tsvwg-archive@odin.ietf.org>; Mon, 25 Feb 2002 08:17:12 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA18596
	for tsvwg-archive@odin.ietf.org; Mon, 25 Feb 2002 08:17:15 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA17988;
	Mon, 25 Feb 2002 08:03:15 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA17957
	for <tsvwg@optimus.ietf.org>; Mon, 25 Feb 2002 08:03:12 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA08329
	for <tsvwg@ietf.org>; Mon, 25 Feb 2002 08:03:03 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir04nok.ntc.nokia.com (esvir04nokt.ntc.nokia.com [172.21.143.36])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1PD3FZ08132
	for <tsvwg@ietf.org>; Mon, 25 Feb 2002 15:03:15 +0200 (EET)
Received: from esebh003.NOE.Nokia.com (unverified) by esvir04nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T594a39b51dac158f24077@esvir04nok.ntc.nokia.com>;
 Mon, 25 Feb 2002 15:03:05 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh003.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Mon, 25 Feb 2002 15:03:05 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] More questions about the AddIP extension
Date: Mon, 25 Feb 2002 15:03:04 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF26308EED9@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] More questions about the AddIP extension
Thread-Index: AcG98SMhzz/rGcvPRHmXGMS1c9YzRQACWXaw
To: <randall@stewart.chicago.il.us>
Cc: <tsvwg@ietf.org>
X-OriginalArrivalTime: 25 Feb 2002 13:03:05.0038 (UTC) FILETIME=[C36096E0:01C1BDFC]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id IAA17958
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hei!

> -----Original Message-----
> From: ext Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: 25. February 2002 13:35
> To: Arias-Rodriguez Ivan (NRC/Helsinki)
> Cc: tsvwg@ietf.org
> Subject: Re: [Tsvwg] More questions about the AddIP extension
> 
> 
> Ivan.Arias-Rodriguez@nokia.com wrote:
> > 
> >         Hi again!
> > 
> >         Suddenly I realized about one thing. The AddIP 
> extension is designed in a way that if the ASCONF ACK chunk 
> is lost and the ASCONF chunk is resent, the ASCONF receiver 
> should answer in the same way as if it would have never 
> received the first ASCONF chunk, not modifying anything as it 
> should have done it the first time. This is of course the 
> right way of doing.
> > 
> 
> Yep...
> 
> >         But, there is an implementation note in page 15 
> that suggests keeping the ASCONF ACK chunk sent as an 
> "optimization"... I tend to think that this is not an 
> optimization but a compulsory thing to do... The reason is 
> that when parsing the ASCONF chunk for the second time the 
> answer might be different because:
> > 
> >         - If we agree that we treat as a success adding an 
> address already in use or deleting an address that was not in 
> use, then the answer to the second ASCONF (including 
> additions and deletions of addresses) will be effectively the 
> same. However, in this case I think that this fact 
> (considering these situations as successes) should be stated 
> in the draft.
> > 
> >         - I can think about another weird but still valid 
> situation. If in the ASCONF the peer tells us to set our 
> primary address to A, and in the same chunk it tells us in a 
> subsequent parameter to delete that address. Upon the receipt 
> of the second ASCONF, the "Set Primary Address" parameter 
> would fail, since that address is not in use... The ASCONF 
> sender is doing strange things, but for example Peter Lei's 
> implementation would produce a different answer for the 
> second ASCONF... :-/
> 
> No... you are required to give the same answer.. 

	Yes. But then, for example, answering with an "Unresolvable Address" error cause to a Set Primary if the address stated is not in use can cause problems. In the example where there is a (weird) ASCONF that requires to set one address as the primary and then it deletes it, if we receive a duplicate ASCONF and re-parse it, we will generate an "Unresolvable Address" while the first time we simply accepted it... Thus, you really need to know what happened before...

	Well, now that I think about it, you could always "undo" the changes required by the ASCONF to be in the same state you were... But then you are not sure either if you did that last time... :-0

	In any case, I think that this discussion is not really important, as these things happen when the ASCONF sender is playing strange games... :-)

	However, that case in the other mail about adding an address that is in use in another association, but due to a deletion in becomes a valid addition in the second ASCONF is a little bit more tricky... The ASCONF sender could receive two ASCONF ACK, one saying that the address was not added, and another one stating the opposite... But I'm not sure I'm young enough to see such corner case situation ever... :-)

> > 
> >         So, I think that either keeping the ASCONF ACK 
> chunk is compulsory, or the behavior in these "corner case" 
> situations (actually the first case is not a corner case at 
> all) should be clearly stated in the draft...
> > 
> 
> 
> I disagree here Ivan... there may be some better way than keeping the
> asconf-ack around. Peter and I do this in our implementation but
> let us not limit implementors. It is required they they always
> answer the same but let's not dictate how you implement that.

	Yes sure...

	Then, if you add to the IMPLEMENTATION NOTE something like "But no matter how it is implemented, the ASCONF ACK sender must be sure that the ASCONF ACK chunk sent upon the receipt of a duplicate ASCONF chunk is exactly the same it sends when it received the original ASCONF chunk", I won't disturb you anymore ;-). I you don't, I won't disturb you either :->

	Thanks for your answer!

	BR Iván Arias Rodríguez

> I could also image an implementation that keeps a state history
> of some kind and thus can reconstruct a state (ip addresses, primaries
> and such) at any given time. So when an ASCONF-ACK arrives it
> can send back the exact same answer. Remember you can only
> have ONE ASCONF out at a time.. so in a state keeping implementation
> you would only have to keep one set of state (if you were worried
> about space)..
> 
> 
> R
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 25 12:00:13 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA18636
	for <tsvwg-archive@odin.ietf.org>; Mon, 25 Feb 2002 12:00:13 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id MAA03108
	for tsvwg-archive@odin.ietf.org; Mon, 25 Feb 2002 12:00:17 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA01784;
	Mon, 25 Feb 2002 11:47:58 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id LAA01707
	for <tsvwg@optimus.ietf.org>; Mon, 25 Feb 2002 11:47:54 -0500 (EST)
Received: from rwcrmhc53.attbi.com (rwcrmhc53.attbi.com [204.127.198.39])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA17757
	for <tsvwg@ietf.org>; Mon, 25 Feb 2002 11:47:50 -0500 (EST)
Received: from ieee.org ([12.248.77.94]) by rwcrmhc53.attbi.com
          (InterMail vM.4.01.03.27 201-229-121-127-20010626) with ESMTP
          id <20020225164723.RNQG2951.rwcrmhc53.attbi.com@ieee.org>;
          Mon, 25 Feb 2002 16:47:23 +0000
Message-ID: <3C7A6A7D.7A6ECBB4@ieee.org>
Date: Mon, 25 Feb 2002 10:46:53 -0600
From: Peter Lei <peter.lei@ieee.org>
X-Mailer: Mozilla 4.79 [en] (Windows NT 5.0; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: randall@stewart.chicago.il.us, tsvwg@ietf.org
Subject: Re: [Tsvwg] More questions about the AddIP extension
References: <58AEDE27D48F884285007ABF827FF26308EED9@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Hei!
> 
> > -----Original Message-----
> > From: ext Randall Stewart [mailto:randall@stewart.chicago.il.us]
> > Sent: 25. February 2002 13:35
> > To: Arias-Rodriguez Ivan (NRC/Helsinki)
> > Cc: tsvwg@ietf.org
> > Subject: Re: [Tsvwg] More questions about the AddIP extension
> >
> >
> > Ivan.Arias-Rodriguez@nokia.com wrote:
> > >
> > >         Hi again!
> > >
> > >         Suddenly I realized about one thing. The AddIP
> > extension is designed in a way that if the ASCONF ACK chunk
> > is lost and the ASCONF chunk is resent, the ASCONF receiver
> > should answer in the same way as if it would have never
> > received the first ASCONF chunk, not modifying anything as it
> > should have done it the first time. This is of course the
> > right way of doing.
> > >
> >
> > Yep...
> >
> > >         But, there is an implementation note in page 15
> > that suggests keeping the ASCONF ACK chunk sent as an
> > "optimization"... I tend to think that this is not an
> > optimization but a compulsory thing to do... The reason is
> > that when parsing the ASCONF chunk for the second time the
> > answer might be different because:
> > >
> > >         - If we agree that we treat as a success adding an
> > address already in use or deleting an address that was not in
> > use, then the answer to the second ASCONF (including
> > additions and deletions of addresses) will be effectively the
> > same. However, in this case I think that this fact
> > (considering these situations as successes) should be stated
> > in the draft.
> > >
> > >         - I can think about another weird but still valid
> > situation. If in the ASCONF the peer tells us to set our
> > primary address to A, and in the same chunk it tells us in a
> > subsequent parameter to delete that address. Upon the receipt
> > of the second ASCONF, the "Set Primary Address" parameter
> > would fail, since that address is not in use... The ASCONF
> > sender is doing strange things, but for example Peter Lei's
> > implementation would produce a different answer for the
> > second ASCONF... :-/
> >
> > No... you are required to give the same answer..
> 
>         Yes. But then, for example, answering with an "Unresolvable Address" error cause to a Set Primary if the address stated is not in use can cause problems. In the example where there is a (weird) ASCONF that requires to set one address as the primary and then it deletes it, if we receive a duplicate ASCONF and re-parse it, we will generate an "Unresolvable Address" while the first time we simply accepted it... Thus, you really need to know what happened before...
> 
>         Well, now that I think about it, you could always "undo" the changes required by the ASCONF to be in the same state you were... But then you are not sure either if you did that last time... :-0

Right, the implementation does need to know what the state is before the first
ASCONF was received and processed;  in the scenario described, we still would
NOT send a (second) ACK with a different result.  Since there can only be one
outstanding ASCONF at any given time per association, there is only a single
set of "previous states" to keep around to "undo the changes" before responding.
*How* it does this should be implementation specific...

>         In any case, I think that this discussion is not really important, as these things happen when the ASCONF sender is playing strange games... :-)
> 
>         However, that case in the other mail about adding an address that is in use in another association, but due to a deletion in becomes a valid addition in the second ASCONF is a little bit more tricky... The ASCONF sender could receive two ASCONF ACK, one saying that the address was not added, and another one stating the opposite... But I'm not sure I'm young enough to see such corner case situation ever... :-)
> 
> > >
> > >         So, I think that either keeping the ASCONF ACK
> > chunk is compulsory, or the behavior in these "corner case"
> > situations (actually the first case is not a corner case at
> > all) should be clearly stated in the draft...
> > >
> >
> >
> > I disagree here Ivan... there may be some better way than keeping the
> > asconf-ack around. Peter and I do this in our implementation but
> > let us not limit implementors. It is required they they always
> > answer the same but let's not dictate how you implement that.
> 
>         Yes sure...
> 
>         Then, if you add to the IMPLEMENTATION NOTE something like "But no matter how it is implemented, the ASCONF ACK sender must be sure that the ASCONF ACK chunk sent upon the receipt of a duplicate ASCONF chunk is exactly the same it sends when it received the original ASCONF chunk", I won't disturb you anymore ;-). I you don't, I won't disturb you either :->
> 
>         Thanks for your answer!
> 
>         BR Iván Arias Rodríguez
> 
> > I could also image an implementation that keeps a state history
> > of some kind and thus can reconstruct a state (ip addresses, primaries
> > and such) at any given time. So when an ASCONF-ACK arrives it
> > can send back the exact same answer. Remember you can only
> > have ONE ASCONF out at a time.. so in a state keeping implementation
> > you would only have to keep one set of state (if you were worried
> > about space)..
> >
> >
> > R
> >
> > --
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> >
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 25 13:53:39 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA24586
	for <tsvwg-archive@odin.ietf.org>; Mon, 25 Feb 2002 13:53:39 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id NAA10618
	for tsvwg-archive@odin.ietf.org; Mon, 25 Feb 2002 13:53:42 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA09915;
	Mon, 25 Feb 2002 13:38:36 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id NAA09876
	for <tsvwg@optimus.ietf.org>; Mon, 25 Feb 2002 13:38:33 -0500 (EST)
Received: from mgw-x3.nokia.com (mgw-x3.nokia.com [131.228.20.26])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA23937;
	Mon, 25 Feb 2002 13:38:29 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir01nok.ntc.nokia.com (esvir01nokt.ntc.nokia.com [172.21.143.33])
	by mgw-x3.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1PIcvi04454;
	Mon, 25 Feb 2002 20:38:58 +0200 (EET)
Received: from esebh003.NOE.Nokia.com (unverified) by esvir01nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T594b6cb5c8ac158f21082@esvir01nok.ntc.nokia.com>;
 Mon, 25 Feb 2002 20:38:24 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh003.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Mon, 25 Feb 2002 20:38:24 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Date: Mon, 25 Feb 2002 20:38:24 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF26308EEE1@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] Some other questions about the AddIP extension
Thread-Index: AcG98b+Pzc2IQ4mEREaT1z6qiMSjMwAN+caA
To: <tsvwg@ietf.org>
Cc: <sigtran@ietf.org>
X-OriginalArrivalTime: 25 Feb 2002 18:38:24.0811 (UTC) FILETIME=[9BB24FB0:01C1BE2B]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id NAA09877
Subject: [Tsvwg] My Master's Thesis about SCTP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi all!

	I finished my Master's Thesis about SCTP some weeks ago. It is mostly about the history of its design since it was a little MDTP baby.

	Andreas Jungmaier has made some space for the .pdf version in its nice SCTP page. If you are not especially an SCTP guru and you are afraid of RFCs, go and check his excellent "SCTP For Beginners" (not only) page.

	http://tdrwww.exp-math.uni-essen.de/pages/forschung/sctp_fb/index.html

	And if you feel like wanting to know why something is like it is go and check 

	http://tdrwww.exp-math.uni-essen.de/pages/forschung/sctp_fb/sctp_links.html

it could be that I even wrote something about it in my Master's Thesis (at the end of the "Drafts and Documents" subsection)...

	Thanks!

	BR Iván Arias Rodríguez

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 25 14:15:04 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA25638
	for <tsvwg-archive@odin.ietf.org>; Mon, 25 Feb 2002 14:15:04 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA12120
	for tsvwg-archive@odin.ietf.org; Mon, 25 Feb 2002 14:15:07 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA11469;
	Mon, 25 Feb 2002 14:02:44 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA11400
	for <tsvwg@optimus.ietf.org>; Mon, 25 Feb 2002 14:02:41 -0500 (EST)
Received: from stewartlin.chicago.il.us (host-64-47-24-166.masergy.com [64.47.24.166] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA25118;
	Mon, 25 Feb 2002 14:02:36 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewartlin.chicago.il.us (8.11.0/8.11.0) with ESMTP id g1PJ3wr30205;
	Mon, 25 Feb 2002 13:03:58 -0600
Message-ID: <3C7A8A4C.5DB8C477@stewart.chicago.il.us>
Date: Mon, 25 Feb 2002 13:02:36 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: Ivan.Arias-Rodriguez@nokia.com
CC: tsvwg@ietf.org, sigtran@ietf.org
Subject: Re: [Tsvwg] My Master's Thesis about SCTP
References: <58AEDE27D48F884285007ABF827FF26308EEE1@esebe017.NOE.Nokia.com>
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

Ivan:

Thanks for the reminder :> I also have it available on

http://www.sctp.org

R

Ivan.Arias-Rodriguez@nokia.com wrote:
> 
>         Hi all!
> 
>         I finished my Master's Thesis about SCTP some weeks ago. It is mostly about the history of its design since it was a little MDTP baby.
> 
>         Andreas Jungmaier has made some space for the .pdf version in its nice SCTP page. If you are not especially an SCTP guru and you are afraid of RFCs, go and check his excellent "SCTP For Beginners" (not only) page.
> 
>         http://tdrwww.exp-math.uni-essen.de/pages/forschung/sctp_fb/index.html
> 
>         And if you feel like wanting to know why something is like it is go and check
> 
>         http://tdrwww.exp-math.uni-essen.de/pages/forschung/sctp_fb/sctp_links.html
> 
> it could be that I even wrote something about it in my Master's Thesis (at the end of the "Drafts and Documents" subsection)...
> 
>         Thanks!
> 
>         BR Iván Arias Rodríguez
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 25 14:54:49 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA27304
	for <tsvwg-archive@odin.ietf.org>; Mon, 25 Feb 2002 14:54:48 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA14187
	for tsvwg-archive@odin.ietf.org; Mon, 25 Feb 2002 14:54:52 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA13773;
	Mon, 25 Feb 2002 14:44:31 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA13742
	for <tsvwg@optimus.ietf.org>; Mon, 25 Feb 2002 14:44:29 -0500 (EST)
Received: from sj-msg-core-3.cisco.com (sj-msg-core-3.cisco.com [171.70.157.152])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA26875;
	Mon, 25 Feb 2002 14:44:21 -0500 (EST)
Received: from mira-sjcm-3.cisco.com (mira-sjcm-3.cisco.com [171.69.24.15])
	by sj-msg-core-3.cisco.com (8.11.3/8.9.1) with ESMTP id g1PJhPw22550;
	Mon, 25 Feb 2002 11:43:25 -0800 (PST)
Received: from ALLYN-W2K.cisco.com (altiga-vpn-client-19.cisco.com [10.64.194.19])
	by mira-sjcm-3.cisco.com (Mirapoint)
	with ESMTP id ADS82932;
	Mon, 25 Feb 2002 11:32:56 -0800 (PST)
Message-Id: <4.3.2.7.2.20020225111728.02338928@mira-sjcm-3.cisco.com>
X-Sender: allyn@mira-sjcm-3.cisco.com
X-Mailer: QUALCOMM Windows Eudora Version 4.3.2
Date: Mon, 25 Feb 2002 11:43:18 -0800
To: ietf@ietf.org, tsvwg@ietf.org, ips@ece.cmu.edu,
        dafs-discussions@yahoogroups.com, rdma@yahoogroups.com
From: allyn romanow <allyn@cisco.com>
Cc: allyn@cisco.com, stephen Bailey <steph@cs.uchicago.edu>, sob@harvard.edu,
        mankin@isi.edu
Mime-Version: 1.0
Content-Type: multipart/alternative;
	boundary="=====================_7536957==_.ALT"
Subject: [Tsvwg] RDMA over IP Internet Draft and BOF at 53rd IETF
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

--=====================_7536957==_.ALT
Content-Type: text/plain; charset="us-ascii"; format=flowed

There is a new problem statement for RDMA over IP.
Discussion is welcome on the RDMA over IP mailing list at
         http://groups.yahoo.com/group/rdma/

A New Internet-Draft is available from the on-line Internet-Drafts directories.


         Title           : RDMA over IP Problem Statement
         Author(s)       : A. Romanow et al.
         Filename        : draft-romanow-rdma-over-ip-problem-statement-00.txt
         Pages           : 17
         Date            : 21-Feb-02

This draft describes the problem that copying in network I/O
typically causes high system costs in end-hosts at high speeds.
The problem is due to the high cost of memory bandwidth, and it can
be substantially improved using 'copy avoidance.' The high overhead
has prevented TCP/IP from being used as an interconnection network,
and instead special purpose memory-to-memory fabrics have been
developed and widely used.  An IP-based solution, developed within
the IETF, is desirable for interoperability of various network
fabrics. It is also particularly important for the IETF to guide
the standardization because interconnection technology will soon
start to be used over the wide area in the Internet

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-romanow-rdma-over-ip-problem-statement-00.txt



There will be a BOF for RDMA over IP (ROI) at the next IETF, scheduled for 
Tuesday March 19 13:00-14:00. The announcement is at

http://ietf.org/ietf/02mar/roi.txt

The BOF agenda is subject to change.

There will be a proposed charter available for discussion later this week at
http://www.cs.uchicago.edu/~steph/roi_charter.txt



--=====================_7536957==_.ALT
Content-Type: text/html; charset="us-ascii"

<html>
There is a new problem statement for RDMA over IP. <br>
Discussion is welcome on the RDMA over IP mailing list at<br>
&nbsp;<x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab><a href="http://groups.yahoo.com/group/rdma/" eudora="autourl">http://groups.yahoo.com/group/rdma/<br>
<br>
</a>A New Internet-Draft is available from the on-line Internet-Drafts
directories.<br>
<br>
<br>
<x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab>Title<x-tab>&nbsp;&nbsp;&nbsp;</x-tab><x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab>:
RDMA over IP Problem Statement<br>
<x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab>Author(s)<x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab>:
A. Romanow et al.<br>
<x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab>Filename<x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab>:
draft-romanow-rdma-over-ip-problem-statement-00.txt<br>
<x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab>Pages<x-tab>&nbsp;&nbsp;&nbsp;</x-tab><x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab>:
17<br>
<x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab>Date<x-tab>&nbsp;&nbsp;&nbsp;&nbsp;</x-tab><x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab>:
21-Feb-02<br>
<x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab><br>
This draft describes the problem that copying in network I/O<br>
typically causes high system costs in end-hosts at high speeds.<br>
The problem is due to the high cost of memory bandwidth, and it can<br>
be substantially improved using 'copy avoidance.' The high overhead<br>
has prevented TCP/IP from being used as an interconnection network,<br>
and instead special purpose memory-to-memory fabrics have been<br>
developed and widely used.&nbsp; An IP-based solution, developed
within<br>
the IETF, is desirable for interoperability of various network<br>
fabrics. It is also particularly important for the IETF to guide<br>
the standardization because interconnection technology will soon<br>
start to be used over the wide area in the Internet<br>
<br>
A URL for this Internet-Draft is:<br>
<a href="http://www.ietf.org/internet-drafts/draft-romanow-rdma-over-ip-problem-statement-00.txt" eudora="autourl">http://www.ietf.org/internet-drafts/draft-romanow-rdma-over-ip-problem-statement-00.</a><a href="http://www.ietf.org/internet-drafts/draft-romanow-rdma-over-ip-problem-statement-00.txt" eudora="autourl">txt<br>
<br>
<br>
<br>
</a>There will be a BOF for RDMA over IP (ROI) at the next IETF,
scheduled for Tuesday March 19 13:00-14:00. The announcement is at<br>
<br>
<a href="http://ietf.org/ietf/02mar/roi.txt" eudora="autourl">http://ietf.org/ietf/02mar/roi.txt</a><br>
&nbsp;<br>
The BOF agenda is subject to change.<br>
<br>
There will be a proposed charter available for discussion later this week
at <br>
<font color="#0000FF"><u><a href="http://www.cs.uchicago.edu/~steph/roi_charter.txt" eudora="autourl">http://www.cs.uchicago.edu/~s</a></u>teph/roi_charter.<a href="http://www.cs.uchicago.edu/~steph/roi_charter.txt" eudora="autourl">txt<br>
<br>
</a></font>&nbsp;<br>
</html>

--=====================_7536957==_.ALT--


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Mon Feb 25 18:30:08 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA03980
	for <tsvwg-archive@odin.ietf.org>; Mon, 25 Feb 2002 18:30:08 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id SAA24773
	for tsvwg-archive@odin.ietf.org; Mon, 25 Feb 2002 18:30:11 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA24288;
	Mon, 25 Feb 2002 18:18:56 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA24259
	for <tsvwg@optimus.ietf.org>; Mon, 25 Feb 2002 18:18:55 -0500 (EST)
Received: from atlrel8.hp.com (atlrel8.hp.com [156.153.255.206])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA03610
	for <tsvwg@ietf.org>; Mon, 25 Feb 2002 18:18:50 -0500 (EST)
Received: from xatlrelay2.atl.hp.com (xatlrelay2.atl.hp.com [15.45.89.191])
	by atlrel8.hp.com (Postfix) with ESMTP id EF637A0023C
	for <tsvwg@ietf.org>; Mon, 25 Feb 2002 18:18:24 -0500 (EST)
Received: from xatlbh3.atl.hp.com (xatlbh3.atl.hp.com [15.45.89.188])
	by xatlrelay2.atl.hp.com (Postfix) with ESMTP id 6FBFB400085
	for <tsvwg@ietf.org>; Mon, 25 Feb 2002 18:18:24 -0500 (EST)
Received: by xatlbh3.atl.hp.com with Internet Mail Service (5.5.2653.19)
	id <FQLBT3KM>; Mon, 25 Feb 2002 18:18:24 -0500
Message-ID: <499DC368E25AD411B3F100902740AD650AFC7DC9@xrose03.rose.hp.com>
From: "WENDT,JIM (HP-Roseville,ex1)" <jim_wendt@hp.com>
To: tsvwg@ietf.org
Date: Mon, 25 Feb 2002 18:18:23 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: [Tsvwg] SCTP and IP Frag aliasing on LFNs ?
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Based upon following in RFC-2960:

> 6.9 Fragmentation and Reassembly
>   Note: Once a message is fragmented it cannot be re-fragmented.
>   Instead if the PMTU has been reduced, then IP fragmentation must be
>   used.  Please see Section 7.3 for details of PMTU discovery.

> 7.3 Path MTU Discovery
>   4) Since data transmission in SCTP is naturally structured in terms
>      of TSNs rather than bytes (as is the case for TCP), the discussion
>      in Section 6.5 of RFC 1191 applies: When retransmitting an IP
>      datagram to a remote address for which the IP datagram appears too
>      large for the path MTU to that address, the IP datagram SHOULD be
>      retransmitted without the DF bit set, allowing it to possibly be
>      fragmented.  Transmissions of new IP datagrams MUST have DF set.

Am I correct in understanding that IP fragmentation can occur
on SCTP packets across a IPv4 PMTU degrade, and that worst case,
IP fragmentation can occur for all packets in flight up to the
roundtrip pipe size between the sending endpoint and the last
hop router?

If so, and the pipe is large, has there been any discussion or
concern regarding aliasing or wrap-around of the 16-bit
Identification (ID) field in the IP header?  Is this an
issue, and if so, how is it being addressed?

Thx,
Jim


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 26 06:36:38 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA23386
	for <tsvwg-archive@odin.ietf.org>; Tue, 26 Feb 2002 06:36:38 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA06974
	for tsvwg-archive@odin.ietf.org; Tue, 26 Feb 2002 06:36:38 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA05868;
	Tue, 26 Feb 2002 06:14:32 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA05840
	for <tsvwg@optimus.ietf.org>; Tue, 26 Feb 2002 06:14:30 -0500 (EST)
Received: from c003.snv.cp.net (c003-h004.c003.snv.cp.net [209.228.32.218])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id GAA23082
	for <tsvwg@ietf.org>; Tue, 26 Feb 2002 06:14:25 -0500 (EST)
Received: (cpmta 12466 invoked from network); 26 Feb 2002 03:13:59 -0800
Received: from 64.130.130.105 (HELO dougrmt)
  by smtp.telocity.com (209.228.32.218) with SMTP; 26 Feb 2002 03:13:59 -0800
X-Sent: 26 Feb 2002 11:13:59 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: "Tsvwg" <tsvwg@ietf.org>
Date: Tue, 26 Feb 2002 03:16:58 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJMEOICPAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Importance: Normal
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] SCTP-DDP Shim Posted
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Files related to implementing a DDP shim for SCTP has been posted and can be
immedately seen at:

http://groups.yahoo.com/group/SCTP-DDP/files/draft-otis-network-overhead-00.
txt

http://groups.yahoo.com/group/SCTP-DDP/files/draft-stewart-otis-sctp-ddp-00.
txt

This shim has been simplifed to rely on the Payload Protocol Identifer as a
means of asserting ULP extensions such as RDMA or RPC as example.

Doug


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 26 07:14:40 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA25039
	for <tsvwg-archive@odin.ietf.org>; Tue, 26 Feb 2002 07:14:40 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA08987
	for tsvwg-archive@odin.ietf.org; Tue, 26 Feb 2002 07:14:42 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA08311;
	Tue, 26 Feb 2002 07:02:38 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA08285
	for <tsvwg@optimus.ietf.org>; Tue, 26 Feb 2002 07:02:36 -0500 (EST)
Received: from CNRI.Reston.VA.US (localhost [127.0.0.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA23996;
	Tue, 26 Feb 2002 07:02:33 -0500 (EST)
Message-Id: <200202261202.HAA23996@ietf.org>
Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce: ;
CC: tsvwg@ietf.org
From: Internet-Drafts@ietf.org
Reply-to: Internet-Drafts@ietf.org
Date: Tue, 26 Feb 2002 07:02:33 -0500
Subject: [Tsvwg] I-D ACTION:draft-riegel-tuexen-mobile-sctp-00.txt
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.


	Title		: Mobile SCTP
	Author(s)	: M. Riegel, M. Tuexen
	Filename	: draft-riegel-tuexen-mobile-sctp-00.txt
	Pages		: 8
	Date		: 25-Feb-02
	
Transport layer mobility management is proposed as an alternative to
Mobile IP for realizing seamless mobility in the Internet. Mobility
management is accomplished by use of SCTP (Stream Control Transmission
Protocol) and some of its currently proposed extensions. With mobile
SCTP seamless handover even crossing physical, link and network layer
technologies is fully realized in the mobile client at the end of the
network not needing any additional provisions in the network. Mobile
services can be provided to such clients by mobile SCTP enabled servers
from everywhere in the Internet.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-riegel-tuexen-mobile-sctp-00.txt

To remove yourself from the IETF Announcement list, send a message to 
ietf-announce-request with the word unsubscribe in the body of the message.

Internet-Drafts are also available by anonymous FTP. Login with the username
"anonymous" and a password of your e-mail address. After logging in,
type "cd internet-drafts" and then
	"get draft-riegel-tuexen-mobile-sctp-00.txt".

A list of Internet-Drafts directories can be found in
http://www.ietf.org/shadow.html 
or ftp://ftp.ietf.org/ietf/1shadow-sites.txt


Internet-Drafts can also be obtained by e-mail.

Send a message to:
	mailserv@ietf.org.
In the body type:
	"FILE /internet-drafts/draft-riegel-tuexen-mobile-sctp-00.txt".
	
NOTE:	The mail server at ietf.org can return the document in
	MIME-encoded form by using the "mpack" utility.  To use this
	feature, insert the command "ENCODING mime" before the "FILE"
	command.  To decode the response(s), you will need "munpack" or
	a MIME-compliant mail reader.  Different MIME-compliant mail readers
	exhibit different behavior, especially when dealing with
	"multipart" MIME messages (i.e. documents which have been split
	up into multiple messages), so check your local documentation on
	how to manipulate these messages.
		
		
Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Multipart/Alternative; Boundary="OtherAccess"

--OtherAccess
Content-Type: Message/External-body;
	access-type="mail-server";
	server="mailserv@ietf.org"

Content-Type: text/plain
Content-ID:	<20020225142641.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-riegel-tuexen-mobile-sctp-00.txt

--OtherAccess
Content-Type: Message/External-body;
	name="draft-riegel-tuexen-mobile-sctp-00.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID:	<20020225142641.I-D@ietf.org>

--OtherAccess--

--NextPart--



_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 26 07:17:04 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA25060
	for <tsvwg-archive@odin.ietf.org>; Tue, 26 Feb 2002 07:17:04 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id HAA09019
	for tsvwg-archive@odin.ietf.org; Tue, 26 Feb 2002 07:17:06 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA08650;
	Tue, 26 Feb 2002 07:05:34 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA08625
	for <tsvwg@optimus.ietf.org>; Tue, 26 Feb 2002 07:05:32 -0500 (EST)
Received: from stewartlin.chicago.il.us (host-64-47-24-166.masergy.com [64.47.24.166] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA24748
	for <tsvwg@ietf.org>; Tue, 26 Feb 2002 07:05:29 -0500 (EST)
Received: from stewart.chicago.il.us (stewlap [10.1.1.5])
	by stewartlin.chicago.il.us (8.11.0/8.11.0) with ESMTP id g1QC6er31888;
	Tue, 26 Feb 2002 06:06:40 -0600
Message-ID: <3C7B7A03.3CFDF8B@stewart.chicago.il.us>
Date: Tue, 26 Feb 2002 06:05:23 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: "WENDT,JIM (HP-Roseville,ex1)" <jim_wendt@hp.com>
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP and IP Frag aliasing on LFNs ?
References: <499DC368E25AD411B3F100902740AD650AFC7DC9@xrose03.rose.hp.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Jim:

Some thoughts below...

"WENDT,JIM (HP-Roseville,ex1)" wrote:
> 
> Based upon following in RFC-2960:
> 
> > 6.9 Fragmentation and Reassembly
> >   Note: Once a message is fragmented it cannot be re-fragmented.
> >   Instead if the PMTU has been reduced, then IP fragmentation must be
> >   used.  Please see Section 7.3 for details of PMTU discovery.
> 
> > 7.3 Path MTU Discovery
> >   4) Since data transmission in SCTP is naturally structured in terms
> >      of TSNs rather than bytes (as is the case for TCP), the discussion
> >      in Section 6.5 of RFC 1191 applies: When retransmitting an IP
> >      datagram to a remote address for which the IP datagram appears too
> >      large for the path MTU to that address, the IP datagram SHOULD be
> >      retransmitted without the DF bit set, allowing it to possibly be
> >      fragmented.  Transmissions of new IP datagrams MUST have DF set.
> 
> Am I correct in understanding that IP fragmentation can occur
> on SCTP packets across a IPv4 PMTU degrade, and that worst case,
> IP fragmentation can occur for all packets in flight up to the
> roundtrip pipe size between the sending endpoint and the last
> hop router?
> 
I would think this can occur for any transport doing  PMTU discovery.
I do have  a hard time envisioning a very very highspeed inter-net
link that has a PMTU event without some sort of accompanying routing
re-convergence. And it seems to me if you hit a routing reconvergence
you are in for at LEAST 1 timeout putting you way back in slow start.

But lets do some rough calculations on a napkin anyway and see what
happens if a PMTU change occurs on say a 1 Gig-a-bit link with
lets say a 300ms round trip time. And just for fun we will assume
you are the ONLY user on this 1 Gig-a-bit link and you have fast
machines on both sides that can use it all.

Now lets say a 1500 byte MTU before the change and a 500 byte
MTU after the change. Now on a napkin, I come up with about 6.7
packets per ms... so we show up with 2,010 packets in flight.

Note I am rounding here a 1,000,000,000 bit -> 100,000,000 Mega Byte.
This may not be exact but it is quite close and allows for the
overhead in the e-net and anthing else :>

Now if I have only 2,010 packets all that need to be resent and with an
IP id of 16 bits or 65,535 I should be ok.. 

In fact I should be ok up to a RTT time of about 4.8 seconds.
-----

Now lets look at the same case with a 10 Gig a bit link.
We jump by a factor of 10 so... 

20,100 packets in flight .. no problem... 

Good up to 480ms round trip time..

So I guess there may be an issue if one has greater than a 480ms
round trip time and a 10 gigabit connection where you are using
the whole link all by yourself.

> If so, and the pipe is large, has there been any discussion or
> concern regarding aliasing or wrap-around of the 16-bit
> Identification (ID) field in the IP header?  Is this an
> issue, and if so, how is it being addressed?
> 

Considering that the scenario is quite far fetched in todays
technology over an inter-net. Maybe in a Intra-net.. but then
one would hope you would not get more than a 400ms RTT...

And besides all that .. in the future if this really becomes
a problem I think SCTP is exandable enough that we can figure
out a new chunk or some such to help with this type of situation.

I think we can worry about this problem when we get a bit closer
to having a realistic scenario where a problem would exist...

I just don't see it yet... 

Maybe I have missed something in my back-of-the napkin calculations
or I am mis-interpreting what circumstances you are envisioning. If
so please correct me..


R


-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 26 08:21:41 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA27921
	for <tsvwg-archive@odin.ietf.org>; Tue, 26 Feb 2002 08:21:41 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id IAA11957
	for tsvwg-archive@odin.ietf.org; Tue, 26 Feb 2002 08:21:44 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA10555;
	Tue, 26 Feb 2002 08:00:01 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id HAA10504
	for <tsvwg@optimus.ietf.org>; Tue, 26 Feb 2002 07:59:59 -0500 (EST)
Received: from hrtades9.atea.be (viap068.atea.be [194.78.143.68])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA27165;
	Tue, 26 Feb 2002 07:59:54 -0500 (EST)
Received: from hrtades10.atea.be (siemens.atea.be [139.10.143.141]) by hrtades9.atea.be with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13)
	id DF4YVDG6; Tue, 26 Feb 2002 13:59:28 +0100
Received: by siemens.atea.be with Internet Mail Service (5.5.2653.19)
	id <FNAN862V>; Tue, 26 Feb 2002 13:59:48 +0100
Message-ID: <E76F715C0429D5118F2100508BB9EDEE0795D1@hrtades7.atea.be>
From: Coene Lode <Lode.Coene@siemens.atea.be>
To: "SIGTRAN (E-mail)" <sigtran@ietf.org>,
        "Transport Area wg (E-mail)"
	 <tsvwg@ietf.org>
Date: Tue, 26 Feb 2002 13:56:36 +0100
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: [Tsvwg] Multihoming issues in SCTP
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Folks,

The latest iteration of the Multihoming Issues draft with SCTP is out.
Main improvements were(?):
- "downloading routing tables" was replaced with "taking part in the routing
information distribution" 
- a few editorial comments.

Thanks for all the comments.
If you would have further comments, let me know.

I will ask the TSV AD's to publish this as a informational RFC.

A New Internet-Draft is available from the on-line Internet-Drafts
directories.


	Title		: Multihoming issues in the Stream Control
Transmission 
                          Protocol
	Author(s)	: L. Coene
	Filename	: draft-coene-sctp-multihome-03.txt
	Pages		: 11
	Date		: 20-Feb-02
	
This document describes issues of the Stream Control Transmission
Protocol (SCTP)[RFC2960] in regard to multihoming on the
Internet. It explores cases where through situations in the
internet, single points-of-failure can occur even when using
multihoming and what the impact is of multihoming on the routing
tables of the host.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-coene-sctp-multihome-03.txt



yours sincerly,
Lode Coene

Siemens atea
atealaan 34          2200 Herentals, Belgium
E-mail: lode.coene@siemens.atea.be
Tel: +32-14-252081
Fax: +32-14-253212

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 26 11:03:15 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA05067
	for <tsvwg-archive@odin.ietf.org>; Tue, 26 Feb 2002 11:03:15 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id LAA19968
	for tsvwg-archive@odin.ietf.org; Tue, 26 Feb 2002 11:03:18 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA18935;
	Tue, 26 Feb 2002 10:42:37 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id KAA18905
	for <tsvwg@optimus.ietf.org>; Tue, 26 Feb 2002 10:42:34 -0500 (EST)
Received: from albatross.wise.edt.ericsson.se (albatross-ext.wise.edt.ericsson.se [193.180.251.46])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA04187
	for <tsvwg@ietf.org>; Tue, 26 Feb 2002 10:42:30 -0500 (EST)
Received: from eed.ericsson.se (mailhost.eed.ericsson.se [164.48.133.33])
	by albatross.wise.edt.ericsson.se (8.12.1/8.12.1/WIREfire-1.4) with ESMTP id g1QFgXZc001752
	for <tsvwg@ietf.org>; Tue, 26 Feb 2002 16:42:33 +0100 (MET)
Received: from res0010384da36.ericsson.com (dhcp5-236 [164.48.135.236])
	by eed.ericsson.se (8.8.8+Sun/1.1.mit) with ESMTP id QAA09870
	for <tsvwg@ietf.org>; Tue, 26 Feb 2002 16:42:32 +0100 (MET)
Message-Id: <5.1.0.14.0.20020226162306.00b1c8f8@mailhost.eed.ericsson.se>
X-Sender: eedrel@mailhost.eed.ericsson.se
X-Mailer: QUALCOMM Windows Eudora Version 5.1
Date: Tue, 26 Feb 2002 16:43:39 +0100
To: tsvwg@ietf.org
From: Reiner Ludwig <Reiner.Ludwig@ericsson.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Subject: [Tsvwg] Eifel Detection Algorithm
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi,

we finalized the Eifel detection algorithm for TCP:
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-tcp-eifel-alg-03.txt

We consider the document pretty much done, and would welcome review and 
feedback from the TSV working group.

Main changes since the last revision:

- No more failure cases:
There was recently a discussion on the mailing list about failure cases of 
the Eifel detection algorithm. This had to do with how the TCP receiver 
implements the TCP Timestamps option defined in RFC1323. This has been 
addressed in the current ID, and we now believe that the algorithm is 
reliable and robust.

- Protections against lying TCP receivers:
We outline an optional restricted version of the algorithm that would even 
work in case the TCP receiver tried forging timestamps to fool the TCP 
sender into believing a genuine packet loss was spurious.

- No more SACK support:
The 02 version defined a (less capable) alternative of the algorithm based 
on the TCP SACK option. That part has been taken out.

///Reiner


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 26 16:09:07 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA21751
	for <tsvwg-archive@odin.ietf.org>; Tue, 26 Feb 2002 16:09:07 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA20414
	for tsvwg-archive@odin.ietf.org; Tue, 26 Feb 2002 16:09:11 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA18989;
	Tue, 26 Feb 2002 15:51:26 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA18962
	for <tsvwg@optimus.ietf.org>; Tue, 26 Feb 2002 15:51:23 -0500 (EST)
Received: from c003.snv.cp.net (c003-h000.c003.snv.cp.net [209.228.32.214])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id PAA21138
	for <tsvwg@ietf.org>; Tue, 26 Feb 2002 15:51:18 -0500 (EST)
Received: (cpmta 17059 invoked from network); 26 Feb 2002 12:50:50 -0800
Received: from 66.123.21.234 (HELO littlejoy)
  by smtp.telocity.com (209.228.32.214) with SMTP; 26 Feb 2002 12:50:50 -0800
X-Sent: 26 Feb 2002 20:50:50 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: "Tsvwg" <tsvwg@ietf.org>, "Rdma" <rdma@yahoogroups.com>
Date: Tue, 26 Feb 2002 12:51:00 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJMEILCOAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0)
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Content-Transfer-Encoding: 7bit
Subject: [Tsvwg] SCTP DDP Shim update
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit


Updates to "SCTP DDP/RDMA Adaptation"
draft-stewart-otis-sctp-ddp-rdma-01.txt
have been published and placed at:
http://groups.yahoo.com/group/SCTP-DDP/files/

Flag names were changed to prevent conflicts with RDMA conventions.
Non-Placement mode was renamed to ANONYMOUS to accurately reflect use.
Payload Protocol Identifier with IANA registration was included to allow
higher-level protocol extensions without need to modify shim definitions.
RDMA or other higher level protocol extensions can include PPI as a means to
negotiate related features or limitations.  The use of AWS was removed as
being redundant with the use of a common Stream as unordered delivery does
not suffer from Head of Queue issues.

A SCTP-DDP specific problem statement that did not meet the -00 deadline is
also included and the link should be functional for non-members.

Doug


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 26 16:11:44 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA21827
	for <tsvwg-archive@odin.ietf.org>; Tue, 26 Feb 2002 16:11:44 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA20935
	for tsvwg-archive@odin.ietf.org; Tue, 26 Feb 2002 16:11:47 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA19372;
	Tue, 26 Feb 2002 15:57:19 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA19341
	for <tsvwg@optimus.ietf.org>; Tue, 26 Feb 2002 15:57:17 -0500 (EST)
Received: from ztxmail05.ztx.compaq.com (ztxmail05.ztx.compaq.com [161.114.1.209])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA21336
	for <tsvwg@ietf.org>; Tue, 26 Feb 2002 15:57:12 -0500 (EST)
Received: from cceexg12.americas.cpqcorp.net (cceexg12.americas.cpqcorp.net [16.110.250.124])
	by ztxmail05.ztx.compaq.com (Postfix) with ESMTP
	id DE7D9438B; Tue, 26 Feb 2002 14:56:45 -0600 (CST)
Received: from cceexc16.americas.cpqcorp.net ([16.110.250.58]) by cceexg12.americas.cpqcorp.net with Microsoft SMTPSVC(5.0.2195.2966);
	 Tue, 26 Feb 2002 14:56:41 -0600
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C1BF08.174B4A11"
Date: Tue, 26 Feb 2002 14:56:41 -0600
Message-ID: <07DEBFB3DCB18A49B06B637199358674F83F6F@cceexc16.americas.cpqcorp.net>
Thread-Topic: [rdma] SCTP DDP Shim update
Thread-Index: AcG/B00ORM18W4R8QA2Ac7A70yQoggAAHaBA
From: "Hilland, Jeff (Sys Hou, HW Tech)" <Jeff.Hilland@Compaq.com>
To: "Douglas Otis" <dotis@sanlight.net>
Cc: "Tsvwg" <tsvwg@ietf.org>, "Rdma" <rdma@yahoogroups.com>
X-OriginalArrivalTime: 26 Feb 2002 20:56:41.0941 (UTC) FILETIME=[1795A050:01C1BF08]
Subject: [Tsvwg] RE: [rdma] SCTP DDP Shim update
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

This is a multi-part message in MIME format.

------_=_NextPart_001_01C1BF08.174B4A11
Content-Type: text/plain;
	charset="US-ASCII"
Content-Transfer-Encoding: quoted-printable

Could you please post these files to where they are publicly available?
The only way to get these files is to subscribe to the SCTP-DDP yahoo
group and I have no desire to do so. (I'm already getting too much
spam.)  Could you please at least make them available on the RDMA
reflector (to which this message was posted).
=20
Thanks,
Jeff Hilland

	-----Original Message-----
	From: Douglas Otis [mailto:dotis@sanlight.net]=20
	Sent: Tuesday, February 26, 2002 2:51 PM
	To: Tsvwg; Rdma
	Subject: [rdma] SCTP DDP Shim update
=09
=09
=09
	Updates to "SCTP DDP/RDMA Adaptation"
	draft-stewart-otis-sctp-ddp-rdma-01.txt
	have been published and placed at:
	http://groups.yahoo.com/group/SCTP-DDP/files/
=09
	Flag names were changed to prevent conflicts with RDMA
conventions.
	Non-Placement mode was renamed to ANONYMOUS to accurately
reflect use.
	Payload Protocol Identifier with IANA registration was included
to allow
	higher-level protocol extensions without need to modify shim
definitions.
	RDMA or other higher level protocol extensions can include PPI
as a means to
	negotiate related features or limitations.  The use of AWS was
removed as
	being redundant with the use of a common Stream as unordered
delivery does
	not suffer from Head of Queue issues.
=09
	A SCTP-DDP specific problem statement that did not meet the -00
deadline is
	also included and the link should be functional for non-members.
=09
	Doug
=09
=09
	To unsubscribe from this group, send an email to:
	rdma-unsubscribe@egroups.com
=09
=09
=09
	Your use of Yahoo! Groups is subject to the Yahoo! Terms of
Service <http://docs.yahoo.com/info/terms/> .=20
=09


------_=_NextPart_001_01C1BF08.174B4A11
Content-Type: text/html;
	charset="US-ASCII"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD><TITLE>Message</TITLE>
<META http-equiv=3DContent-Type content=3D"text/html; =
charset=3Dus-ascii">
<META content=3D"MSHTML 6.00.2713.1100" name=3DGENERATOR></HEAD>
<BODY>
<DIV><FONT face=3DArial color=3D#0000ff size=3D2><SPAN =
class=3D013215420-26022002>Could=20
you please post these files to where they are publicly available?&nbsp; =
The only=20
way to get these files is to subscribe to the SCTP-DDP yahoo group and I =
have no=20
desire to do so. (I'm already getting too much spam.)&nbsp; Could you =
please at=20
least make them available on the RDMA reflector (to which this message =
was=20
posted).</SPAN></FONT></DIV>
<DIV><FONT face=3DArial color=3D#0000ff size=3D2><SPAN=20
class=3D013215420-26022002></SPAN></FONT><FONT face=3DArial =
color=3D#0000ff=20
size=3D2><SPAN class=3D013215420-26022002></SPAN></FONT>&nbsp;</DIV>
<DIV><FONT face=3DArial color=3D#0000ff size=3D2><SPAN=20
class=3D013215420-26022002>Thanks,</SPAN></FONT></DIV>
<DIV><FONT face=3DArial color=3D#0000ff size=3D2><SPAN =
class=3D013215420-26022002>Jeff=20
Hilland</SPAN></FONT></DIV>
<BLOCKQUOTE style=3D"MARGIN-RIGHT: 0px">
  <DIV></DIV>
  <DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr =
align=3Dleft><FONT=20
  face=3DTahoma size=3D2>-----Original Message-----<BR><B>From:</B> =
Douglas Otis=20
  [mailto:dotis@sanlight.net] <BR><B>Sent:</B> Tuesday, February 26, =
2002 2:51=20
  PM<BR><B>To:</B> Tsvwg; Rdma<BR><B>Subject:</B> [rdma] SCTP DDP Shim=20
  update<BR><BR></FONT></DIV><TT><BR>Updates to "SCTP DDP/RDMA=20
  Adaptation"<BR>draft-stewart-otis-sctp-ddp-rdma-01.txt<BR>have been =
published=20
  and placed at:<BR><A=20
  =
href=3D"http://groups.yahoo.com/group/SCTP-DDP/files/">http://groups.yaho=
o.com/group/SCTP-DDP/files/</A><BR><BR>Flag=20
  names were changed to prevent conflicts with RDMA=20
  conventions.<BR>Non-Placement mode was renamed to ANONYMOUS to =
accurately=20
  reflect use.<BR>Payload Protocol Identifier with IANA registration was =

  included to allow<BR>higher-level protocol extensions without need to =
modify=20
  shim definitions.<BR>RDMA or other higher level protocol extensions =
can=20
  include PPI as a means to<BR>negotiate related features or =
limitations.&nbsp;=20
  The use of AWS was removed as<BR>being redundant with the use of a =
common=20
  Stream as unordered delivery does<BR>not suffer from Head of Queue=20
  issues.<BR><BR>A SCTP-DDP specific problem statement that did not meet =
the -00=20
  deadline is<BR>also included and the link should be functional for=20
  non-members.<BR><BR>Doug<BR><BR></TT><BR><TT>To unsubscribe from this =
group,=20
  send an email =
to:<BR>rdma-unsubscribe@egroups.com<BR><BR></TT><BR><BR><TT>Your=20
  use of Yahoo! Groups is subject to the <A=20
  href=3D"http://docs.yahoo.com/info/terms/">Yahoo! Terms of =
Service</A>.</TT>=20
<BR></BLOCKQUOTE></BODY></HTML>
=00
------_=_NextPart_001_01C1BF08.174B4A11--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 26 16:11:49 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA21842
	for <tsvwg-archive@odin.ietf.org>; Tue, 26 Feb 2002 16:11:44 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA20950
	for tsvwg-archive@odin.ietf.org; Tue, 26 Feb 2002 16:11:48 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA19434;
	Tue, 26 Feb 2002 15:58:22 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA19402
	for <tsvwg@optimus.ietf.org>; Tue, 26 Feb 2002 15:58:20 -0500 (EST)
Received: from palrel13.hp.com (palrel13.hp.com [156.153.255.238])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA21397
	for <tsvwg@ietf.org>; Tue, 26 Feb 2002 15:58:12 -0500 (EST)
Received: from xparelay2.corp.hp.com (xparelay2.corp.hp.com [15.58.137.112])
	by palrel13.hp.com (Postfix) with ESMTP
	id 34DFE400624; Tue, 26 Feb 2002 12:57:45 -0800 (PST)
Received: from xpabh4.corp.hp.com (xpabh4.corp.hp.com [15.58.136.1])
	by xparelay2.corp.hp.com (Postfix) with ESMTP
	id E721ABB; Tue, 26 Feb 2002 12:57:39 -0800 (PST)
Received: by xpabh4.corp.hp.com with Internet Mail Service (5.5.2653.19)
	id <FQ4AT9CS>; Tue, 26 Feb 2002 12:57:39 -0800
Message-ID: <499DC368E25AD411B3F100902740AD650AFC7DCC@xrose03.rose.hp.com>
From: "WENDT,JIM (HP-Roseville,ex1)" <jim_wendt@hp.com>
To: "'Randall Stewart'" <randall@stewart.chicago.il.us>
Cc: tsvwg@ietf.org
Subject: RE: [Tsvwg] SCTP and IP Frag aliasing on LFNs ?
Date: Tue, 26 Feb 2002 12:57:38 -0800
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Randy,
Hi. I do come up with different numbers, and they are
very different from yours so please double check
me on them. Sorry I didn't include them initially.

Using your parameters:
  link-rate = 1 Giga-bit-per-sec
            = 100,000,000 bytes-per-sec (approx)
  RTT = 300ms
  MTU = 500 bytes

I come up with:
  bytes-per-ms = bytes-per-sec / 1000
               = 100,000,000 / 1000
               = 100,000 bytes-per-ms

  packets-per-ms = bytes-per-ms / MTU
                 = 100,000 / 500
                 = 200 packets-per-ms

  packets-in-pipe = packet-per-ms * RTT(ms)
                  = 200 * 300
                  = 60,000 packets in pipe
                    IP-ID doesn't wrap in the pipe

And at 10 Gbps (10x):
  600,000 packets in pipe
  IP-ID wraps 9.15 times in the pipe

And at 100 Gbps (100x):
  6,000,000 packets in pipe
  IP-ID wraps 91.5 times in the pipe

And yes it would be the same issue for any transport
that relies on IP (v4) fragmentation.

So I stil have questions about this issue in general.
Is it going to be an actual problem in the
forseeable future (even for smaller RTTs in the
50ms range), or are there other mitigating factors
that make it a non-issue?

Thx again,
Jim

P.S. Another way to calc this is:
  pipe-distance-in-sec = 300ms = 0.3 sec

  bytes-in-pipe = bytes-per-sec * pipe-distance-in-sec
                = 100,000,000 * 0.3
                = 30,000,000 bytes-in-pipe

  packets-in-pipe = bytes-in-pipe / MTU
                  = 30,000,000 / 500
                  = 60,000 packets-in-pipe (on 1Gbps link)

---

> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Tuesday, February 26, 2002 4:05 AM
> To: WENDT,JIM (HP-Roseville,ex1)
> Cc: tsvwg@ietf.org
> Subject: Re: [Tsvwg] SCTP and IP Frag aliasing on LFNs ?
> 
> 
> Jim:
> 
> Some thoughts below...
> 
> "WENDT,JIM (HP-Roseville,ex1)" wrote:
> > 
> > Based upon following in RFC-2960:
> > 
> > > 6.9 Fragmentation and Reassembly
> > >   Note: Once a message is fragmented it cannot be re-fragmented.
> > >   Instead if the PMTU has been reduced, then IP 
> fragmentation must be
> > >   used.  Please see Section 7.3 for details of PMTU discovery.
> > 
> > > 7.3 Path MTU Discovery
> > >   4) Since data transmission in SCTP is naturally 
> structured in terms
> > >      of TSNs rather than bytes (as is the case for TCP), 
> the discussion
> > >      in Section 6.5 of RFC 1191 applies: When retransmitting an IP
> > >      datagram to a remote address for which the IP 
> datagram appears too
> > >      large for the path MTU to that address, the IP 
> datagram SHOULD be
> > >      retransmitted without the DF bit set, allowing it to 
> possibly be
> > >      fragmented.  Transmissions of new IP datagrams MUST 
> have DF set.
> > 
> > Am I correct in understanding that IP fragmentation can occur
> > on SCTP packets across a IPv4 PMTU degrade, and that worst case,
> > IP fragmentation can occur for all packets in flight up to the
> > roundtrip pipe size between the sending endpoint and the last
> > hop router?
> > 
> I would think this can occur for any transport doing  PMTU discovery.
> I do have  a hard time envisioning a very very highspeed inter-net
> link that has a PMTU event without some sort of accompanying routing
> re-convergence. And it seems to me if you hit a routing reconvergence
> you are in for at LEAST 1 timeout putting you way back in slow start.
> 
> But lets do some rough calculations on a napkin anyway and see what
> happens if a PMTU change occurs on say a 1 Gig-a-bit link with
> lets say a 300ms round trip time. And just for fun we will assume
> you are the ONLY user on this 1 Gig-a-bit link and you have fast
> machines on both sides that can use it all.
> 
> Now lets say a 1500 byte MTU before the change and a 500 byte
> MTU after the change. Now on a napkin, I come up with about 6.7
> packets per ms... so we show up with 2,010 packets in flight.
> 
> Note I am rounding here a 1,000,000,000 bit -> 100,000,000 Mega Byte.
> This may not be exact but it is quite close and allows for the
> overhead in the e-net and anthing else :>
> 
> Now if I have only 2,010 packets all that need to be resent 
> and with an
> IP id of 16 bits or 65,535 I should be ok.. 
> 
> In fact I should be ok up to a RTT time of about 4.8 seconds.
> -----
> 
> Now lets look at the same case with a 10 Gig a bit link.
> We jump by a factor of 10 so... 
> 
> 20,100 packets in flight .. no problem... 
> 
> Good up to 480ms round trip time..
> 
> So I guess there may be an issue if one has greater than a 480ms
> round trip time and a 10 gigabit connection where you are using
> the whole link all by yourself.
> 
> > If so, and the pipe is large, has there been any discussion or
> > concern regarding aliasing or wrap-around of the 16-bit
> > Identification (ID) field in the IP header?  Is this an
> > issue, and if so, how is it being addressed?
> > 
> 
> Considering that the scenario is quite far fetched in todays
> technology over an inter-net. Maybe in a Intra-net.. but then
> one would hope you would not get more than a 400ms RTT...
> 
> And besides all that .. in the future if this really becomes
> a problem I think SCTP is exandable enough that we can figure
> out a new chunk or some such to help with this type of situation.
> 
> I think we can worry about this problem when we get a bit closer
> to having a realistic scenario where a problem would exist...
> 
> I just don't see it yet... 
> 
> Maybe I have missed something in my back-of-the napkin calculations
> or I am mis-interpreting what circumstances you are envisioning. If
> so please correct me..
> 
> 
> R
> 
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 26 16:37:20 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA22461
	for <tsvwg-archive@odin.ietf.org>; Tue, 26 Feb 2002 16:37:19 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id QAA22594
	for tsvwg-archive@odin.ietf.org; Tue, 26 Feb 2002 16:37:23 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA21460;
	Tue, 26 Feb 2002 16:21:28 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id QAA21424
	for <tsvwg@optimus.ietf.org>; Tue, 26 Feb 2002 16:21:26 -0500 (EST)
Received: from c003.snv.cp.net (c003-h013.c003.snv.cp.net [209.228.32.227])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id QAA22041
	for <tsvwg@ietf.org>; Tue, 26 Feb 2002 16:21:17 -0500 (EST)
Received: (cpmta 8834 invoked from network); 26 Feb 2002 13:20:50 -0800
Received: from 66.123.21.234 (HELO littlejoy)
  by smtp.telocity.com (209.228.32.227) with SMTP; 26 Feb 2002 13:20:50 -0800
X-Sent: 26 Feb 2002 21:20:50 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: "Hilland, Jeff \(Sys Hou, HW Tech\)" <Jeff.Hilland@Compaq.com>
Cc: "Tsvwg" <tsvwg@ietf.org>, "Rdma" <rdma@yahoogroups.com>
Subject: RE: [Tsvwg] RE: [rdma] SCTP DDP Shim update
Date: Tue, 26 Feb 2002 13:20:59 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJIEIOCOAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0)
Importance: Normal
In-Reply-To: <07DEBFB3DCB18A49B06B637199358674F83F6F@cceexc16.americas.cpqcorp.net>
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Jeff,

Ignore the join invitation and scroll down to see the file links.  You will
see this file post to the IETF shortly.

Doug


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 26 20:05:13 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA26814
	for <tsvwg-archive@odin.ietf.org>; Tue, 26 Feb 2002 20:05:12 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA03352
	for tsvwg-archive@odin.ietf.org; Tue, 26 Feb 2002 20:05:15 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA02155;
	Tue, 26 Feb 2002 19:45:33 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA02124
	for <tsvwg@optimus.ietf.org>; Tue, 26 Feb 2002 19:45:31 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA26557
	for <tsvwg@ietf.org>; Tue, 26 Feb 2002 19:45:28 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir04nok.ntc.nokia.com (esvir04nokt.ntc.nokia.com [172.21.143.36])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1R0jdZ07148
	for <tsvwg@ietf.org>; Wed, 27 Feb 2002 02:45:39 +0200 (EET)
Received: from esebh004.NOE.Nokia.com (unverified) by esvir04nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T5951e32749ac158f24077@esvir04nok.ntc.nokia.com>;
 Wed, 27 Feb 2002 02:45:30 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh004.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Wed, 27 Feb 2002 02:45:29 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] SCTP and IP Frag aliasing on LFNs ?
Date: Wed, 27 Feb 2002 02:45:29 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF26308EEE4@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] SCTP and IP Frag aliasing on LFNs ?
Thread-Index: AcG/CGDF6H6ZERHgQPGa5YbCsugLcQAHrwcA
To: <jim_wendt@hp.com>, <randall@stewart.chicago.il.us>
Cc: <tsvwg@ietf.org>
X-OriginalArrivalTime: 27 Feb 2002 00:45:29.0897 (UTC) FILETIME=[0E157990:01C1BF28]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id TAA02125
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Just a comment...

	Remember that if you are able of having 60,000 packets of 500 bytes on flight, it means that the receiver must have a buffer of at least 30,000,000... It is not an incredibly big buffer but again, it makes things a little bit more difficult... In the 100 Gbps network it would be a buffer of 3 Gbytes...

	Moreover, I think that before all these big numbers become a reality we will have ICMPv6 that includes also the TSNs of the DATA chunks inside the SCTP datagram...

	BR Iván Arias Rodríguez

-----Original Message-----
From: ext WENDT,JIM (HP-Roseville,ex1) [mailto:jim_wendt@hp.com]
Sent: 26 February 2002 22:58
To: 'Randall Stewart'
Cc: tsvwg@ietf.org
Subject: RE: [Tsvwg] SCTP and IP Frag aliasing on LFNs ?


Randy,
Hi. I do come up with different numbers, and they are
very different from yours so please double check
me on them. Sorry I didn't include them initially.

Using your parameters:
  link-rate = 1 Giga-bit-per-sec
            = 100,000,000 bytes-per-sec (approx)
  RTT = 300ms
  MTU = 500 bytes

I come up with:
  bytes-per-ms = bytes-per-sec / 1000
               = 100,000,000 / 1000
               = 100,000 bytes-per-ms

  packets-per-ms = bytes-per-ms / MTU
                 = 100,000 / 500
                 = 200 packets-per-ms

  packets-in-pipe = packet-per-ms * RTT(ms)
                  = 200 * 300
                  = 60,000 packets in pipe
                    IP-ID doesn't wrap in the pipe

And at 10 Gbps (10x):
  600,000 packets in pipe
  IP-ID wraps 9.15 times in the pipe

And at 100 Gbps (100x):
  6,000,000 packets in pipe
  IP-ID wraps 91.5 times in the pipe

And yes it would be the same issue for any transport
that relies on IP (v4) fragmentation.

So I stil have questions about this issue in general.
Is it going to be an actual problem in the
forseeable future (even for smaller RTTs in the
50ms range), or are there other mitigating factors
that make it a non-issue?

Thx again,
Jim

P.S. Another way to calc this is:
  pipe-distance-in-sec = 300ms = 0.3 sec

  bytes-in-pipe = bytes-per-sec * pipe-distance-in-sec
                = 100,000,000 * 0.3
                = 30,000,000 bytes-in-pipe

  packets-in-pipe = bytes-in-pipe / MTU
                  = 30,000,000 / 500
                  = 60,000 packets-in-pipe (on 1Gbps link)

---

> -----Original Message-----
> From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> Sent: Tuesday, February 26, 2002 4:05 AM
> To: WENDT,JIM (HP-Roseville,ex1)
> Cc: tsvwg@ietf.org
> Subject: Re: [Tsvwg] SCTP and IP Frag aliasing on LFNs ?
> 
> 
> Jim:
> 
> Some thoughts below...
> 
> "WENDT,JIM (HP-Roseville,ex1)" wrote:
> > 
> > Based upon following in RFC-2960:
> > 
> > > 6.9 Fragmentation and Reassembly
> > >   Note: Once a message is fragmented it cannot be re-fragmented.
> > >   Instead if the PMTU has been reduced, then IP 
> fragmentation must be
> > >   used.  Please see Section 7.3 for details of PMTU discovery.
> > 
> > > 7.3 Path MTU Discovery
> > >   4) Since data transmission in SCTP is naturally 
> structured in terms
> > >      of TSNs rather than bytes (as is the case for TCP), 
> the discussion
> > >      in Section 6.5 of RFC 1191 applies: When retransmitting an IP
> > >      datagram to a remote address for which the IP 
> datagram appears too
> > >      large for the path MTU to that address, the IP 
> datagram SHOULD be
> > >      retransmitted without the DF bit set, allowing it to 
> possibly be
> > >      fragmented.  Transmissions of new IP datagrams MUST 
> have DF set.
> > 
> > Am I correct in understanding that IP fragmentation can occur
> > on SCTP packets across a IPv4 PMTU degrade, and that worst case,
> > IP fragmentation can occur for all packets in flight up to the
> > roundtrip pipe size between the sending endpoint and the last
> > hop router?
> > 
> I would think this can occur for any transport doing  PMTU discovery.
> I do have  a hard time envisioning a very very highspeed inter-net
> link that has a PMTU event without some sort of accompanying routing
> re-convergence. And it seems to me if you hit a routing reconvergence
> you are in for at LEAST 1 timeout putting you way back in slow start.
> 
> But lets do some rough calculations on a napkin anyway and see what
> happens if a PMTU change occurs on say a 1 Gig-a-bit link with
> lets say a 300ms round trip time. And just for fun we will assume
> you are the ONLY user on this 1 Gig-a-bit link and you have fast
> machines on both sides that can use it all.
> 
> Now lets say a 1500 byte MTU before the change and a 500 byte
> MTU after the change. Now on a napkin, I come up with about 6.7
> packets per ms... so we show up with 2,010 packets in flight.
> 
> Note I am rounding here a 1,000,000,000 bit -> 100,000,000 Mega Byte.
> This may not be exact but it is quite close and allows for the
> overhead in the e-net and anthing else :>
> 
> Now if I have only 2,010 packets all that need to be resent 
> and with an
> IP id of 16 bits or 65,535 I should be ok.. 
> 
> In fact I should be ok up to a RTT time of about 4.8 seconds.
> -----
> 
> Now lets look at the same case with a 10 Gig a bit link.
> We jump by a factor of 10 so... 
> 
> 20,100 packets in flight .. no problem... 
> 
> Good up to 480ms round trip time..
> 
> So I guess there may be an issue if one has greater than a 480ms
> round trip time and a 10 gigabit connection where you are using
> the whole link all by yourself.
> 
> > If so, and the pipe is large, has there been any discussion or
> > concern regarding aliasing or wrap-around of the 16-bit
> > Identification (ID) field in the IP header?  Is this an
> > issue, and if so, how is it being addressed?
> > 
> 
> Considering that the scenario is quite far fetched in todays
> technology over an inter-net. Maybe in a Intra-net.. but then
> one would hope you would not get more than a 400ms RTT...
> 
> And besides all that .. in the future if this really becomes
> a problem I think SCTP is exandable enough that we can figure
> out a new chunk or some such to help with this type of situation.
> 
> I think we can worry about this problem when we get a bit closer
> to having a realistic scenario where a problem would exist...
> 
> I just don't see it yet... 
> 
> Maybe I have missed something in my back-of-the napkin calculations
> or I am mis-interpreting what circumstances you are envisioning. If
> so please correct me..
> 
> 
> R
> 
> 
> -- 
> Randall R. Stewart
> randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Tue Feb 26 20:05:44 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA26834
	for <tsvwg-archive@odin.ietf.org>; Tue, 26 Feb 2002 20:05:43 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id UAA03384
	for tsvwg-archive@odin.ietf.org; Tue, 26 Feb 2002 20:05:45 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA02095;
	Tue, 26 Feb 2002 19:44:25 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA02065
	for <tsvwg@optimus.ietf.org>; Tue, 26 Feb 2002 19:44:22 -0500 (EST)
Received: from bfgbhome.inetint.com (IDENT:root@tnt-dal-42-200.dallas.net [209.44.42.200])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA26539
	for <tsvwg@ietf.org>; Tue, 26 Feb 2002 19:44:18 -0500 (EST)
Received: (from brian@localhost)
	by bfgbhome.inetint.com (8.9.3/8.9.3) id SAA00776;
	Tue, 26 Feb 2002 18:44:03 -0600
Date: Tue, 26 Feb 2002 18:44:03 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: "WENDT,JIM (HP-Roseville,ex1)" <jim_wendt@hp.com>
Cc: "'Randall Stewart'" <randall@stewart.chicago.il.us>, tsvwg@ietf.org
Subject: Re: [Tsvwg] SCTP and IP Frag aliasing on LFNs ?
Message-ID: <20020226184403.B32648@openss7.org>
Reply-To: bidulock@openss7.org
References: <499DC368E25AD411B3F100902740AD650AFC7DCC@xrose03.rose.hp.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.2i
In-Reply-To: <499DC368E25AD411B3F100902740AD650AFC7DCC@xrose03.rose.hp.com>; from jim_wendt@hp.com on Tue, Feb 26, 2002 at 12:57:38PM -0800
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
Content-Transfer-Encoding: 8bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

JIM,

The RFC requires an implementation to chunk down DATA chunks
to fit into a path MTU, but AFAIK there is no requirement
stopping an implementation from chunking down DATA chunks to
less than a path MTU and then bundling them back up.  Our
implementation, for example, splits DATA chunks to 1/2 PMTU or
less.  This mitigates falling MTUs in that they must fall
faster (and smaller) to cause fragmentation on retransmission.

--brian

On Tue, 26 Feb 2002, WENDT,JIM (HP-Roseville,ex1) wrote:

> Randy,
> Hi. I do come up with different numbers, and they are
> very different from yours so please double check
> me on them. Sorry I didn't include them initially.
> 
> Using your parameters:
>   link-rate = 1 Giga-bit-per-sec
>             = 100,000,000 bytes-per-sec (approx)
>   RTT = 300ms
>   MTU = 500 bytes
> 
> I come up with:
>   bytes-per-ms = bytes-per-sec / 1000
>                = 100,000,000 / 1000
>                = 100,000 bytes-per-ms
> 
>   packets-per-ms = bytes-per-ms / MTU
>                  = 100,000 / 500
>                  = 200 packets-per-ms
> 
>   packets-in-pipe = packet-per-ms * RTT(ms)
>                   = 200 * 300
>                   = 60,000 packets in pipe
>                     IP-ID doesn't wrap in the pipe
> 
> And at 10 Gbps (10x):
>   600,000 packets in pipe
>   IP-ID wraps 9.15 times in the pipe
> 
> And at 100 Gbps (100x):
>   6,000,000 packets in pipe
>   IP-ID wraps 91.5 times in the pipe
> 
> And yes it would be the same issue for any transport
> that relies on IP (v4) fragmentation.
> 
> So I stil have questions about this issue in general.
> Is it going to be an actual problem in the
> forseeable future (even for smaller RTTs in the
> 50ms range), or are there other mitigating factors
> that make it a non-issue?
> 
> Thx again,
> Jim
> 
> P.S. Another way to calc this is:
>   pipe-distance-in-sec = 300ms = 0.3 sec
> 
>   bytes-in-pipe = bytes-per-sec * pipe-distance-in-sec
>                 = 100,000,000 * 0.3
>                 = 30,000,000 bytes-in-pipe
> 
>   packets-in-pipe = bytes-in-pipe / MTU
>                   = 30,000,000 / 500
>                   = 60,000 packets-in-pipe (on 1Gbps link)
> 
> ---
> 
> > -----Original Message-----
> > From: Randall Stewart [mailto:randall@stewart.chicago.il.us]
> > Sent: Tuesday, February 26, 2002 4:05 AM
> > To: WENDT,JIM (HP-Roseville,ex1)
> > Cc: tsvwg@ietf.org
> > Subject: Re: [Tsvwg] SCTP and IP Frag aliasing on LFNs ?
> > 
> > 
> > Jim:
> > 
> > Some thoughts below...
> > 
> > "WENDT,JIM (HP-Roseville,ex1)" wrote:
> > > 
> > > Based upon following in RFC-2960:
> > > 
> > > > 6.9 Fragmentation and Reassembly
> > > >   Note: Once a message is fragmented it cannot be re-fragmented.
> > > >   Instead if the PMTU has been reduced, then IP 
> > fragmentation must be
> > > >   used.  Please see Section 7.3 for details of PMTU discovery.
> > > 
> > > > 7.3 Path MTU Discovery
> > > >   4) Since data transmission in SCTP is naturally 
> > structured in terms
> > > >      of TSNs rather than bytes (as is the case for TCP), 
> > the discussion
> > > >      in Section 6.5 of RFC 1191 applies: When retransmitting an IP
> > > >      datagram to a remote address for which the IP 
> > datagram appears too
> > > >      large for the path MTU to that address, the IP 
> > datagram SHOULD be
> > > >      retransmitted without the DF bit set, allowing it to 
> > possibly be
> > > >      fragmented.  Transmissions of new IP datagrams MUST 
> > have DF set.
> > > 
> > > Am I correct in understanding that IP fragmentation can occur
> > > on SCTP packets across a IPv4 PMTU degrade, and that worst case,
> > > IP fragmentation can occur for all packets in flight up to the
> > > roundtrip pipe size between the sending endpoint and the last
> > > hop router?
> > > 
> > I would think this can occur for any transport doing  PMTU discovery.
> > I do have  a hard time envisioning a very very highspeed inter-net
> > link that has a PMTU event without some sort of accompanying routing
> > re-convergence. And it seems to me if you hit a routing reconvergence
> > you are in for at LEAST 1 timeout putting you way back in slow start.
> > 
> > But lets do some rough calculations on a napkin anyway and see what
> > happens if a PMTU change occurs on say a 1 Gig-a-bit link with
> > lets say a 300ms round trip time. And just for fun we will assume
> > you are the ONLY user on this 1 Gig-a-bit link and you have fast
> > machines on both sides that can use it all.
> > 
> > Now lets say a 1500 byte MTU before the change and a 500 byte
> > MTU after the change. Now on a napkin, I come up with about 6.7
> > packets per ms... so we show up with 2,010 packets in flight.
> > 
> > Note I am rounding here a 1,000,000,000 bit -> 100,000,000 Mega Byte.
> > This may not be exact but it is quite close and allows for the
> > overhead in the e-net and anthing else :>
> > 
> > Now if I have only 2,010 packets all that need to be resent 
> > and with an
> > IP id of 16 bits or 65,535 I should be ok.. 
> > 
> > In fact I should be ok up to a RTT time of about 4.8 seconds.
> > -----
> > 
> > Now lets look at the same case with a 10 Gig a bit link.
> > We jump by a factor of 10 so... 
> > 
> > 20,100 packets in flight .. no problem... 
> > 
> > Good up to 480ms round trip time..
> > 
> > So I guess there may be an issue if one has greater than a 480ms
> > round trip time and a 10 gigabit connection where you are using
> > the whole link all by yourself.
> > 
> > > If so, and the pipe is large, has there been any discussion or
> > > concern regarding aliasing or wrap-around of the 16-bit
> > > Identification (ID) field in the IP header?  Is this an
> > > issue, and if so, how is it being addressed?
> > > 
> > 
> > Considering that the scenario is quite far fetched in todays
> > technology over an inter-net. Maybe in a Intra-net.. but then
> > one would hope you would not get more than a 400ms RTT...
> > 
> > And besides all that .. in the future if this really becomes
> > a problem I think SCTP is exandable enough that we can figure
> > out a new chunk or some such to help with this type of situation.
> > 
> > I think we can worry about this problem when we get a bit closer
> > to having a realistic scenario where a problem would exist...
> > 
> > I just don't see it yet... 
> > 
> > Maybe I have missed something in my back-of-the napkin calculations
> > or I am mis-interpreting what circumstances you are envisioning. If
> > so please correct me..
> > 
> > 
> > R
> > 
> > 
> > -- 
> > Randall R. Stewart
> > randall@stewart.chicago.il.us 815-342-5222 (cell phone)
> > 
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Brian F. G. Bidulock    ¦ The reasonable man adapts himself to the ¦
bidulock@openss7.org    ¦ world; the unreasonable one persists in  ¦
http://www.openss7.org/ ¦ trying  to adapt the  world  to himself. ¦
                        ¦ Therefore  all  progress  depends on the ¦
                        ¦ unreasonable man. -- George Bernard Shaw ¦

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 27 06:13:05 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA16284
	for <tsvwg-archive@odin.ietf.org>; Wed, 27 Feb 2002 06:13:05 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA11401
	for tsvwg-archive@odin.ietf.org; Wed, 27 Feb 2002 06:13:09 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA10838;
	Wed, 27 Feb 2002 06:01:08 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA10808
	for <tsvwg@optimus.ietf.org>; Wed, 27 Feb 2002 06:01:04 -0500 (EST)
Received: from mail1.hd.intel.com (hdfdns01.hd.intel.com [192.52.58.10])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA16128
	for <tsvwg@ietf.org>; Wed, 27 Feb 2002 06:00:59 -0500 (EST)
Received: from pysmsxvs01.py.intel.com (pysmsxvs01.py.intel.com [146.152.3.51])
	by mail1.hd.intel.com (8.9.1a+p1/8.9.1/d: relay.m4,v 1.51 2002/02/19 21:12:32 root Exp $) with SMTP id LAA24389
	for <tsvwg@ietf.org>; Wed, 27 Feb 2002 11:00:32 GMT
Received: from pysmsx030.py.intel.com ([146.152.3.52])
 by pysmsxvs01.py.intel.com (NAVGW 2.5.1.16) with SMTP id M2002022706003207628
 for <tsvwg@ietf.org>; Wed, 27 Feb 2002 06:00:32 -0500
Received: by pysmsx030.py.intel.com with Internet Mail Service (5.5.2653.19)
	id <DVHFP9H5>; Wed, 27 Feb 2002 06:00:32 -0500
Message-ID: <A1171C2850D8D4119D7B0002A508C7640146683B@bgsmsx91.iind.intel.com>
From: "Buraga, Partha S" <p_buraga@trillium.com>
To: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Date: Wed, 27 Feb 2002 05:51:53 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: [Tsvwg] Regarding IPv6 address scope
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi

If source or destination address of INIT chunk ( IPv6 header ) is SITE LOCAL
address, can i allow the LINK local address as INIT chunk parameter ? 
other way, if source or destination address of INIT chunk ( IPv6 header ) is
LINK LOCAL address, can i allow the SITE local address as INIT chunk
parameter ?


thanks
partha


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 27 06:28:02 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA16515
	for <tsvwg-archive@odin.ietf.org>; Wed, 27 Feb 2002 06:28:01 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA12185
	for tsvwg-archive@odin.ietf.org; Wed, 27 Feb 2002 06:28:04 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA11510;
	Wed, 27 Feb 2002 06:13:55 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id GAA11434
	for <tsvwg@optimus.ietf.org>; Wed, 27 Feb 2002 06:13:51 -0500 (EST)
Received: from ilsa.franken.de (postfix@ilsa.franken.de [193.175.24.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA16306
	for <tsvwg@ietf.org>; Wed, 27 Feb 2002 06:13:45 -0500 (EST)
Received: from localhost (unknown [212.234.160.52])
	by ilsa.franken.de (Postfix) with ESMTP
	id 13D77245C3; Wed, 27 Feb 2002 12:13:47 +0100 (CET)
Date: Wed, 27 Feb 2002 12:11:18 +0100
Subject: Re: [Tsvwg] Regarding IPv6 address scope
Content-Type: text/plain; charset=US-ASCII; format=flowed
Mime-Version: 1.0 (Apple Message framework v481)
Cc: "'tsvwg@ietf.org'" <tsvwg@ietf.org>
To: "Buraga, Partha S" <p_buraga@trillium.com>
From: Michael Tuexen <Michael.Tuexen@icn.siemens.de>
In-Reply-To: <A1171C2850D8D4119D7B0002A508C7640146683B@bgsmsx91.iind.intel.com>
Message-Id: <B8E224D8-2B72-11D6-8EA3-0030654C1AB6@icn.siemens.de>
Content-Transfer-Encoding: 7bit
X-Mailer: Apple Mail (2.481)
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Hi,

this topic is covered in
http://www.ietf.org/internet-drafts/draft-stewart-tsvwg-sctpipv6-00.txt

best regards
Michael
On Wednesday, February 27, 2002, at 11:51 AM, Buraga, Partha S wrote:

> Hi
>
> If source or destination address of INIT chunk ( IPv6 header ) is SITE 
> LOCAL
> address, can i allow the LINK local address as INIT chunk parameter ?
> other way, if source or destination address of INIT chunk ( IPv6 
> header ) is
> LINK LOCAL address, can i allow the SITE local address as INIT chunk
> parameter ?
>
>
> thanks
> partha
>
>
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
>
Michael.Tuexen@micmac.franken.de


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 27 09:19:50 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA23520
	for <tsvwg-archive@odin.ietf.org>; Wed, 27 Feb 2002 09:19:49 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA21990
	for tsvwg-archive@odin.ietf.org; Wed, 27 Feb 2002 09:19:54 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA20811;
	Wed, 27 Feb 2002 08:59:48 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id IAA20782
	for <tsvwg@optimus.ietf.org>; Wed, 27 Feb 2002 08:59:45 -0500 (EST)
Received: from chuckie.dgms.com (angelica.ulticom.com [208.255.120.2])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA22685
	for <tsvwg@ietf.org>; Wed, 27 Feb 2002 08:59:37 -0500 (EST)
Received: from ulticom.com (localhost [127.0.0.1])
	by chuckie.dgms.com (8.9.3/8.9.3) with ESMTP id IAA06342;
	Wed, 27 Feb 2002 08:58:55 -0500 (EST)
Message-ID: <3C7CE61E.32EB0C7B@ulticom.com>
Date: Wed, 27 Feb 2002 08:58:54 -0500
From: David Lehmann <lehmann@ulticom.com>
X-Mailer: Mozilla 4.77 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Michael Tuexen <Michael.Tuexen@icn.siemens.de>
CC: "Buraga, Partha S" <p_buraga@trillium.com>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>
Subject: Re: [Tsvwg] Regarding IPv6 address scope
References: <B8E224D8-2B72-11D6-8EA3-0030654C1AB6@icn.siemens.de>
Content-Type: multipart/mixed;
 boundary="------------56C6C191D76ECFFDC637DD4E"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


This is a multi-part message in MIME format.
--------------56C6C191D76ECFFDC637DD4E
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit



Michael Tuexen wrote:
> 
> Hi,
> 
> this topic is covered in
> http://www.ietf.org/internet-drafts/draft-stewart-tsvwg-sctpipv6-00.txt

This draft has expired.  Will this draft become an RFC or
be incorporated into another draft which will become an RFC?

Now about the document text...

Can I get a clarification on A2 and B2?
If I have two machines connected by two separate LANs 
(e.g. two crossover ethernet cables), I can have an 
association using both LANs.  The wording of A2/B2 
seems to restrict the association to the single physical 
link used by the INIT packet.

What if I had two machines connect by two separate
networks.  One is a site network and the other is
the global Internet.  Can I have an association which
uses both networks?  The working of A5/B5 seem to
exclude this configuration.

-David
--------------56C6C191D76ECFFDC637DD4E
Content-Type: text/x-vcard; charset=us-ascii;
 name="lehmann.vcf"
Content-Description: Card for David Lehmann
Content-Disposition: attachment;
 filename="lehmann.vcf"
Content-Transfer-Encoding: 7bit

begin:vcard 
n:Lehmann;David
tel;work:856-787-2729
x-mozilla-html:FALSE
url:www.ulticom.com
org:<IMG SRC="http://www.ulticom.com/images/ulticom_logoNS.gif">;Software Engineering
adr:;;1020 Briggs Road;Mt. Laurel;NJ;08054;USA
version:2.1
email;internet:lehmann@ulticom.com
title:AOL IM: davidULCM
fn:David Lehmann
end:vcard

--------------56C6C191D76ECFFDC637DD4E--


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 27 09:46:59 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA25047
	for <tsvwg-archive@odin.ietf.org>; Wed, 27 Feb 2002 09:46:57 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id JAA23782
	for tsvwg-archive@odin.ietf.org; Wed, 27 Feb 2002 09:47:02 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA22586;
	Wed, 27 Feb 2002 09:29:41 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA22557
	for <tsvwg@optimus.ietf.org>; Wed, 27 Feb 2002 09:29:39 -0500 (EST)
Received: from ilsa.franken.de (postfix@ilsa.franken.de [193.175.24.42])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA24059
	for <tsvwg@ietf.org>; Wed, 27 Feb 2002 09:29:34 -0500 (EST)
Received: from localhost (unknown [212.234.160.52])
	by ilsa.franken.de (Postfix) with ESMTP
	id 1E69B245C3; Wed, 27 Feb 2002 15:29:31 +0100 (CET)
Date: Wed, 27 Feb 2002 15:27:01 +0100
Subject: Re: [Tsvwg] Regarding IPv6 address scope
Content-Type: text/plain; charset=US-ASCII; format=flowed
Mime-Version: 1.0 (Apple Message framework v481)
Cc: "Buraga, Partha S" <p_buraga@trillium.com>,
        "'tsvwg@ietf.org'" <tsvwg@ietf.org>
To: David Lehmann <lehmann@ulticom.com>
From: Michael Tuexen <Michael.Tuexen@icn.siemens.de>
In-Reply-To: <3C7CE61E.32EB0C7B@ulticom.com>
Message-Id: <107A6652-2B8E-11D6-8EA3-0030654C1AB6@icn.siemens.de>
Content-Transfer-Encoding: 7bit
X-Mailer: Apple Mail (2.481)
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

David,

I personally think that it would be good to resubmit the ID after
inclusion of the experience get will get at the connectathon
(www.connectathon.org) testing SCTP. It starts tomorrow, SCTP
testing will start on Friday.

These issues will definitely be discussed there. Will be be at
connectathon?

For further comments see below.

Best regards
Michael

On Wednesday, February 27, 2002, at 02:58 PM, David Lehmann wrote:

>
>
> Michael Tuexen wrote:
>>
>> Hi,
>>
>> this topic is covered in
>> http://www.ietf.org/internet-drafts/draft-stewart-tsvwg-sctpipv6-00.txt
>
> This draft has expired.  Will this draft become an RFC or
> be incorporated into another draft which will become an RFC?
>
> Now about the document text...
>
> Can I get a clarification on A2 and B2?
> If I have two machines connected by two separate LANs
> (e.g. two crossover ethernet cables), I can have an
> association using both LANs.  The wording of A2/B2
> seems to restrict the association to the single physical
> link used by the INIT packet.
>
I think this is a problem. Because you could even use
the same link local address on both interfaces, I think.

> What if I had two machines connect by two separate
> networks.  One is a site network and the other is
> the global Internet.  Can I have an association which
> uses both networks?  The working of A5/B5 seem to
> exclude this configuration.
>
You can include the global addresses if you send the INIT
from the site local to the site local address. The INIT-ACk
will also use the site local one as from/to addresses.
If you send the INIT from and to the global one, you can not
include the site local one.

> -David
>
>
Michael.Tuexen@micmac.franken.de


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 27 10:18:46 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA26843
	for <tsvwg-archive@odin.ietf.org>; Wed, 27 Feb 2002 10:18:45 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA25561
	for tsvwg-archive@odin.ietf.org; Wed, 27 Feb 2002 10:18:48 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA24389;
	Wed, 27 Feb 2002 09:59:08 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA24362
	for <tsvwg@optimus.ietf.org>; Wed, 27 Feb 2002 09:59:06 -0500 (EST)
Received: from chuckie.dgms.com (angelica.ulticom.com [208.255.120.2])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA25690
	for <tsvwg@ietf.org>; Wed, 27 Feb 2002 09:59:02 -0500 (EST)
Received: from ulticom.com (localhost [127.0.0.1])
	by chuckie.dgms.com (8.9.3/8.9.3) with ESMTP id JAA19146;
	Wed, 27 Feb 2002 09:58:24 -0500 (EST)
Message-ID: <3C7CF40F.772AFFDB@ulticom.com>
Date: Wed, 27 Feb 2002 09:58:23 -0500
From: David Lehmann <lehmann@ulticom.com>
X-Mailer: Mozilla 4.77 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Michael Tuexen <Michael.Tuexen@icn.siemens.de>,
        IETF Transport Services Working Group <tsvwg@ietf.org>
Subject: Re: [Tsvwg] Regarding IPv6 address scope
References: <107A6652-2B8E-11D6-8EA3-0030654C1AB6@icn.siemens.de>
Content-Type: multipart/mixed;
 boundary="------------15457B08A66A449B93A8027C"
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


This is a multi-part message in MIME format.
--------------15457B08A66A449B93A8027C
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit



Michael Tuexen wrote:
> 
> David,
> 
> I personally think that it would be good to resubmit the ID after
> inclusion of the experience get will get at the connectathon
> (www.connectathon.org) testing SCTP. It starts tomorrow, SCTP
> testing will start on Friday.
> 
> These issues will definitely be discussed there. Will be be at
> connectathon?

No, I will not be there.  I am waiting for the next bake-off.
(At the time of registration, SCTP looked like an after-thought
as it did the previous year.)

Please publish the key points of the discussion on tsvwg.

> > Can I get a clarification on A2 and B2?
> > If I have two machines connected by two separate LANs
> > (e.g. two crossover ethernet cables), I can have an
> > association using both LANs.  The wording of A2/B2
> > seems to restrict the association to the single physical
> > link used by the INIT packet.
> >
> I think this is a problem. Because you could even use
> the same link local address on both interfaces, I think.

I think you are right.  Sun uses the same ethernet address
for all of the interfaces.  Thus, the link-local addresses
will be the same for all... I think.

> > What if I had two machines connect by two separate
> > networks.  One is a site network and the other is
> > the global Internet.  Can I have an association which
> > uses both networks?  The working of A5/B5 seem to
> > exclude this configuration.
> >
> You can include the global addresses if you send the INIT
> from the site local to the site local address. The INIT-ACk
> will also use the site local one as from/to addresses.
> If you send the INIT from and to the global one, you can not
> include the site local one.

I don't see why this should be the case.  Each destination is 
an equally valid path to the receiver.  Thus, both of the 
sender's addresses should be valid, regardless of the path
taken by the INIT.

IMHO, if the user wants to only use the global addresses,
the user should bind only to the global addresses.

-David
--------------15457B08A66A449B93A8027C
Content-Type: text/x-vcard; charset=us-ascii;
 name="lehmann.vcf"
Content-Description: Card for David Lehmann
Content-Disposition: attachment;
 filename="lehmann.vcf"
Content-Transfer-Encoding: 7bit

begin:vcard 
n:Lehmann;David
tel;work:856-787-2729
x-mozilla-html:FALSE
url:www.ulticom.com
org:<IMG SRC="http://www.ulticom.com/images/ulticom_logoNS.gif">;Software Engineering
adr:;;1020 Briggs Road;Mt. Laurel;NJ;08054;USA
version:2.1
email;internet:lehmann@ulticom.com
title:AOL IM: davidULCM
fn:David Lehmann
end:vcard

--------------15457B08A66A449B93A8027C--


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Wed Feb 27 19:43:16 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA26122
	for <tsvwg-archive@odin.ietf.org>; Wed, 27 Feb 2002 19:43:15 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id TAA11893
	for tsvwg-archive@odin.ietf.org; Wed, 27 Feb 2002 19:43:18 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA10769;
	Wed, 27 Feb 2002 19:17:07 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id TAA10740
	for <tsvwg@optimus.ietf.org>; Wed, 27 Feb 2002 19:17:05 -0500 (EST)
Received: from localhost ([211.217.173.10])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id TAA25462
	for <tsvwg@ietf.org>; Wed, 27 Feb 2002 19:17:01 -0500 (EST)
Message-Id: <200202280017.TAA25462@ietf.org>
Reply-To: gotours@dreamwiz.com
From: ¿©Çà°¡ÀÚ<test@test.com>
To: tsvwg@ietf.org
Mime-Version: 1.0
Content-Type: text/html; charset="ks_c_5601-1987"
Date: Thu, 28 Feb 2002 09:10:59 +0900
Subject: [Tsvwg] [ÀÌº¥Æ®±¤°í]¹«·á È¸¿ø°¡ÀÔ ÇÏ½Ã¸é ÇØ¿Ü ¹«·á ¿©Çà±ÇÀ» µå¸³´Ï´Ù.
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=euc-kr">
<title>¿©Çà°¡ÀÚ ¿ÀÇÂ ±â³äÀ¸·Î È¸¿ø °¡ÀÔ½Ã [ÅÂ±¹,ÇÊ¸®ÇÉ,»çÀÌÆÇ] ¹«·á ÇØ¿Ü ¿©Çà±ÇÀ» µå¸³´Ï´Ù.</title>
<meta name="generator" content="Namo WebEditor v5.0">
</head>
<body bgcolor="white" text="black" link="blue" vlink="purple" alink="red">
<table align="center" cellpadding="0" cellspacing="0" width="600">
<tr>
<td width="974">
<p align="center"><img src="http://my.dreamwiz.com/merzone/6.jpg" width="600" height="152" border="0" usemap="#ImageMap1"></p>
</td>
</tr>
<tr>
<td width="974">
<p align="center"><a href="http://www.honeymoonbest.com" target="_blank"><img src="http://my.dreamwiz.com/merzone/7.jpg" width="600" height="170" border="0"></a></p>
</td>
</tr>
<tr>
<td width="974" height="255">
<p align="center"><a href="http://www.honeymoonbest.com" target="_blank"><img src="http://my.dreamwiz.com/merzone/3.jpg" width="600" height="254" border="0"></a></p>
</td>
</tr>
<tr>
<td width="974">
<p align="center"><a href="http://www.honeymoonbest.com" target="_blank"><img src="http://my.dreamwiz.com/merzone/4.jpg" width="600" height="221" border="0"></a></p>
</td>
</tr>
<tr>
<td width="974">
<p align="center"><img src="http://my.dreamwiz.com/merzone/5.jpg" width="600" height="83" border="0"></p>
</td>
</tr>
<tr>
<td width="974">
<P align=center><span style="FONT-SIZE: 9pt"><br></span><span style="FONT-SIZE: 10pt"
><b><font color="blue">[ÀÌº¥Æ®1] = ¿©Çà°¡ÀÚ È¨ÆäÀÌÁö ¿ÀÇÂ ±â³äÀ¸·Î È¸¿ø °¡ÀÔ½Ã [ÅÂ±¹,ÇÊ¸®ÇÉ,»çÀÌÆÇ] <BR>¹«·á
ÇØ¿Ü ¿©Çà±ÇÀ» µå¸³´Ï´Ù.&nbsp;&nbsp; [<FONT color=#ff0080>Ç×°ø±Ç º°µµ</FONT>
]<br>±â°£ : 2002³â 2¿ù18ÀÏ ~ 4¿ù 30ÀÏ ±îÁö ´ë»ó : ¿©Çà°¡ÀÚ ½Å±ÔÈ¸¿ø °¡ÀÔÀÚ
Àü¿øÁõÁ¤<BR><BR>[ÀÌº¥Æ®2] =
ÁÖÀ§ÀÇ&nbsp;°áÈ¥ÇÏ½Ã´Â&nbsp;ºÐµé¿¡°Ô&nbsp;ÀúÈñ&nbsp;¿©Çà°¡ÀÚ¸¦&nbsp;¼Ò°³½ÃÄÑ&nbsp;ÁÖ¼Å¼­&nbsp;½ÅÈ¥¿©ÇàÀÌ&nbsp;<BR>°è¾àµÇ¸é&nbsp;°è¾à°ú&nbsp;µ¿½Ã¿¡&nbsp;°èÁÂ·Î&nbsp;ÀÏ±Ý&nbsp;50,000¿øÀ»&nbsp;ÀÔ±ÝÇØ&nbsp;µå¸³´Ï´Ù.<BR>(´Ü&nbsp;Á¦ÁÖµµ&nbsp;½ÅÈ¥¿©ÇàÀº&nbsp;20,000¿ø)&nbsp;</font></b></span><span style="FONT-SIZE: 9pt"><br></span><span style="FONT-SIZE: 12pt"
><b><a href="http://www.honeymoonbest.com" target="_blank">&nbsp;¿©Çà°¡ÀÚ
È¨ÆäÀÌÁö ¹Ù·Î°¡±â</a><br>&nbsp;</b></span></P>
</td>
</tr>
<tr>
<td width="974"> <p> <span style="FONT-SIZE: 10pt"
><b>±ÍÇÏÀÇ ½Â¶ô¾øÀÌ È«º¸¼º ÀüÀÚ ¿ìÆíÀ» º¸³»°Ô µÈ Á¡ Á¤ÁßÈ÷ »ç°ú µå¸³´Ï´Ù.<BR>Á¤º¸Åë½Å¸ÁÀÌ¿ëÃËÁø¹ý ±ÔÁ¤À» ÁØ¼öÇÏ¿© ±¤°í¸ÞÀÏÀÓÀ» Ç¥½ÃÇÏ¿´À¸¸ç, ¼ö½Å°ÅºÎ ÀåÄ¡¸¦ ¸¶·ÃÇÏ°í ÀÖ½À´Ï´Ù.<BR>±ÍÇÏÀÇ ÀüÀÚ ¿ìÆí ÁÖ¼Ò´Â ÀÎÅÍ³Ý »óÀÇ
°ø°³µÈ Àå¼Ò¿¡¼­ ½ÀµæÇÏ¿´À¸¸ç, ÀúÈñ´Â ±ÍÇÏÀÇ ÀüÀÚ¿ìÆí ÁÖ¼Ò ¿Ü ¾î¶°ÇÑ °³ÀÎÁ¤º¸µµ °¡Áö°í ÀÖÁö ¾ÊÀ¸¹Ç·Î ¾È½ÉÇÏ½Ã±â ¹Ù¶ø´Ï´Ù. ¼ö½ÅÀ» ¿øÄ¡ ¾ÊÀ¸½Ã¸é </b></span><A href="mailto:gotour@hanmir.com"><span style="FONT-SIZE: 10pt"
><b><IMG align=absMiddle border=0 height=21
src="http://my.dreamwiz.com/merzone/Btn_opt_out.gif"
width=57></b></span></A><span style="FONT-SIZE: 10pt"
><b>
¸¦ Å¬¸¯ÇØ ÁÖ½Ê½Ã¿ä. &nbsp;</b></span></p>
</td>
</tr>
<tr>
<td width="974">
<p>&nbsp;</p>
</td>
</tr>
</table>
<map name="ImageMap1">
<area shape="RECT" coords="346,27,570,60" href   ="http://www.honeymoonbest.com" target="_blank">
</map>
</body>
</html>

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Wed Feb 27 22:38:43 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA02185
	for <tsvwg-archive@odin.ietf.org>; Wed, 27 Feb 2002 22:38:43 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id WAA20523
	for tsvwg-archive@odin.ietf.org; Wed, 27 Feb 2002 22:38:47 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA19490;
	Wed, 27 Feb 2002 22:15:00 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id WAA19461
	for <tsvwg@ns.ietf.org>; Wed, 27 Feb 2002 22:14:58 -0500 (EST)
Received: from imo-d01.mx.aol.com (imo-d01.mx.aol.com [205.188.157.33])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA00654
	for <tsvwg@ietf.org>; Wed, 27 Feb 2002 22:14:54 -0500 (EST)
From: DGPickett@aol.com
Received: from DGPickett@aol.com
	by imo-d01.mx.aol.com (mail_out_v32.5.) id l.d7.13a8dbd4 (3973)
	 for <tsvwg@ietf.org>; Wed, 27 Feb 2002 22:14:22 -0500 (EST)
Message-ID: <d7.13a8dbd4.29aefa8e@aol.com>
Date: Wed, 27 Feb 2002 22:14:22 EST
To: tsvwg@ietf.org
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="part1_d7.13a8dbd4.29aefa8e_boundary"
X-Mailer: AOL 6.0 for Windows US sub 10556
Subject: [Tsvwg] Re: SCTP and IP Frag aliasing on LFNs
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


--part1_d7.13a8dbd4.29aefa8e_boundary
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit

I have read of some routers doing what I will casually term:

 - TCP fragmentation (spreading the TCP level payload into two consecutively 
sequenced and not fragmented TCP-IP packets) and

 - TCP aggregation (merging sequentially adjacent TCP packets into fewer 
packets that fit the local MTU).

This can help bridge an unexpected MTU bottleneck, and does not degrade 
packet reliability as much as IP fragmentation, since if the second packet is 
lost, you get 50% throughput (short term) not 0%.

Is it possible that routers could be written to do this sort of gratuitous 
packet rewriting with SCTP, or does the SCTP protocol and checking implicitly 
prevent it?

--part1_d7.13a8dbd4.29aefa8e_boundary
Content-Type: text/html; charset="US-ASCII"
Content-Transfer-Encoding: 7bit

<HTML><FONT FACE=arial,helvetica><FONT  SIZE=2>I have read of some routers doing what I will casually term:
<BR>
<BR> - TCP fragmentation (spreading the TCP level payload into two consecutively sequenced and not fragmented TCP-IP packets) and
<BR>
<BR> - TCP aggregation (merging sequentially adjacent TCP packets into fewer packets that fit the local MTU).
<BR>
<BR>This can help bridge an unexpected MTU bottleneck, and does not degrade packet reliability as much as IP fragmentation, since if the second packet is lost, you get 50% throughput (short term) not 0%.
<BR>
<BR>Is it possible that routers could be written to do this sort of gratuitous packet rewriting with SCTP, or does the SCTP protocol and checking implicitly prevent it?</FONT></HTML>

--part1_d7.13a8dbd4.29aefa8e_boundary--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@ns.ietf.org  Thu Feb 28 06:05:05 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA16762
	for <tsvwg-archive@odin.ietf.org>; Thu, 28 Feb 2002 06:05:01 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id GAA21104
	for tsvwg-archive@odin.ietf.org; Thu, 28 Feb 2002 06:05:05 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA20164;
	Thu, 28 Feb 2002 05:45:20 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id FAA20134
	for <tsvwg@ns.ietf.org>; Thu, 28 Feb 2002 05:45:18 -0500 (EST)
Received: from mail.clavister.com (mail.enternet.se [193.13.79.249])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id FAA16529
	for <tsvwg@ietf.org>; Thu, 28 Feb 2002 05:45:12 -0500 (EST)
Received: (qmail 28141 invoked from network); 28 Feb 2002 10:38:13 -0000
Received: from unknown (HELO smtp-gw) (193.13.79.241)
  by 0 with SMTP; 28 Feb 2002 10:38:13 -0000
	id FA169PT2; Thu, 28 Feb 2002 11:45:05 +0100
Message-ID: <3C7E0A53.59CB0E24@clavister.se>
Date: Thu, 28 Feb 2002 11:45:39 +0100
From: Mikael Olsson <mikael.olsson@clavister.se>
Organization: Clavister AB
X-Mailer: Mozilla 4.79 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: tsvwg@ietf.org
CC: lln@cdt.luth.se, micke@cs.arizona.edu, steve@cs.arizona.edu
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Content-Transfer-Encoding: 8bit
Subject: [Tsvwg] Bad cryptography suggestion in draft-larzon-udplite-04.txt
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit


(I'm not on the tsvwg mailing list; I just caught an oddity in
draft-larzon-udplite-04.txt and thought I'd notify you of it.
Respond with me in the cc list if you want me to read your
response).

It is quite possible that this issue has already been raised
on the mailing list or elsewhere, but I really don't have 
the time to go looking for it, so I thought I'd just go ahead
and send this to the list. -Rather safe than sorry in these 
areas, right?

Quoted from draft-larzon-udplite-04.txt, on the section on 
encryption that does not spread damange across the entire
"non-sensitive" payload of a packet:
   "A class of such transforms would be transforms where the 
   sensitive part is encrypted using a strong transform as 
   usual, and the insensitive part is encrypted by XORing it 
   with a cryptographic hash computed over the cleartext of the 
   sensitive part.  However, it is clear that with the transforms 
   in use today, encryption eliminates the benefits that UDP
   Lite might bring."

From a security standpoint, this is an extraordinarily bad idea.
This is, in effect, a keystream cipher, with a potentially _very_
short keystream keyspace to boot, that may very well re-use the 
key stream twice if/when the "checksummed" portion of UDP Lite 
remains the same between two packets. 

In instances where only the UDP Lite header is checksummed (and
consequently encrypted with strong encrypion), this is exactly what 
will be happening, as one communication stream will always be using
the same ports, and getting the same checksum due to the UDP Lite
checksumming scheme.

As any cryptographer will tell you, re-using the same keystream
pretty much voids the cryptography:

- Let K be the keystream that results from hashing the sensitive part
  of a packet. Assume that the sensitive part of packets A and B
  is the same. This results in K becoming the same for both packets.
- Let A be the "non-sensitive" plaintext from the first packet
- Let B be the "non-sensitive" plaintext from the second packet

The cryptotext of A becomes: A xor K
The cryptotext of B becomes: B xor K

Now, if we take both cryptotexts and xor then with eachother,
we get:
  A xor K xor B xor K
simplified as
  A xor B xor (K xor K)
which, of course, becomes
  A xor B

This is _VERY_ sensitive to statistical analysis. Please reference 
the mistake that microsoft did with their SAM password encryption, 
where they used the same keystrem to encrypt both the NT hash
and the LM hash of the password and instantly made it crackable
with only a slight processing power increase over existing
password crackers.

There are also other potential problems as one can typically expect 
that the "sensitive" area of UDP Lite is much shorter than the
"non-sensitive" area. This would necessitate quite a bit of 
keystream stretching, with degradation of the keystream strength
as a result. Also, if only the UDP header is used as keystream
entropy, you get a maximum keyspace size of 32 bits (source port
and destination port -- the length and checksum are already known /
may be deduced), which can probably be reduced even further by making 
assumptions about the port spans.


I realize that the passage I'm referencing in the draft is only
a suggestion, but it is my belief that even suggestions of
using bad cryptography should be addressed as potientially
dangerous. Given the space allotted to cryptography considerations
in the current draft, I feel that suggested uses of cryptography
are better left out entirely and addressed separately, to allow
for more detailed analysis or perhaps even inclusion in work
being done by other ietf working groups.

However, there is, of course, no harm done in noting the problems
that the cryptography practices of today may cause for the 
indended operation of UDP Lite.


Regards,
Mikael Olsson

-- 
Mikael Olsson, Clavister AB
Storgatan 12, Box 393, SE-891 28 ÖRNSKÖLDSVIK, Sweden
Phone: +46 (0)660 29 92 00   Mobile: +46 (0)70 26 222 05
Fax: +46 (0)660 122 50       WWW: http://www.clavister.com

"Senex semper diu dormit"

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 28 10:15:30 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA27452
	for <tsvwg-archive@odin.ietf.org>; Thu, 28 Feb 2002 10:14:26 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id KAA05476
	for tsvwg-archive@odin.ietf.org; Thu, 28 Feb 2002 10:14:29 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA04313;
	Thu, 28 Feb 2002 09:59:12 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id JAA04198
	for <tsvwg@optimus.ietf.org>; Thu, 28 Feb 2002 09:58:08 -0500 (EST)
Received: from sj-msg-core-3.cisco.com (sj-msg-core-3.cisco.com [171.70.157.152])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA26518
	for <tsvwg@ietf.org>; Thu, 28 Feb 2002 09:58:04 -0500 (EST)
Received: from mira-sjc5-2.cisco.com (mira-sjc5-2.cisco.com [171.71.163.16])
	by sj-msg-core-3.cisco.com (8.11.3/8.9.1) with ESMTP id g1SEvVw22253;
	Thu, 28 Feb 2002 06:57:32 -0800 (PST)
Received: from stewart.chicago.il.us (ssh-sj1.cisco.com [171.68.225.134])
	by mira-sjc5-2.cisco.com (Mirapoint)
	with ESMTP id ABM28304;
	Thu, 28 Feb 2002 06:57:32 -0800 (PST)
Message-ID: <3C7E455B.98BB66E0@stewart.chicago.il.us>
Date: Thu, 28 Feb 2002 08:57:31 -0600
From: Randall Stewart <randall@stewart.chicago.il.us>
X-Mailer: Mozilla 4.76 [en] (X11; U; Linux 2.2.12 i386)
X-Accept-Language: en
MIME-Version: 1.0
To: David Lehmann <lehmann@ulticom.com>
CC: Michael Tuexen <Michael.Tuexen@icn.siemens.de>,
        IETF Transport Services Working Group <tsvwg@ietf.org>
Subject: Re: [Tsvwg] Regarding IPv6 address scope
References: <107A6652-2B8E-11D6-8EA3-0030654C1AB6@icn.siemens.de> <3C7CF40F.772AFFDB@ulticom.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

David/Michael:

Thre are problems with link/site/global addresses in IPv6
for sure... The draft does need an update... I re-read it
a few weeks back and planned on getting to it... but I ran
out of cycles :<

Even at that there are things I am sure I have not thought
of or discovered in my playing... hopefully at connect-a-thon
we can come up with a more complete set of rules...

A couple of thoughts:


If I send From/To a link local...
  - I can NOT necessarily include all my other link
    locals because:
     a) How do I know that my peer also has interfaces 
        on the same lan's that I do?
     b) Even if I know that my peer DOES have interfaces
        in common with ALL my interfaces, how can I
        know or inform my peer what is the valid scope id
        for his end of our common interfaces... its not
        a problem when he receives from a link local, he can
        dig down and find the scope id off the receive interface..
        but the other interfaces unless the local endpoint is
        somehow configured... and it still can tell the peer..
 
        Remember I may be using fe80::1 on say my second interface
        but my peer may have interfaces a, b, and c. Maybe my
        second interface is on b, but my peer also has a neighbor
        (not me) on c using the same address.  If he gets a INIT
        chunk from me listing fe80::1 he has no idea if I am the
        neighbor on b or c?.

For the above reasons our current BSD implementation in KAME
will only allow link locals to be included if:
   - You are sending to a link local.. but the link
     local is NOT put in the list of addresses.. it gets
     in by inference as being a source address on the INIT/INIT-ACK.

   - Or if you are sending to a loopback address. This we have
     not played with to much though.


For site locals its a similar thing.. I must have a site local
for a destination before I allow the scoping to list site locals. Right
now we have only one site.. but once you start getting the use of
scope'd sites.. the same issues start to creep in like link locals.

Its really a gummy mess.. and we do need to think on this and
have another spin of the draft... 

R
      
     


David Lehmann wrote:
> 
> Michael Tuexen wrote:
> >
> > David,
> >
> > I personally think that it would be good to resubmit the ID after
> > inclusion of the experience get will get at the connectathon
> > (www.connectathon.org) testing SCTP. It starts tomorrow, SCTP
> > testing will start on Friday.
> >
> > These issues will definitely be discussed there. Will be be at
> > connectathon?
> 
> No, I will not be there.  I am waiting for the next bake-off.
> (At the time of registration, SCTP looked like an after-thought
> as it did the previous year.)
> 
> Please publish the key points of the discussion on tsvwg.
> 
> > > Can I get a clarification on A2 and B2?
> > > If I have two machines connected by two separate LANs
> > > (e.g. two crossover ethernet cables), I can have an
> > > association using both LANs.  The wording of A2/B2
> > > seems to restrict the association to the single physical
> > > link used by the INIT packet.
> > >
> > I think this is a problem. Because you could even use
> > the same link local address on both interfaces, I think.
> 
> I think you are right.  Sun uses the same ethernet address
> for all of the interfaces.  Thus, the link-local addresses
> will be the same for all... I think.
> 
> > > What if I had two machines connect by two separate
> > > networks.  One is a site network and the other is
> > > the global Internet.  Can I have an association which
> > > uses both networks?  The working of A5/B5 seem to
> > > exclude this configuration.
> > >
> > You can include the global addresses if you send the INIT
> > from the site local to the site local address. The INIT-ACk
> > will also use the site local one as from/to addresses.
> > If you send the INIT from and to the global one, you can not
> > include the site local one.
> 
> I don't see why this should be the case.  Each destination is
> an equally valid path to the receiver.  Thus, both of the
> sender's addresses should be valid, regardless of the path
> taken by the INIT.
> 
> IMHO, if the user wants to only use the global addresses,
> the user should bind only to the global addresses.
> 
> -David

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 28 14:43:46 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA15465
	for <tsvwg-archive@odin.ietf.org>; Thu, 28 Feb 2002 14:43:46 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id OAA24869
	for tsvwg-archive@odin.ietf.org; Thu, 28 Feb 2002 14:43:49 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA23585;
	Thu, 28 Feb 2002 14:24:35 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id OAA23553
	for <tsvwg@optimus.ietf.org>; Thu, 28 Feb 2002 14:24:31 -0500 (EST)
Received: from motgate.mot.com (motgate.mot.com [129.188.136.100])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA14365
	for <tsvwg@ietf.org>; Thu, 28 Feb 2002 14:24:27 -0500 (EST)
Received: [from pobox3.mot.com (pobox3.mot.com [10.64.251.242]) by motgate.mot.com (motgate 2.1) with ESMTP id MAA06162; Thu, 28 Feb 2002 12:24:25 -0700 (MST)]
Received: [from relay1.cig.mot.com (relay1.cig.mot.com [136.182.15.23]) by pobox3.mot.com (MOT-pobox3 2.0) with ESMTP id MAA13324; Thu, 28 Feb 2002 12:12:46 -0700 (MST)]
Received: from quantix20.cig.mot.com (quantix20 [160.47.58.31])
	by relay1.cig.mot.com (8.11.4/8.11.4) with ESMTP id g1SJOH922701;
	Thu, 28 Feb 2002 13:24:17 -0600 (CST)
Received: from email.mot.com (d50-384a.cig.mot.com [160.47.56.74])
	by quantix20.cig.mot.com (8.8.8+Sun/8.8.8) with ESMTP id NAA26947;
	Thu, 28 Feb 2002 13:29:15 -0600 (CST)
Message-ID: <3C7E84D5.2CFCD625@email.mot.com>
Date: Thu, 28 Feb 2002 13:28:21 -0600
From: Qiaobing Xie <qxie1@email.mot.com>
X-Mailer: Mozilla 4.77 [en] (Win98; U)
X-Accept-Language: en
MIME-Version: 1.0
To: DGPickett@aol.com
CC: tsvwg@ietf.org
Subject: Re: [Tsvwg] Re: SCTP and IP Frag aliasing on LFNs
References: <d7.13a8dbd4.29aefa8e@aol.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 7bit

Similar (but not exact) thing can be done with SCTP. For example, if a
router gets an SCTP packet which carries multiple data chunks and is too
big for the MTU of the next hop, it could in theory split the
multi-data-chunk SCTP packet into multiple SCTP packets of smaller size
and forward them on. But there is no guarantee that each of the
resultant packets will fit into the MTU. If you want to further fragment
an oversized data chunk into small pieces, there will be problems with
SCTP. This is because SCTP is message-oriented and it has to preserve
the boundaries of the original user message. TCP has no such constrain
and can therefore split (or combine) data segments without worrying
break the protocol semantics.

-Qiaobing

DGPickett@aol.com wrote:
> 
> I have read of some routers doing what I will casually term:
> 
> - TCP fragmentation (spreading the TCP level payload into two
> consecutively sequenced and not fragmented TCP-IP packets) and
> 
> - TCP aggregation (merging sequentially adjacent TCP packets into
> fewer packets that fit the local MTU).
> 
> This can help bridge an unexpected MTU bottleneck, and does not
> degrade packet reliability as much as IP fragmentation, since if the
> second packet is lost, you get 50% throughput (short term) not 0%.
> 
> Is it possible that routers could be written to do this sort of
> gratuitous packet rewriting with SCTP, or does the SCTP protocol and
> checking implicitly prevent it?

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 28 15:23:55 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA17748
	for <tsvwg-archive@odin.ietf.org>; Thu, 28 Feb 2002 15:23:54 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA27660
	for tsvwg-archive@odin.ietf.org; Thu, 28 Feb 2002 15:23:58 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA26799;
	Thu, 28 Feb 2002 15:12:02 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA26767
	for <tsvwg@optimus.ietf.org>; Thu, 28 Feb 2002 15:11:59 -0500 (EST)
Received: from c003.snv.cp.net (c003-h000.c003.snv.cp.net [209.228.32.214])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id PAA17029
	for <tsvwg@ietf.org>; Thu, 28 Feb 2002 15:11:54 -0500 (EST)
Received: (cpmta 16210 invoked from network); 28 Feb 2002 12:11:25 -0800
Received: from 66.123.21.234 (HELO littlejoy)
  by smtp.telocity.com (209.228.32.214) with SMTP; 28 Feb 2002 12:11:25 -0800
X-Sent: 28 Feb 2002 20:11:25 GMT
From: "Douglas Otis" <dotis@sanlight.net>
To: "Tsvwg" <tsvwg@ietf.org>
Date: Thu, 28 Mar 2002 12:11:35 -0800
Message-ID: <NEBBJGDMMLHHCIKHGBEJAEJMCOAA.dotis@sanlight.net>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0)
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4807.1700
Content-Transfer-Encoding: 8bit
Subject: [Tsvwg] Intel Forum presentation of RDMA
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

FWIW: Intel Developer Forum Conference, Spring 2002 Feb, 27th.

James (Jim) Pickerton presenting for Microsoft endorsed TUF-TCP/RDMA as the
solution for the future and stated TUF was winning approval within the IETF.
When I asked why TUF rather than SCTP, Jim responded that SCTP was too new
and buggy and would not be ready for at least five years.  He also added
that TUF could be implemented by changing as few as 105 lines of code within
the TCP stack.  The 'call to action' was all applications should incorporate
TUF within their next generation.

TUF is not enough.  This draft does not elucidate what lines of code are to
be changed within TCP.  The transport must be seen as the means of accessing
the IP layer.  How this layer is accessed should be included in a document
adding additional mechanisms for either the IP or TCP layer.  From this
draft it is not clear what layer is even being modified.

The presentation referenced Bernard Aboda's web site for further
information:
http://www.drizzle.com/~aboba/RDMA/rdma-faq.html

Once again a touted RDMA FAQ mis-information that should be corrected.

"2.3 What about SCTP? Can’t SCTP be used to solve the problem?
SCTP addresses the head-of-line blocking issue. SCTP chunking does not solve
the header alignment problem because only the first chunk will be
self-describing. SCTP also does not provide RDMA semantics, for example
random access memory operations.

SCTP streams (or TCP connections) can not be used to effect header data
splitting because ULPs generally require in-order delivery between headers
and data."

The SCTP TLV structure is the type of structure used in IPv6 as an effective
way of bundling objects.  Within SCTP these objects are aligned to 32 bit
boundaries as well.  The benefit is consistent packet size.  I would also
hope that the SCTP-DDP shim draft makes the issue of header/payload
splitting clear.  Each packet within this scheme is independent of the next
and does not require ordered delivery to incorporate DDP headers.  The
number of exceptional conditions created by the TUF scheme together with a
tacit recommendation of improved error checking makes hardware
implementations of TUF-TCP problematic and application specific.  RDMA was
presented as an ultimate general transport of which other protocols must
adopt.  When asked from the audience about the relative lack of security
offered by TCP/RDMA, "It is no worse than TCP and one could always use
IPSEC."

Doug


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 28 15:27:31 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA17966
	for <tsvwg-archive@odin.ietf.org>; Thu, 28 Feb 2002 15:27:31 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA27879
	for tsvwg-archive@odin.ietf.org; Thu, 28 Feb 2002 15:27:35 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA26921;
	Thu, 28 Feb 2002 15:13:11 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA26892
	for <tsvwg@optimus.ietf.org>; Thu, 28 Feb 2002 15:13:09 -0500 (EST)
Received: from mx10.radisys.com (mx10.radisys.com [206.102.10.36])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA17088
	for <tsvwg@ietf.org>; Thu, 28 Feb 2002 15:13:04 -0500 (EST)
From: Simon.Zhuang@radisys.com
Received: by mx10.radisys.com (Postfix, from userid 5)
	id 8483C142EEA; Thu, 28 Feb 2002 12:12:36 -0800 (PST)
Received: from UNKNOWN(206.103.52.220), claiming to be "notes.radisys.com"
 via SMTP by mx10, id smtpdAAAjl0_R_; Thu Feb 28 12:12:34 2002
To: tsvwg@ietf.org
X-Mailer: Lotus Notes Release 5.0.8  June 18, 2001
Message-ID: <OFEFD9D5AF.BB550802-ON85256B6E.0068CF6F@radisys.com>
Date: Thu, 28 Feb 2002 15:01:07 -0500
X-MIMETrack: Serialize by Router on HQ_ACS_1/Radisys_Corporation/US(Release 5.0.8 |June
 18, 2001) at 02/28/2002 12:06:23 PM
MIME-Version: 1.0
Content-type: text/plain; charset=us-ascii
Subject: [Tsvwg] Path heartbeat
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hi,

I have a question about the heartbeat.

Assumption:
Endpoint A is multihomed with 2 ipaddresses (addrA1 and addrA2)
Endpoint B is multihomed with 2 ipaddress(addrB1 and addrB2) as well
There is only one association established between A and B
No user data traffic

In RFC2960 (Page 93) in section 8.3:
  " By default, an SCTP endpoint shall monitor the reachability of the
   idle destination transport address(es) of its peer by sending a
   HEARTBEAT chunk periodically to the destination transport
   address(es)."

Based on the RFC, because the source address is not considered, it is not
clear if heartbeat should be sent from A to B for all 4 paths:
addrA1 to addrB1
addrA1 to addrB2
addrA2 to addrB1
addrA2 to addrB2

Can any one in the list help clarify it?

Thanks in advance.

Simon Zhuang


_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 28 15:44:32 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA18787
	for <tsvwg-archive@odin.ietf.org>; Thu, 28 Feb 2002 15:44:27 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id PAA29065
	for tsvwg-archive@odin.ietf.org; Thu, 28 Feb 2002 15:44:31 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA28006;
	Thu, 28 Feb 2002 15:30:05 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id PAA27958
	for <tsvwg@optimus.ietf.org>; Thu, 28 Feb 2002 15:30:02 -0500 (EST)
Received: from mgw-x2.nokia.com (mgw-x2.nokia.com [131.228.20.22])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA18128
	for <tsvwg@ietf.org>; Thu, 28 Feb 2002 15:29:57 -0500 (EST)
From: Ivan.Arias-Rodriguez@nokia.com
Received: from esvir04nok.ntc.nokia.com (esvir04nokt.ntc.nokia.com [172.21.143.36])
	by mgw-x2.nokia.com (Switch-2.1.0/Switch-2.1.0) with ESMTP id g1SKU8Z21177
	for <tsvwg@ietf.org>; Thu, 28 Feb 2002 22:30:08 +0200 (EET)
Received: from esebh001.NOE.Nokia.com (unverified) by esvir04nok.ntc.nokia.com
 (Content Technologies SMTPRS 4.2.5) with ESMTP id <T595b45f610ac158f24077@esvir04nok.ntc.nokia.com>;
 Thu, 28 Feb 2002 22:30:00 +0200
Received: from esebe017.NOE.Nokia.com ([172.21.138.56]) by esebh001.NOE.Nokia.com with Microsoft SMTPSVC(5.0.2195.3779);
	 Thu, 28 Feb 2002 22:29:59 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.0.5762.3
content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Subject: RE: [Tsvwg] Path heartbeat
Date: Thu, 28 Feb 2002 22:29:59 +0200
Message-ID: <58AEDE27D48F884285007ABF827FF263257B9C@esebe017.NOE.Nokia.com>
Thread-Topic: [Tsvwg] Path heartbeat
Thread-Index: AcHAlGg6hizCclrWTWefSUWkb1kirgAAXxoQ
To: <Simon.Zhuang@radisys.com>, <tsvwg@ietf.org>
X-OriginalArrivalTime: 28 Feb 2002 20:29:59.0502 (UTC) FILETIME=[B14856E0:01C1C096]
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by optimus.ietf.org id PAA27959
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org
Content-Transfer-Encoding: 8bit

	Hi Simon!

	This issue about source address has been debated several times in the list already...

	The conclusion is that basically, once you set the destination IP address, your routing tables are the ones that choose the souce address (and network adapter), and thus there isn't much you can do about it.

	So, basically you only have two paths. And this is what the RFC talks about. For every destination address you just have to send from one source address (the one chosen by the routing tables).

	I hope this helps to clarify (and I hope not to be wrong either! :->)

	BR Iván Arias Rodríguez

> -----Original Message-----
> From: ext Simon.Zhuang@radisys.com [mailto:Simon.Zhuang@radisys.com]
> Sent: 28 February 2002 22:01
> To: tsvwg@ietf.org
> Subject: [Tsvwg] Path heartbeat
> 
> 
> Hi,
> 
> I have a question about the heartbeat.
> 
> Assumption:
> Endpoint A is multihomed with 2 ipaddresses (addrA1 and addrA2)
> Endpoint B is multihomed with 2 ipaddress(addrB1 and addrB2) as well
> There is only one association established between A and B
> No user data traffic
> 
> In RFC2960 (Page 93) in section 8.3:
>   " By default, an SCTP endpoint shall monitor the reachability of the
>    idle destination transport address(es) of its peer by sending a
>    HEARTBEAT chunk periodically to the destination transport
>    address(es)."
> 
> Based on the RFC, because the source address is not 
> considered, it is not
> clear if heartbeat should be sent from A to B for all 4 paths:
> addrA1 to addrB1
> addrA1 to addrB2
> addrA2 to addrB1
> addrA2 to addrB2
> 
> Can any one in the list help clarify it?
> 
> Thanks in advance.
> 
> Simon Zhuang
> 
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/tsvwg
> 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 28 17:56:37 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA24707
	for <tsvwg-archive@odin.ietf.org>; Thu, 28 Feb 2002 17:56:36 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id RAA07733
	for tsvwg-archive@odin.ietf.org; Thu, 28 Feb 2002 17:56:40 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA07010;
	Thu, 28 Feb 2002 17:41:37 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id RAA06981
	for <tsvwg@optimus.ietf.org>; Thu, 28 Feb 2002 17:41:35 -0500 (EST)
Received: from yourwebsite.com (evrtwa1-ar13-4-3-033-042.evrtwa1.vz.dsl.gtei.net [4.3.33.42])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id RAA24049
	for <tsvwg@ietf.org>; Thu, 28 Feb 2002 17:41:30 -0500 (EST)
From: dailythrower@yahoo.com
Message-Id: <200202282241.RAA24049@ietf.org>
Reply-To: makemoney@big-marketing.com
To: tsvwg@ietf.org
Mime-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"
Date: Thu, 28 Feb 2002 14:20:35 -0800
Subject: [Tsvwg] Make $1250 a day with this confidential report
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org

Hello, 

Just thought I'd fill you in on an amazing Million 
Dollar report that uncovers confidential information
you are not supposed to know about. 

It reveals hush, hush info like...

*How a banking secret, which bankers don't want
published, can help you get $5,000 to $15,000 to
use for any purpose within 30 days. 

*How to use a "Back Door" method to get listed on 
Yahoo. There is a back door to Yahoo that has been 
kept hush hush until recently. Best of all it takes 
only 1 minute to do. 

*How to use an extremely powerful technique to get 
your web site listed at the top of the major search 
engines and directories like Yahoo, AOL, Lycos and 
more within FIVE Business days. This technique is 
so new it was just introduced to the internet in 
November 2001. 

*Get a list of over 100 major corporations offering 
legitimate work opportunities for telecommute home 
workers paying $500 to $1,500 per week. In today's 
economy corporations like Bank of America, Dell 
Computer, Motorala and many other such companies 
finding it less costly to hire people to work from 
their homes. The list includes contact information 
for the corporations that need your help.

*How to open a checking account if you currently are 
unable to because you have a bad rating with Chex-
systems or Telecheck.  

*How you can get 1,000,000 hits to your web site in 
1 year. Even Microsoft has tried to patent this one
but can't. 

*How to get $3,000 to $50,000 within 1 week regard-
less of your credit if you currently have a merchant account. 

*How you can resell this information with a tested, 
proven Sales Letter that is EXPLODING wiht sales
everywhere. 

WARNING! Check out this updated, FREE REport before
you try any other of the many business plans on the
internet. 

This information may not be available much longer.
This may be THE LAST TIME you see this report!

Get yours before the deadline by sending a Blank
Email to:

mailto: makemoney@big-marketing.com

It will arrive in just a few minutes. 

Yours truly,

Chase 

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



From daemon@optimus.ietf.org  Thu Feb 28 21:17:52 2002
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA29376
	for <tsvwg-archive@odin.ietf.org>; Thu, 28 Feb 2002 21:17:51 -0500 (EST)
Received: (from daemon@localhost)
	by optimus.ietf.org (8.9.1a/8.9.1) id VAA17331
	for tsvwg-archive@odin.ietf.org; Thu, 28 Feb 2002 21:17:55 -0500 (EST)
Received: from optimus.ietf.org (localhost [127.0.0.1])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA16210;
	Thu, 28 Feb 2002 20:56:47 -0500 (EST)
Received: from ietf.org (odin [132.151.1.176])
	by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id UAA16185
	for <tsvwg@optimus.ietf.org>; Thu, 28 Feb 2002 20:56:45 -0500 (EST)
Received: from imo-d09.mx.aol.com (imo-d09.mx.aol.com [205.188.157.41])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA28979
	for <tsvwg@ietf.org>; Thu, 28 Feb 2002 20:56:41 -0500 (EST)
From: DGPickett@aol.com
Received: from DGPickett@aol.com
	by imo-d09.mx.aol.com (mail_out_v32.5.) id l.97.23d8088f (3975);
	Thu, 28 Feb 2002 20:56:03 -0500 (EST)
Message-ID: <97.23d8088f.29b039b3@aol.com>
Date: Thu, 28 Feb 2002 20:56:03 EST
Subject: Re: [Tsvwg] Re: SCTP and IP Frag aliasing on LFNs
To: qxie1@email.mot.com
CC: tsvwg@ietf.org
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="part1_97.23d8088f.29b039b3_boundary"
X-Mailer: AOL 6.0 for Windows US sub 10556
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
X-BeenThere: tsvwg@ietf.org


--part1_97.23d8088f.29b039b3_boundary
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit

By my understanding, the router doing the TCP packet split/merge act is 
implicitly aware of the local MTU limits.  It could take 3 1500 byte packets 
and make a 4K packet to send out now and a 500 byte packet to send out later 
if nothing showed up to merge.  Going the other way, it could take a 4K 
packet and make 2 1500 packets to send now and a 1K packet to send later if 
nothing showed to merge it with.  A router might even have a bias for sending 
max MTU packets first, then smallest first, to maximize the opportunity for 
packet merge, which my example shows helps bandwidth loss from header and 
other per packet overheads when the MTU changes either way, even more than it 
helps for flat MTU.

Personally, I don't see that many 'dribbly' TCP applications that develop a 
lot of sequential sub-MTU packets in the same direction with acknowledgment, 
just those with big file transfer packets or mini packets of keystrokes, 
keystroke echoes, small messages, keepalive, or ack.

--part1_97.23d8088f.29b039b3_boundary
Content-Type: text/html; charset="US-ASCII"
Content-Transfer-Encoding: 7bit

<HTML><FONT FACE=arial,helvetica><FONT  SIZE=2>By my understanding, the router doing the TCP packet split/merge act is implicitly aware of the local MTU limits. &nbsp;It could take 3 1500 byte packets and make a 4K packet to send out now and a 500 byte packet to send out later if nothing showed up to merge. &nbsp;Going the other way, it could take a 4K packet and make 2 1500 packets to send now and a 1K packet to send later if nothing showed to merge it with. &nbsp;A router might even have a bias for sending max MTU packets first, then smallest first, to maximize the opportunity for packet merge, which my example shows helps bandwidth loss from header and other per packet overheads when the MTU changes either way, even more than it helps for flat MTU.
<BR>
<BR>Personally, I don't see that many 'dribbly' TCP applications that develop a lot of sequential sub-MTU packets in the same direction with acknowledgment, just those with big file transfer packets or mini packets of keystrokes, keystroke echoes, small messages, keepalive, or ack.</FONT></HTML>

--part1_97.23d8088f.29b039b3_boundary--

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
https://www1.ietf.org/mailman/listinfo/tsvwg



