From tsvwg-bounces@ietf.org Mon Apr 03 16:12:44 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQVPl-0001Gs-PM; Mon, 03 Apr 2006 16:12:41 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQVPj-0001ET-SF; Mon, 03 Apr 2006 16:12:39 -0400
Received: from mailer1.psc.edu ([128.182.58.100])
	by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FQVPi-0002z2-JN; Mon, 03 Apr 2006 16:12:39 -0400
Received: from [128.182.160.77] (tibook.psc.edu [128.182.160.77])
	(authenticated bits=0)
	by mailer1.psc.edu (8.13.5.20060308/8.13.3) with ESMTP id
	k33KCYKn005513
	(version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO);
	Mon, 3 Apr 2006 16:12:36 -0400 (EDT)
Message-ID: <443181B1.9030107@psc.edu>
Date: Mon, 03 Apr 2006 16:12:33 -0400
From: John Heffner <jheffner@psc.edu>
User-Agent: Thunderbird 1.5 (Macintosh/20051201)
MIME-Version: 1.0
To: tsvwg@ietf.org, int-area@ietf.org
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 79899194edc4f33a41f49410777972f8
Cc: 
Subject: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

We're moving this doc through the AD-sponsored informational track. 
Though it's not a working group item, we'd like to solicit feedback 
(from both tsvwg and int-area).  Please note that the purpose of this 
document is only to describe a known problem, not a solution.

Abstract:
    IPv4 fragmentation is not sufficiently robust for general use in
    today's Internet.  The 16-bit IP identification field is not large
    enough to prevent frequent incorrectly assembled IP fragments, and
    the TCP and UDP checksums are insufficient to prevent the resulting
    corrupted datagrams from being delivered to higher protocol layers.
    This note describes some easily reproduced experiments demonstrating
    the problem, and discusses some of the operational implications of
    these observations.

Currently available at:
http://www.ietf.org/internet-drafts/draft-heffner-frag-harmful-00.txt

Thanks,
   -John




From tsvwg-bounces@ietf.org Mon Apr 03 16:36:57 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQVnC-000666-Ch; Mon, 03 Apr 2006 16:36:54 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQVnB-00065y-DZ; Mon, 03 Apr 2006 16:36:53 -0400
Received: from palrel13.hp.com ([156.153.255.238])
	by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FQVn8-0004QC-Uo; Mon, 03 Apr 2006 16:36:53 -0400
Received: from tardy.cup.hp.com (tardy.cup.hp.com [15.244.56.217])
	by palrel13.hp.com (Postfix) with ESMTP id 615B434504;
	Mon,  3 Apr 2006 13:36:50 -0700 (PDT)
Received: from hp.com (localhost [127.0.0.1])
	by tardy.cup.hp.com (8.9.3 (PHNE_28810)/8.9.3 SMKit7.02) with ESMTP id
	NAA14906; Mon, 3 Apr 2006 13:36:49 -0700 (PDT)
Message-ID: <44318761.6090500@hp.com>
Date: Mon, 03 Apr 2006 13:36:49 -0700
From: Rick Jones <rick.jones2@hp.com>
User-Agent: Mozilla/5.0 (X11; U; HP-UX 9000/785; en-US; rv:1.6) Gecko/20040304
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: John Heffner <jheffner@psc.edu>
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
References: <443181B1.9030107@psc.edu>
In-Reply-To: <443181B1.9030107@psc.edu>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: f60d0f7806b0c40781eee6b9cd0b2135
Cc: int-area@ietf.org, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

John Heffner wrote:
> We're moving this doc through the AD-sponsored informational track. 
> Though it's not a working group item, we'd like to solicit feedback 
> (from both tsvwg and int-area).  Please note that the purpose of this 
> document is only to describe a known problem, not a solution.
> 
> Abstract:
>    IPv4 fragmentation is not sufficiently robust for general use in
>    today's Internet.  The 16-bit IP identification field is not large
>    enough to prevent frequent incorrectly assembled IP fragments, and
>    the TCP and UDP checksums are insufficient to prevent the resulting
>    corrupted datagrams from being delivered to higher protocol layers.
>    This note describes some easily reproduced experiments demonstrating
>    the problem, and discusses some of the operational implications of
>    these observations.
> 
> Currently available at:
> http://www.ietf.org/internet-drafts/draft-heffner-frag-harmful-00.txt

Ah, Frankengrams :) Fun things.  I remember we had "issues" wrt back in 
'91 or so even on 10 Mbit/s Ethernet.

> The case of particular concern occurs when the first fragment of a
> datagram is lost in the network.  The remaining fragments will be
> stored in the fragment reassembly buffer.  When the sender wraps the
> ID field, the first fragment of the new datagram will be mis-
> associated with the rest of the old datagram.  Assuming the fragments
> are delivered in order, the rest of the new datagram will be now be
> incomplete (since it is missing the first fragment), so they will be
> saved in the fragment reassembly buffer, forming a cycle that repeats
> every 65536 datagrams.  It is possible to have a number of
> simultaneous cycles, bounded by the size of the fragment reassembly
> buffer.

Does it really matter if it is the first fragment?  I would think that 
cycling is possible so long as (any of) the lost fragments are simply 
not the last fragment when one assumes the fragments are flowing in-order.

Also, to ass-u-me that fragments flow through the network in-order is 
probably not a good thing?  Load-balanced link aggregates may send 
fragments of one datagram across more than one link in the aggregate, 
and I know of at least one NIC, albeit old and likely unused today, that 
would send the first fragment last as part of accumulating an offloaded 
Checksum (HP-PB FDDI NIC).

A while back, over in the "netdev" mailing list, the we were discussing 
ways to minimize the risk of frankengrams springing to life based on 
ass-u-me-ing :) that if one saw N or more datagrams beyond the one(s) 
being reassembled, the likelihood of those datagrams successfully 
reassembling was epsilon. While the specifics of that may be more 
"linux-centric" than an RFC should be, I suspect the discussion might 
provide a starting point for a useful, additional section in the RFC 
discussing ways transport stacks might minimize the likelihood of 
Frankengrams.

I'd probably add to that list what we put into HP-UX 8.07 and what may 
be in several stacks today - drivers accepting "packet trains" whereby 
the driver either queues the entire train, or none of it.  That was 
implemented to solve the problem of transmit queues filling part way 
through the enqueing of the fragments of a given datagram.  Without that 
behaviour, it is possible to send nothing but unreassemblable/potential 
Frankengram datagrams.  At least for the duration of a burst of sending 
faster than the link.  An intermediate device doing fragmentation might 
want to employ that as well.

sincerely,

rick jones




From tsvwg-bounces@ietf.org Mon Apr 03 18:09:16 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQXEX-000649-6N; Mon, 03 Apr 2006 18:09:13 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQXEW-000641-0X; Mon, 03 Apr 2006 18:09:12 -0400
Received: from mailer1.psc.edu ([128.182.58.100])
	by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FQXEV-000858-Ni; Mon, 03 Apr 2006 18:09:11 -0400
Received: from [128.182.160.77] (tibook.psc.edu [128.182.160.77])
	(authenticated bits=0)
	by mailer1.psc.edu (8.13.5.20060308/8.13.3) with ESMTP id
	k33M99Y5007894
	(version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO);
	Mon, 3 Apr 2006 18:09:09 -0400 (EDT)
Message-ID: <44319D02.1050306@psc.edu>
Date: Mon, 03 Apr 2006 18:09:06 -0400
From: John Heffner <jheffner@psc.edu>
User-Agent: Thunderbird 1.5 (Macintosh/20051201)
MIME-Version: 1.0
To: Rick Jones <rick.jones2@hp.com>
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
References: <443181B1.9030107@psc.edu> <44318761.6090500@hp.com>
In-Reply-To: <44318761.6090500@hp.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 7aafa0432175920a4b3e118e16c5cb64
Cc: int-area@ietf.org, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Rick Jones wrote:
>> The case of particular concern occurs when the first fragment of a
>> datagram is lost in the network.  The remaining fragments will be
>> stored in the fragment reassembly buffer.  When the sender wraps the
>> ID field, the first fragment of the new datagram will be mis-
>> associated with the rest of the old datagram.  Assuming the fragments
>> are delivered in order, the rest of the new datagram will be now be
>> incomplete (since it is missing the first fragment), so they will be
>> saved in the fragment reassembly buffer, forming a cycle that repeats
>> every 65536 datagrams.  It is possible to have a number of
>> simultaneous cycles, bounded by the size of the fragment reassembly
>> buffer.
> 
> Does it really matter if it is the first fragment?  I would think that 
> cycling is possible so long as (any of) the lost fragments are simply 
> not the last fragment when one assumes the fragments are flowing in-order.

It could matter.  If it's a middle fragment, you'll receive some 
duplicate earlier ones before the datagram can be completed.  If the OS 
has a consistency check here (that the data is the same), it could toss 
the whole thing when it fails.  I'm not aware of any implementation that 
does this, but if you're just talking about the first fragment, it's 
more clear that it's a problem.


> Also, to ass-u-me that fragments flow through the network in-order is 
> probably not a good thing?  Load-balanced link aggregates may send 
> fragments of one datagram across more than one link in the aggregate, 
> and I know of at least one NIC, albeit old and likely unused today, that 
> would send the first fragment last as part of accumulating an offloaded 
> Checksum (HP-PB FDDI NIC).

However, data is often delivered in order, so it serves as a good 
example. :)  Assuming in-order delivery is partially for the sake of 
clarity.  Any order (reverse, first-packet-last) that's always repeated 
has the same property.  A randomized delivery order is not as bad as the 
worst case.


> A while back, over in the "netdev" mailing list, the we were discussing 
> ways to minimize the risk of frankengrams springing to life based on 
> ass-u-me-ing :) that if one saw N or more datagrams beyond the one(s) 
> being reassembled, the likelihood of those datagrams successfully 
> reassembling was epsilon. While the specifics of that may be more 
> "linux-centric" than an RFC should be, I suspect the discussion might 
> provide a starting point for a useful, additional section in the RFC 
> discussing ways transport stacks might minimize the likelihood of 
> Frankengrams.

It would be a good place to start for a *future* draft on the subject. 
We decided it is out of scope for this one.

Thanks,
   -John




From tsvwg-bounces@ietf.org Mon Apr 03 18:24:39 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQXTR-0005CU-9v; Mon, 03 Apr 2006 18:24:37 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQXTP-0005CH-Ow; Mon, 03 Apr 2006 18:24:35 -0400
Received: from palrel10.hp.com ([156.153.255.245])
	by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FQXTP-0000mK-Dj; Mon, 03 Apr 2006 18:24:35 -0400
Received: from tardy.cup.hp.com (tardy.cup.hp.com [15.244.56.217])
	by palrel10.hp.com (Postfix) with ESMTP id D4AF434F91;
	Mon,  3 Apr 2006 15:24:32 -0700 (PDT)
Received: from hp.com (localhost [127.0.0.1])
	by tardy.cup.hp.com (8.9.3 (PHNE_28810)/8.9.3 SMKit7.02) with ESMTP id
	PAA15456; Mon, 3 Apr 2006 15:24:32 -0700 (PDT)
Message-ID: <4431A0A0.4040806@hp.com>
Date: Mon, 03 Apr 2006 15:24:32 -0700
From: Rick Jones <rick.jones2@hp.com>
User-Agent: Mozilla/5.0 (X11; U; HP-UX 9000/785; en-US; rv:1.6) Gecko/20040304
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: John Heffner <jheffner@psc.edu>
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
References: <443181B1.9030107@psc.edu> <44318761.6090500@hp.com>
	<44319D02.1050306@psc.edu>
In-Reply-To: <44319D02.1050306@psc.edu>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 5a9a1bd6c2d06a21d748b7d0070ddcb8
Cc: int-area@ietf.org, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

>> Does it really matter if it is the first fragment?  I would think that 
>> cycling is possible so long as (any of) the lost fragments are simply 
>> not the last fragment when one assumes the fragments are flowing 
>> in-order.
> 
> 
> It could matter.  If it's a middle fragment, you'll receive some 
> duplicate earlier ones before the datagram can be completed.  If the OS 
> has a consistency check here (that the data is the same), it could toss 
> the whole thing when it fails.  I'm not aware of any implementation that 
> does this, but if you're just talking about the first fragment, it's 
> more clear that it's a problem.

Given the overhead such a consistency check would add, I doubt it would 
ever be implemented?  And the conssitency check rather depends on the 
transmission order remaining the same right?

>> Also, to ass-u-me that fragments flow through the network in-order is 
>> probably not a good thing?  Load-balanced link aggregates may send 
>> fragments of one datagram across more than one link in the aggregate, 
>> and I know of at least one NIC, albeit old and likely unused today, 
>> that would send the first fragment last as part of accumulating an 
>> offloaded Checksum (HP-PB FDDI NIC).
> 
> 
> However, data is often delivered in order, so it serves as a good 
> example. :)  Assuming in-order delivery is partially for the sake of 
> clarity.  Any order (reverse, first-packet-last) that's always repeated 
> has the same property.  A randomized delivery order is not as bad as the 
> worst case.

But it is still bad, and I worry (fear, uncertainty and doubt, oh my) 
that someone reading the draft will say "oh, good, my  stack does things 
in a different order so it must not be so bad"

Maybe just some "while we use in-order as an example the problem is 
still present and bad with other packet orderings" text.


>> A while back, over in the "netdev" mailing list, the we were 
>> discussing ways to minimize the risk of frankengrams springing to life 
>> based on ass-u-me-ing :) that if one saw N or more datagrams beyond 
>> the one(s) being reassembled, the likelihood of those datagrams 
>> successfully reassembling was epsilon. While the specifics of that may 
>> be more "linux-centric" than an RFC should be, I suspect the 
>> discussion might provide a starting point for a useful, additional 
>> section in the RFC discussing ways transport stacks might minimize the 
>> likelihood of Frankengrams.
> 
> 
> It would be a good place to start for a *future* draft on the subject. 
> We decided it is out of scope for this one.

Maybe it is drifting a bit, but isn't it rather like announcing a 
vulnerabilty without announcing how to mitigate it?

rick jones




From tsvwg-bounces@ietf.org Mon Apr 03 18:44:57 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQXn5-0003gP-1M; Mon, 03 Apr 2006 18:44:55 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQXn4-0003gH-9P; Mon, 03 Apr 2006 18:44:54 -0400
Received: from mailer1.psc.edu ([128.182.58.100])
	by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FQXn3-0001uy-2K; Mon, 03 Apr 2006 18:44:54 -0400
Received: from [128.182.160.77] (tibook.psc.edu [128.182.160.77])
	(authenticated bits=0)
	by mailer1.psc.edu (8.13.5.20060308/8.13.3) with ESMTP id
	k33MipEW015028
	(version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO);
	Mon, 3 Apr 2006 18:44:51 -0400 (EDT)
Message-ID: <4431A560.2000504@psc.edu>
Date: Mon, 03 Apr 2006 18:44:48 -0400
From: John Heffner <jheffner@psc.edu>
User-Agent: Thunderbird 1.5 (Macintosh/20051201)
MIME-Version: 1.0
To: Rick Jones <rick.jones2@hp.com>
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
References: <443181B1.9030107@psc.edu> <44318761.6090500@hp.com>
	<44319D02.1050306@psc.edu> <4431A0A0.4040806@hp.com>
In-Reply-To: <4431A0A0.4040806@hp.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 7baded97d9887f7a0c7e8a33c2e3ea1b
Cc: int-area@ietf.org, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Rick Jones wrote:
> But it is still bad, and I worry (fear, uncertainty and doubt, oh my) 
> that someone reading the draft will say "oh, good, my  stack does things 
> in a different order so it must not be so bad"
> 
> Maybe just some "while we use in-order as an example the problem is 
> still present and bad with other packet orderings" text.

Will try to make this a little more clear.  Thanks for the input.


>>> A while back, over in the "netdev" mailing list, the we were 
>>> discussing ways to minimize the risk of frankengrams springing to 
>>> life based on ass-u-me-ing :) that if one saw N or more datagrams 
>>> beyond the one(s) being reassembled, the likelihood of those 
>>> datagrams successfully reassembling was epsilon. While the specifics 
>>> of that may be more "linux-centric" than an RFC should be, I suspect 
>>> the discussion might provide a starting point for a useful, 
>>> additional section in the RFC discussing ways transport stacks might 
>>> minimize the likelihood of Frankengrams.
>>
>>
>> It would be a good place to start for a *future* draft on the subject. 
>> We decided it is out of scope for this one.
> 
> Maybe it is drifting a bit, but isn't it rather like announcing a 
> vulnerabilty without announcing how to mitigate it?

I'd say it's more like when a bridge is out, first putting up a warning 
sign, then worrying about how to fix it. ;)

Seriously, we considered writing a section like this.  But on careful 
thought, it was not obvious how to mitigate the problem robustly.  The 
Linux solution works pretty well under specific conditions, but writing 
a BCP on how to safely use fragmentation for high bandwidth apps could 
get thorny.  The safest thing to do is just not use fragmentation for 
these apps, or if you absolutely must, at least add protection (like IPsec).

Thanks,
   -John




From tsvwg-bounces@ietf.org Mon Apr 03 19:12:06 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQYDK-0003LN-3Y; Mon, 03 Apr 2006 19:12:02 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQYDI-0003LF-FR; Mon, 03 Apr 2006 19:12:00 -0400
Received: from palrel10.hp.com ([156.153.255.245])
	by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FQYDH-00033N-78; Mon, 03 Apr 2006 19:12:00 -0400
Received: from tardy.cup.hp.com (tardy.cup.hp.com [15.244.56.217])
	by palrel10.hp.com (Postfix) with ESMTP id A777034B4D;
	Mon,  3 Apr 2006 16:11:58 -0700 (PDT)
Received: from hp.com (localhost [127.0.0.1])
	by tardy.cup.hp.com (8.9.3 (PHNE_28810)/8.9.3 SMKit7.02) with ESMTP id
	QAA15714; Mon, 3 Apr 2006 16:11:58 -0700 (PDT)
Message-ID: <4431ABBD.5000801@hp.com>
Date: Mon, 03 Apr 2006 16:11:57 -0700
From: Rick Jones <rick.jones2@hp.com>
User-Agent: Mozilla/5.0 (X11; U; HP-UX 9000/785; en-US; rv:1.6) Gecko/20040304
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: John Heffner <jheffner@psc.edu>
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
References: <443181B1.9030107@psc.edu>
	<44318761.6090500@hp.com>	<44319D02.1050306@psc.edu>
	<4431A0A0.4040806@hp.com> <4431A560.2000504@psc.edu>
In-Reply-To: <4431A560.2000504@psc.edu>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 0bc60ec82efc80c84b8d02f4b0e4de22
Cc: int-area@ietf.org, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

>> Maybe it is drifting a bit, but isn't it rather like announcing a 
>> vulnerabilty without announcing how to mitigate it?
> 
> 
> I'd say it's more like when a bridge is out, first putting up a warning 
> sign, then worrying about how to fix it. ;)

Well, most poeple already know how to mitigate the problems from a 
bridge being out :)

And since you've basically told the script kiddies "Hey, here is 
something else for you to try..."

> Seriously, we considered writing a section like this.  But on careful 
> thought, it was not obvious how to mitigate the problem robustly.  The 
> Linux solution works pretty well under specific conditions, but writing 
> a BCP on how to safely use fragmentation for high bandwidth apps could 
> get thorny.  The safest thing to do is just not use fragmentation for 
> these apps, or if you absolutely must, at least add protection (like 
> IPsec).

True enough.  But as you point-out in the RFC, there are cases where 
applications are indeed trying to avoid the use of fragmentation, but 
those intermediate devices are hindering it.

Ain't life grand when the fundamental assumptions shift?-)

rick jones




From tsvwg-bounces@ietf.org Mon Apr 03 19:19:06 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQYK6-0007X4-Pu; Mon, 03 Apr 2006 19:19:02 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQYK5-0007Ww-IL; Mon, 03 Apr 2006 19:19:01 -0400
Received: from pop-borzoi.atl.sa.earthlink.net ([207.69.195.70])
	by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FQYK4-0003KN-9m; Mon, 03 Apr 2006 19:19:01 -0400
Received: from h-68-164-82-140.snvacaid.dynamic.covad.net ([68.164.82.140]
	helo=earthlink.net)
	by pop-borzoi.atl.sa.earthlink.net with esmtp (Exim 3.36 #10)
	id 1FQYK2-0001L8-00; Mon, 03 Apr 2006 19:18:58 -0400
Message-ID: <44319974.50A86DBE@earthlink.net>
Date: Mon, 03 Apr 2006 14:53:56 -0700
From: Erblichs <erblichs@earthlink.net>
X-Sender: "Erblichs" <@smtp.earthlink.net> (Unverified)
X-Mailer: Mozilla 4.72 [en]C-gatewaynet  (Win98; I)
X-Accept-Language: en
MIME-Version: 1.0
To: John Heffner <jheffner@psc.edu>
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
References: <443181B1.9030107@psc.edu>
Content-Type: text/plain; charset=iso-8859-2
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.1 (/)
X-Scan-Signature: 92df29fa99cf13e554b84c8374345c17
Cc: int-area@ietf.org, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

John Heffner,

	Interesting..

	A few Short comments.. Possible fixes in #1, #4, #5, and #6.

	1) Does your 26 Mbps flow rate in section 2, almost imply that
	a connection should be multiplexed so that each sub connection
	has a bps rate of 26Mbps to remove this issue in IPv4 if
	fragmentation is an issue?

	2) How does xmit of say a 9k jumbo segments effect this?
	 Wasn't their work into accepting 9k size segments a few
	 years back?

	3) Does the use of say a 3/4 multiplier for the actual transmit
	size decrease the linklyhood of a fragment? Could we have tacked
	an additional header that caused the fragmentation, that this
	would fix?

	4) If a fragment does occur and out-of-order segments are
	seen should a algor like the fast-xmit used be implimented
	to drop the partial set of recv'd fragments. TCP will force the
	entire set to be rexmit'ed, right?

	5) Some datagrams being delivered to a system are comprised
	   of distinct objects which if modified are detected.
	    
          - compressed files.

	   I assume that the distinct items being created is not in the
	   scope of this memo, but decrease the chance of a TCP checksum
	   being accepted at a higher layer.

	   However, if the above is true, would it be wise on some object
	   types to do a non-lossy compression before xmit knowing that the
	   expand would detect any abnormalities????

	6) It sounds like some type of SACK ack that states
	   fragmented segment dropped, redo your path MTU and decrease
	   segment size; is NEEDED.


	Mitchell Erblich
	------------------------

 

	
John Heffner wrote:
> 
> We're moving this doc through the AD-sponsored informational track.
> Though it's not a working group item, we'd like to solicit feedback
> (from both tsvwg and int-area).  Please note that the purpose of this
> document is only to describe a known problem, not a solution.
> 
> Abstract:
>     IPv4 fragmentation is not sufficiently robust for general use in
>     today's Internet.  The 16-bit IP identification field is not large
>     enough to prevent frequent incorrectly assembled IP fragments, and
>     the TCP and UDP checksums are insufficient to prevent the resulting
>     corrupted datagrams from being delivered to higher protocol layers.
>     This note describes some easily reproduced experiments demonstrating
>     the problem, and discusses some of the operational implications of
>     these observations.
> 
> Currently available at:
> http://www.ietf.org/internet-drafts/draft-heffner-frag-harmful-00.txt
> 
> Thanks,
>    -John




From tsvwg-bounces@ietf.org Mon Apr 03 20:49:38 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQZjj-0002ZF-OX; Mon, 03 Apr 2006 20:49:35 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQZji-0002Z7-N7; Mon, 03 Apr 2006 20:49:34 -0400
Received: from mailer1.psc.edu ([128.182.58.100])
	by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FQZjh-0007Yo-EL; Mon, 03 Apr 2006 20:49:34 -0400
Received: from [192.168.1.4] (ipl-67-013.pppoe.stargate.net [206.210.67.13])
	(authenticated bits=0)
	by mailer1.psc.edu (8.13.5.20060308/8.13.3) with ESMTP id
	k340nVc8003749
	(version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO);
	Mon, 3 Apr 2006 20:49:32 -0400 (EDT)
Message-ID: <4431C29B.409@psc.edu>
Date: Mon, 03 Apr 2006 20:49:31 -0400
From: John Heffner <jheffner@psc.edu>
User-Agent: Mozilla Thunderbird 1.0.7 (Macintosh/20050923)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Erblichs <erblichs@earthlink.net>
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
References: <443181B1.9030107@psc.edu> <44319974.50A86DBE@earthlink.net>
In-Reply-To: <44319974.50A86DBE@earthlink.net>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 25620135586de10c627e3628c432b04a
Cc: int-area@ietf.org, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Erblichs wrote:
> John Heffner,
> 
> 	Interesting..
> 
> 	A few Short comments.. Possible fixes in #1, #4, #5, and #6.
> 
> 	1) Does your 26 Mbps flow rate in section 2, almost imply that
> 	a connection should be multiplexed so that each sub connection
> 	has a bps rate of 26Mbps to remove this issue in IPv4 if
> 	fragmentation is an issue?

The problem occurs independently of flows.  It only depends on src,dst 
address pair.


> 	2) How does xmit of say a 9k jumbo segments effect this?
> 	 Wasn't their work into accepting 9k size segments a few
> 	 years back?

Using 9k frames instead of 1500 as in the examples scale the rate at 
which the problem can occur up by a factor of six.


> 	3) Does the use of say a 3/4 multiplier for the actual transmit
> 	size decrease the linklyhood of a fragment? Could we have tacked
> 	an additional header that caused the fragmentation, that this
> 	would fix?

There are many ways fragments can be created, not just from an extra 
tunnel header.  For example, it's natural to want to send 
disk-block-sized datagrams.  If the block size is >PMTU, then you end up 
fragmenting.


> 	4) If a fragment does occur and out-of-order segments are
> 	seen should a algor like the fast-xmit used be implimented
> 	to drop the partial set of recv'd fragments. TCP will force the
> 	entire set to be rexmit'ed, right?

What you're describing is like what Linux does now.  It works pretty 
well under most realistic conditions.  Aside from some theoretical 
corner cases, a couple practical issues:

1) The sender has no way of knowing that the receiver is implementing this.

2) NATs.


> 	5) Some datagrams being delivered to a system are comprised
> 	   of distinct objects which if modified are detected.
> 	    
>           - compressed files.
> 
> 	   I assume that the distinct items being created is not in the
> 	   scope of this memo, but decrease the chance of a TCP checksum
> 	   being accepted at a higher layer.
> 
> 	   However, if the above is true, would it be wise on some object
> 	   types to do a non-lossy compression before xmit knowing that the
> 	   expand would detect any abnormalities????

You could...  Cryptographic hashes would be better yet.


> 	6) It sounds like some type of SACK ack that states
> 	   fragmented segment dropped, redo your path MTU and decrease
> 	   segment size; is NEEDED.

When MTU discovery is working, and the DF bit is not ignored, it should 
be adequate.

Thanks,
   -John




From tsvwg-bounces@ietf.org Mon Apr 03 21:06:30 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQZya-0002TQ-U1; Mon, 03 Apr 2006 21:04:56 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQZyZ-0002TD-3j; Mon, 03 Apr 2006 21:04:55 -0400
Received: from sj-iport-2-in.cisco.com ([171.71.176.71]
	helo=sj-iport-2.cisco.com) by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FQZyW-00085F-Qe; Mon, 03 Apr 2006 21:04:55 -0400
Received: from sj-core-5.cisco.com ([171.71.177.238])
	by sj-iport-2.cisco.com with ESMTP; 03 Apr 2006 18:04:52 -0700
X-IronPort-AV: i="4.03,160,1141632000"; 
	d="scan'208"; a="318766646:sNHT32119420"
Received: from imail.cisco.com (sjc12-sbr-sw3-3f5.cisco.com [172.19.96.182])
	by sj-core-5.cisco.com (8.12.10/8.12.6) with ESMTP id k3414p7T013947;
	Mon, 3 Apr 2006 18:04:51 -0700 (PDT)
Received: from [10.32.244.220] (stealth-10-32-244-220.cisco.com
	[10.32.244.220])
	by imail.cisco.com (8.12.11/8.12.10) with ESMTP id k34153ss013203;
	Mon, 3 Apr 2006 18:05:03 -0700
In-Reply-To: <4431ABBD.5000801@hp.com>
References: <443181B1.9030107@psc.edu>
	<44318761.6090500@hp.com>	<44319D02.1050306@psc.edu>
	<4431A0A0.4040806@hp.com> <4431A560.2000504@psc.edu>
	<4431ABBD.5000801@hp.com>
Mime-Version: 1.0 (Apple Message framework v749.3)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <3A1F3E27-462B-40A7-8E0B-EE997C1219A2@cisco.com>
Content-Transfer-Encoding: 7bit
From: Fred Baker <fred@cisco.com>
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
Date: Mon, 3 Apr 2006 18:04:50 -0700
To: Rick Jones <rick.jones2@hp.com>
X-Mailer: Apple Mail (2.749.3)
DKIM-Signature: a=rsa-sha1; q=dns; l=464; t=1144112703; x=1144976703;
	c=relaxed/simple; s=oregon;
	h=To:Content-Type:From:Subject:Content-Transfer-Encoding:MIME-Version;
	d=cisco.com; i=fred@cisco.com; z=From:Fred=20Baker=20<fred@cisco.com>
	|Subject:Re=3A=20[Tsvwg]=20Seeking=20comments=20on=20draft-heffner-frag-harmful-0
	0 |To:Rick=20Jones=20<rick.jones2@hp.com>;
	X=v=3Dcisco.com=3B=20h=3DmxKEPPrY+KDvfQYwsESJPrmGvec=3D;
	b=UZW/Rn6O6mQoSagprCcdvbdlKQJd4bwcg4NuVm+JoiTyJNCbfhrhs5rdw+f9FKRawk2jldtK
	+nMpZJ0xSCzYIWO+hfN3Ux0RlauvvCWvUvk5llql01HtJ4na2Ql6o1hN;
Authentication-Results: imail.cisco.com; header.From=fred@cisco.com;
	dkim=pass ( sig from cisco.com verified; ); 
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 7a6398bf8aaeabc7a7bb696b6b0a2aad
Cc: int-area@ietf.org, tsvwg@ietf.org, John Heffner <jheffner@psc.edu>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org


On Apr 3, 2006, at 4:11 PM, Rick Jones wrote:

> And since you've basically told the script kiddies "Hey, here is  
> something else for you to try..."

I could be completely misinformed, but I thought fragmentation  
attacks weren't news? Saying there are issues in an RFC isn't  
anywhere near as bad as naming the vendor in the discussion of the  
vulnerability...

http://www-src.lip6.fr/homepages/Fabrice.Legond-Aubry/www.ouah.org/ 
fragma.html




From tsvwg-bounces@ietf.org Mon Apr 03 21:12:56 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQa6D-0000Gw-K7; Mon, 03 Apr 2006 21:12:49 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQa6B-0000Gl-DO; Mon, 03 Apr 2006 21:12:47 -0400
Received: from palrel11.hp.com ([156.153.255.246])
	by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FQa6A-0008QI-5s; Mon, 03 Apr 2006 21:12:47 -0400
Received: from tardy.cup.hp.com (tardy.cup.hp.com [15.244.56.217])
	by palrel11.hp.com (Postfix) with ESMTP id 7532634D85;
	Mon,  3 Apr 2006 18:12:45 -0700 (PDT)
Received: from hp.com (localhost [127.0.0.1])
	by tardy.cup.hp.com (8.9.3 (PHNE_28810)/8.9.3 SMKit7.02) with ESMTP id
	SAA16465; Mon, 3 Apr 2006 18:12:44 -0700 (PDT)
Message-ID: <4431C80C.7090408@hp.com>
Date: Mon, 03 Apr 2006 18:12:44 -0700
From: Rick Jones <rick.jones2@hp.com>
User-Agent: Mozilla/5.0 (X11; U; HP-UX 9000/785; en-US; rv:1.6) Gecko/20040304
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Erblichs <erblichs@earthlink.net>
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
References: <443181B1.9030107@psc.edu> <44319974.50A86DBE@earthlink.net>
In-Reply-To: <44319974.50A86DBE@earthlink.net>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 08e48e05374109708c00c6208b534009
Cc: int-area@ietf.org, tsvwg@ietf.org, John Heffner <jheffner@psc.edu>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

> 	2) How does xmit of say a 9k jumbo segments effect this?
> 	 Wasn't their work into accepting 9k size segments a few
> 	 years back?

While the IETF may have said something wrt "jumbo frames" I don't think 
the IEEE has sanctioned them.  I'm not sure if they ever will.

rick jones




From tsvwg-bounces@ietf.org Mon Apr 03 21:14:01 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQa7M-0000SQ-6M; Mon, 03 Apr 2006 21:14:00 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQa7K-0000SI-Ll; Mon, 03 Apr 2006 21:13:58 -0400
Received: from palrel11.hp.com ([156.153.255.246])
	by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FQa7K-0008Re-DX; Mon, 03 Apr 2006 21:13:58 -0400
Received: from tardy.cup.hp.com (tardy.cup.hp.com [15.244.56.217])
	by palrel11.hp.com (Postfix) with ESMTP id E7066347FA;
	Mon,  3 Apr 2006 18:13:57 -0700 (PDT)
Received: from hp.com (localhost [127.0.0.1])
	by tardy.cup.hp.com (8.9.3 (PHNE_28810)/8.9.3 SMKit7.02) with ESMTP id
	SAA16510; Mon, 3 Apr 2006 18:13:57 -0700 (PDT)
Message-ID: <4431C855.3090103@hp.com>
Date: Mon, 03 Apr 2006 18:13:57 -0700
From: Rick Jones <rick.jones2@hp.com>
User-Agent: Mozilla/5.0 (X11; U; HP-UX 9000/785; en-US; rv:1.6) Gecko/20040304
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Fred Baker <fred@cisco.com>
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
References: <443181B1.9030107@psc.edu>	<44318761.6090500@hp.com>	<44319D02.1050306@psc.edu>	<4431A0A0.4040806@hp.com>
	<4431A560.2000504@psc.edu>	<4431ABBD.5000801@hp.com>
	<3A1F3E27-462B-40A7-8E0B-EE997C1219A2@cisco.com>
In-Reply-To: <3A1F3E27-462B-40A7-8E0B-EE997C1219A2@cisco.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: d6b246023072368de71562c0ab503126
Cc: int-area@ietf.org, tsvwg@ietf.org, John Heffner <jheffner@psc.edu>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Fred Baker wrote:
> 
> On Apr 3, 2006, at 4:11 PM, Rick Jones wrote:
> 
>> And since you've basically told the script kiddies "Hey, here is  
>> something else for you to try..."
> 
> 
> I could be completely misinformed, but I thought fragmentation  attacks 
> weren't news? Saying there are issues in an RFC isn't  anywhere near as 
> bad as naming the vendor in the discussion of the  vulnerability...
> 
> http://www-src.lip6.fr/homepages/Fabrice.Legond-Aubry/www.ouah.org/ 
> fragma.html

I suspect they are no more or less news than what the RFC is describing 
to begin with since the issue goes back at least 15 years if not more :)

rick jones




From tsvwg-bounces@ietf.org Mon Apr 03 21:42:16 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQaYc-0000ez-Vv; Mon, 03 Apr 2006 21:42:10 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FQaYb-0000ct-DJ
	for tsvwg@ietf.org; Mon, 03 Apr 2006 21:42:09 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FQaYa-0001Js-3u
	for tsvwg@ietf.org; Mon, 03 Apr 2006 21:42:09 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FQaYb-0000S3-7Z; Mon, 03 Apr 2006 19:41:47 -0600
Message-ID: <4431CEF6.5000406@middle.net>
Date: Mon, 03 Apr 2006 19:42:14 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: John Heffner <jheffner@psc.edu>
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
References: <443181B1.9030107@psc.edu>
In-Reply-To: <443181B1.9030107@psc.edu>
Content-Type: multipart/alternative;
	boundary="------------000007020701040308040805"
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.1 (/)
X-Scan-Signature: 8b431ad66d60be2d47c7bfeb879db82c
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

This is a multi-part message in MIME format.
--------------000007020701040308040805
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

The draft states:

In the case of SSH, any stream corruption results in immediate termination of the connection.

While RFC4253 certainly allows this behavior, I don't believe it 
requires it.  Terminating a connection after a single MAC error is a 
little excessive considering the transports used.

As far as possible solutions go, wouldn't it make sense to have 
fragmenting devices (to the degree they are necessary at all) wrap 
fragments in a new header with a 32 bit identification field and a 32 
bit CRC that covers the contents of the entire datagram?

 - Mark B.


--------------000007020701040308040805
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
  <title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
The draft states:<br>
<blockquote>
  <pre>In the case of SSH, any stream corruption results in immediate termination of the connection.</pre>
</blockquote>
While RFC4253 certainly allows this behavior, I don't believe it
requires it.&nbsp; Terminating a connection after a single MAC error is a
little excessive considering the transports used.<br>
<br>
As far as possible solutions go, wouldn't it make sense to have
fragmenting devices (to the degree they are necessary at all) wrap
fragments in a new header with a 32 bit identification field and a 32
bit CRC that covers the contents of the entire datagram?<br>
<br>
&nbsp;- Mark B.<br>
<br>
</body>
</html>

--------------000007020701040308040805--




From tsvwg-bounces@ietf.org Mon Apr 03 22:35:15 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQbNq-0005vK-Ad; Mon, 03 Apr 2006 22:35:06 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FQbNo-0005mG-L6
	for tsvwg@ietf.org; Mon, 03 Apr 2006 22:35:04 -0400
Received: from mailer1.psc.edu ([128.182.58.100])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FQbNn-0004K1-Dk
	for tsvwg@ietf.org; Mon, 03 Apr 2006 22:35:04 -0400
Received: from [192.168.1.4] (ipl-67-013.pppoe.stargate.net [206.210.67.13])
	(authenticated bits=0)
	by mailer1.psc.edu (8.13.5.20060308/8.13.3) with ESMTP id
	k342Z2mQ019626
	(version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO);
	Mon, 3 Apr 2006 22:35:03 -0400 (EDT)
Message-ID: <4431DB56.9080709@psc.edu>
Date: Mon, 03 Apr 2006 22:35:02 -0400
From: John Heffner <jheffner@psc.edu>
User-Agent: Mozilla Thunderbird 1.0.7 (Macintosh/20050923)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Mark Butler <butlerm@middle.net>
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
References: <443181B1.9030107@psc.edu> <4431CEF6.5000406@middle.net>
In-Reply-To: <4431CEF6.5000406@middle.net>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: ea4ac80f790299f943f0a53be7e1a21a
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark Butler wrote:
> The draft states:
> 
> In the case of SSH, any stream corruption results in immediate termination of the connection.
> 
> While RFC4253 certainly allows this behavior, I don't believe it 
> requires it.  Terminating a connection after a single MAC error is a 
> little excessive considering the transports used.

I actually got the citation wrong.  From RFC4251, Section 9.3.5:

    This protocol is designed to be used over a reliable transport.  If
    transmission errors or message manipulation occur, the connection is
    closed.  The connection SHOULD be re-established if this occurs.

OpenSSH terminates with extreme prejudice.  (I expreienced this myself 
when I had a NIC with a bad DMA engine that did TCP checksum 
offloading.)  I'm not aware of any mechanism in the SSH protocol that 
could repair a corrupted connection.


> As far as possible solutions go, wouldn't it make sense to have 
> fragmenting devices (to the degree they are necessary at all) wrap 
> fragments in a new header with a 32 bit identification field and a 32 
> bit CRC that covers the contents of the entire datagram?

You mean SCTP over IPv6? ;)

Thanks,
   -John




From tsvwg-bounces@ietf.org Tue Apr 04 02:20:39 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQety-0001gd-5a; Tue, 04 Apr 2006 02:20:30 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FQetx-0001gY-0a
	for tsvwg@ietf.org; Tue, 04 Apr 2006 02:20:29 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FQetu-0004UC-La
	for tsvwg@ietf.org; Tue, 04 Apr 2006 02:20:28 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FQetr-0001QH-9D; Tue, 04 Apr 2006 00:20:01 -0600
Message-ID: <4432102D.9000801@middle.net>
Date: Tue, 04 Apr 2006 00:20:29 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: John Heffner <jheffner@psc.edu>
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
References: <443181B1.9030107@psc.edu> <4431CEF6.5000406@middle.net>
	<4431DB56.9080709@psc.edu>
In-Reply-To: <4431DB56.9080709@psc.edu>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 8b431ad66d60be2d47c7bfeb879db82c
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

John Heffner wrote:

> Mark Butler wrote:
>
>> The draft states:
>>
>> In the case of SSH, any stream corruption results in immediate 
>> termination of the connection.
>>
>> While RFC4253 certainly allows this behavior, I don't believe it 
>> requires it.  Terminating a connection after a single MAC error is a 
>> little excessive considering the transports used.
>
>
> I actually got the citation wrong.  From RFC4251, Section 9.3.5:
>
>    This protocol is designed to be used over a reliable transport.  If
>    transmission errors or message manipulation occur, the connection is
>    closed.  The connection SHOULD be re-established if this occurs.

Thanks for the reference.  Hypothetically speaking, SSH could be run 
over a transport that allows ULPs to veto messages before they are 
acked, allowing the connection to be preserved. 

>
> OpenSSH terminates with extreme prejudice.  (I expreienced this myself 
> when I had a NIC with a bad DMA engine that did TCP checksum 
> offloading.)  I'm not aware of any mechanism in the SSH protocol that 
> could repair a corrupted connection.
>
>
>> As far as possible solutions go, wouldn't it make sense to have 
>> fragmenting devices (to the degree they are necessary at all) wrap 
>> fragments in a new header with a 32 bit identification field and a 32 
>> bit CRC that covers the contents of the entire datagram?
>
>
> You mean SCTP over IPv6? ;)

Sort of. I was thinking more of the precedent set by IPSEC over IPv4.  
e.g. taking the IPv6 fragment header, assigning a new protocol number, 
adding 32 bits for the CRC, and inserting it after the IPv4 header.  
Support would be negotiated at the transport layer and signalled to 
fragmenting devices via the Reserved bit next to the DF bit in the IPv4 
header.

- Mark B.

>
>





From tsvwg-bounces@ietf.org Tue Apr 04 10:52:16 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQmtA-0000fD-PL; Tue, 04 Apr 2006 10:52:12 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FQmt9-0000f7-0Z
	for tsvwg@ietf.org; Tue, 04 Apr 2006 10:52:11 -0400
Received: from test-iport-1.cisco.com ([171.71.176.117])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FQmt7-0001Ng-EE
	for tsvwg@ietf.org; Tue, 04 Apr 2006 10:52:10 -0400
Received: from sj-core-5.cisco.com ([171.71.177.238])
	by test-iport-1.cisco.com with ESMTP; 04 Apr 2006 07:52:08 -0700
Received: from imail.cisco.com (sjc12-sbr-sw3-3f5.cisco.com [172.19.96.182])
	by sj-core-5.cisco.com (8.12.10/8.12.6) with ESMTP id k34Eq87T000967;
	Tue, 4 Apr 2006 07:52:08 -0700 (PDT)
Received: from [10.32.244.220] (stealth-10-32-244-220.cisco.com
	[10.32.244.220])
	by imail.cisco.com (8.12.11/8.12.10) with ESMTP id k34EqHRK002198;
	Tue, 4 Apr 2006 07:52:17 -0700
In-Reply-To: <4431DB56.9080709@psc.edu>
References: <443181B1.9030107@psc.edu> <4431CEF6.5000406@middle.net>
	<4431DB56.9080709@psc.edu>
Mime-Version: 1.0 (Apple Message framework v749.3)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <D953ADF8-7947-41ED-ABCA-48D912A3F44D@cisco.com>
Content-Transfer-Encoding: 7bit
From: Fred Baker <fred@cisco.com>
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
Date: Tue, 4 Apr 2006 07:52:06 -0700
To: John Heffner <jheffner@psc.edu>, Mark Butler <butlerm@middle.net>
X-Mailer: Apple Mail (2.749.3)
DKIM-Signature: a=rsa-sha1; q=dns; l=1542; t=1144162338; x=1145026338;
	c=relaxed/simple; s=oregon;
	h=To:Content-Type:From:Subject:Content-Transfer-Encoding:MIME-Version;
	d=cisco.com; i=fred@cisco.com; z=From:Fred=20Baker=20<fred@cisco.com>
	|Subject:Re=3A=20[Tsvwg]=20Seeking=20comments=20on=20draft-heffner-frag-harmful-0
	0 |To:John=20Heffner=20<jheffner@psc.edu>,
	=0A=20Mark=20Butler=20<butlerm@middl e.net>;
	X=v=3Dcisco.com=3B=20h=3DUlJiqnzIprPV/Nt7WS97zrG3xl4=3D;
	b=m0GDbPJBcjuPDs8ePcN+Xu9GGdCk/enpqz9yM/TQqqP9SlSuQ3HP+BaFhln/0xZPUXLNMt+g
	7f676ntfZb5Ln4KLJ15Y2zn23H1rwK08X41D53/5o5iE+ImFWKLxhsXJ;
Authentication-Results: imail.cisco.com; header.From=fred@cisco.com;
	dkim=pass ( sig from cisco.com verified; ); 
X-Spam-Score: 0.0 (/)
X-Scan-Signature: bb8f917bb6b8da28fc948aeffb74aa17
Cc: TSVWG <tsvwg@ietf.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org


On Apr 3, 2006, at 6:42 PM, Mark Butler wrote:
> The draft states:
> In the case of SSH, any stream corruption results in immediate  
> termination of the connection.While RFC4253 certainly allows this  
> behavior, I don't believe it requires it.  Terminating a connection  
> after a single MAC error is a little excessive considering the  
> transports used.
>
> As far as possible solutions go, wouldn't it make sense to have  
> fragmenting devices (to the degree they are necessary at all) wrap  
> fragments in a new header with a 32 bit identification field and a  
> 32 bit CRC that covers the contents of the entire datagram?
>
>  - Mark B.

On Apr 3, 2006, at 7:35 PM, John Heffner wrote:
> I actually got the citation wrong.  From RFC4251, Section 9.3.5:
>
>    This protocol is designed to be used over a reliable transport.  If
>    transmission errors or message manipulation occur, the  
> connection is
>    closed.  The connection SHOULD be re-established if this occurs.

This suggests, at least to me, that the data stream being looked at  
by SSH is the TCP data stream, not the MAC data stream. The MAC - and  
IP - remain best effort, which is to say that they can drop,  
duplicate, or reorder anything they pleaase. It is the output of the  
*transport*, which is above IP, that is deemed reliable and  
unreliability in which would close a connection with prejudice.

Right?

Tell me we haven't changed the definition of the word "transport" to  
put it somewhere at or below IP?




From tsvwg-bounces@ietf.org Tue Apr 04 11:07:16 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FQn7b-00017K-P6; Tue, 04 Apr 2006 11:07:07 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FQn7a-00016l-NF
	for tsvwg@ietf.org; Tue, 04 Apr 2006 11:07:06 -0400
Received: from mailer1.psc.edu ([128.182.58.100])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FQn7Z-00022q-GK
	for tsvwg@ietf.org; Tue, 04 Apr 2006 11:07:06 -0400
Received: from [128.182.160.77] (tibook.psc.edu [128.182.160.77])
	(authenticated bits=0)
	by mailer1.psc.edu (8.13.5.20060308/8.13.3) with ESMTP id
	k34F6xpR024097
	(version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO);
	Tue, 4 Apr 2006 11:06:59 -0400 (EDT)
Message-ID: <44328B91.6030208@psc.edu>
Date: Tue, 04 Apr 2006 11:06:57 -0400
From: John Heffner <jheffner@psc.edu>
User-Agent: Thunderbird 1.5 (Macintosh/20051201)
MIME-Version: 1.0
To: Fred Baker <fred@cisco.com>
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
References: <443181B1.9030107@psc.edu> <4431CEF6.5000406@middle.net>
	<4431DB56.9080709@psc.edu>
	<D953ADF8-7947-41ED-ABCA-48D912A3F44D@cisco.com>
In-Reply-To: <D953ADF8-7947-41ED-ABCA-48D912A3F44D@cisco.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 7a6398bf8aaeabc7a7bb696b6b0a2aad
Cc: TSVWG <tsvwg@ietf.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Fred Baker wrote:
> This suggests, at least to me, that the data stream being looked at by 
> SSH is the TCP data stream, not the MAC data stream. The MAC - and IP - 
> remain best effort, which is to say that they can drop, duplicate, or 
> reorder anything they pleaase. It is the output of the *transport*, 
> which is above IP, that is deemed reliable and unreliability in which 
> would close a connection with prejudice.

There's a bit of a terminology overload here. :)  MAC in this context is 
Message Authentication Codes, which SSH uses to ensure data integrity 
(above the reliable transport).

   -John




From tsvwg-bounces@ietf.org Wed Apr 05 20:01:56 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRHwf-0003VB-2P; Wed, 05 Apr 2006 20:01:53 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRHwd-0003V6-4Z
	for tsvwg@ietf.org; Wed, 05 Apr 2006 20:01:51 -0400
Received: from sj-iport-5.cisco.com ([171.68.10.87])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRHwb-0003fI-SX
	for tsvwg@ietf.org; Wed, 05 Apr 2006 20:01:51 -0400
Received: from sj-core-3.cisco.com ([171.68.223.137])
	by sj-iport-5.cisco.com with ESMTP; 05 Apr 2006 17:01:51 -0700
X-IronPort-AV: i="4.04,91,1144047600"; 
	d="scan'208"; a="266891813:sNHT35323848"
Received: from xbh-sjc-211.amer.cisco.com (xbh-sjc-211.cisco.com
	[171.70.151.144])
	by sj-core-3.cisco.com (8.12.10/8.12.6) with ESMTP id k3601m1l024778;
	Wed, 5 Apr 2006 17:01:49 -0700 (PDT)
Received: from xfe-sjc-211.amer.cisco.com ([171.70.151.174]) by
	xbh-sjc-211.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Wed, 5 Apr 2006 17:01:48 -0700
Received: from jmpolk-wxp.cisco.com ([10.89.20.13]) by
	xfe-sjc-211.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Wed, 5 Apr 2006 17:01:47 -0700
Message-Id: <4.3.2.7.2.20060405185529.0296e788@email.cisco.com>
X-Sender: jmpolk@email.cisco.com
X-Mailer: QUALCOMM Windows Eudora Version 4.3.2
Date: Wed, 05 Apr 2006 19:01:46 -0500
To: tsvwg <tsvwg@ietf.org>
From: "James M. Polk" <jmpolk@cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
X-OriginalArrivalTime: 06 Apr 2006 00:01:48.0162 (UTC)
	FILETIME=[4C9FB620:01C6590D]
X-Spam-Score: 1.1 (+)
X-Scan-Signature: 30ac594df0e66ffa5a93eb4c48bcb014
Cc: Lars Eggert <lars.eggert@netlab.nec.de>, magnus.westerlund@ericsson.com
Subject: [Tsvwg] WGLC for draft-ietf-tsvwg-vpn-signaled-preemption-00
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

TSVWG

Due to an oversight on my part, I did not send out a official notice about 
the WGLC for this document when I announced it in/during the TSVWG session 
in Dallas.

I will therefore let the LC go a week longer, until April 12th, 2006.

To the authors of this ID, I apologize for the delay.




From tsvwg-bounces@ietf.org Wed Apr 05 21:18:57 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRJ97-0000II-AJ; Wed, 05 Apr 2006 21:18:49 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRJ96-0000ID-1V
	for tsvwg@ietf.org; Wed, 05 Apr 2006 21:18:48 -0400
Received: from test-iport-3.cisco.com ([171.71.176.78])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRJ94-0007at-Pc
	for tsvwg@ietf.org; Wed, 05 Apr 2006 21:18:48 -0400
Received: from sj-core-2.cisco.com ([171.71.177.254])
	by test-iport-3.cisco.com with ESMTP; 05 Apr 2006 18:18:46 -0700
Received: from xbh-sjc-231.amer.cisco.com (xbh-sjc-231.cisco.com
	[128.107.191.100])
	by sj-core-2.cisco.com (8.12.10/8.12.6) with ESMTP id k361IkGv002524;
	Wed, 5 Apr 2006 18:18:46 -0700 (PDT)
Received: from xfe-sjc-212.amer.cisco.com ([171.70.151.187]) by
	xbh-sjc-231.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Wed, 5 Apr 2006 18:18:45 -0700
Received: from jmpolk-wxp.cisco.com ([10.89.20.13]) by
	xfe-sjc-212.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Wed, 5 Apr 2006 18:18:45 -0700
Message-Id: <4.3.2.7.2.20060405201329.02964e18@email.cisco.com>
X-Sender: jmpolk@email.cisco.com
X-Mailer: QUALCOMM Windows Eudora Version 4.3.2
Date: Wed, 05 Apr 2006 20:18:43 -0500
To: tsvwg <tsvwg@ietf.org>
From: "James M. Polk" <jmpolk@cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
X-OriginalArrivalTime: 06 Apr 2006 01:18:45.0575 (UTC)
	FILETIME=[0CD10570:01C65918]
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 856eb5f76e7a34990d1d457d8e8e5b7f
Cc: Lars Eggert <lars.eggert@netlab.nec.de>, magnus.westerlund@ericsson.com,
	jmpolk@cisco.com
Subject: [Tsvwg] WGLC for draft-ietf-tsvwg-rsvp-ipsec-00.txt NOW
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

TSVWG Group

As agreed to in Dallas during the TSVWG meeting, the WGLC for

	draft-ietf-tsvwg-rsvp-ipsec-00.txt

starts NOW.

All comments (nits, errors, suggestions, comments, approval) regarding this 
ID should be directed to the TSVWG list.

WGLC will end on or about April 20th, 2006, depending on the level of 
discussion at that time.

James
Lars
Magnus
	IETF TSVWG co-chairs




From tsvwg-bounces@ietf.org Fri Apr 07 05:18:33 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRn6p-0004GU-OR; Fri, 07 Apr 2006 05:18:27 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRn6o-0004GP-Et
	for tsvwg@ietf.org; Fri, 07 Apr 2006 05:18:26 -0400
Received: from hostreea1.alcatel.com ([143.209.238.161]
	helo=audl951.usa.alcatel.com)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRn6n-0005ZO-5d
	for tsvwg@ietf.org; Fri, 07 Apr 2006 05:18:26 -0400
Received: from axes-mach01.usa.alcatel.com (axes-mach01.usa.alcatel.com
	[10.255.0.20])
	by audl951.usa.alcatel.com (ALCANET) with ESMTP id k379IMo8012639
	for <tsvwg@ietf.org>; Fri, 7 Apr 2006 04:18:23 -0500
Received: from [10.255.1.166] (ax-sp-pc114.usa.alcatel.com [10.255.1.166])
	by axes-mach01.usa.alcatel.com (8.12.10+Sun/8.12.2) with ESMTP id
	k379LttL006459
	for <tsvwg@ietf.org>; Fri, 7 Apr 2006 14:51:56 +0530 (IST)
Message-ID: <44362D55.1040806@alcatel.com>
Date: Fri, 07 Apr 2006 14:43:57 +0530
From: devayya <devayya.bopaiah@alcatel.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: tsvwg <tsvwg@ietf.org>
Content-Type: multipart/mixed; boundary="------------040804090008040100090808"
X-Scanned-By: MIMEDefang 2.51 on 143.209.238.34
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 92df29fa99cf13e554b84c8374345c17
Subject: [Tsvwg] SCTP multi-homing association
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

This is a multi-part message in MIME format.
--------------040804090008040100090808
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Hi,

I have a question on SCTP multi-homing,

My SCTP multi-homing test set up is as follows,
There are two etherenet ports with different ip address are configured 
for one endpoint(say endpont A). and similarly I have another endpoint 
(say endpoint B) with two ethernet ports with different ip addresses. 
The association between two endpoints is ESTABLISHED.

Using a test tool I transmit traffic between these two endpoints.
With less amount of traffic flowing, if one of the ethernet port IP 
address is lost, then the association between these two endpoints remain 
ACTIVE and the traffic flows on the second ethernet port.(As per 
multi-homing  requirement)
With increase in traffic flowing, if one of the ethernet port IP address 
is lost, then the association between these two endpoints goes to 
INACTIVE. and there is traffic loss. Although the SCTP association is 
still present, the ULP association goes down, with the reason "delay ACK 
timer expiry". Can anybody help find out why?

As per SCTP implemtors guide, there is mention of  updation of cwnd when 
user tries to transmit DATA, cwnd should be modified depending on the 
flightsize and a newly introduced MAX.BURST.(not present in RFC)
               if((flightsize + Max.Burst*MTU) < cwnd)
                    cwnd = flightsize +  Max.Burst*MTU

Is it really needed to implement this MAX.BURST?
Will addition of this MAX.BURST solve the earlier problem mentioned by me?

Thanks and regards,
devayya

P.s: For your reference I have attached a text file of the set up.

--------------040804090008040100090808
Content-Type: text/plain;
 name="sctp_multihm.txt"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="sctp_multihm.txt"




    ----------------                           -----------------
    |               |                          |                | 
    |               |P| <==================> |R|                | 
    |Endpoint A     |                          |  Endpoint  B   |
    |               |Q|                      |S|                |
    |               |                          |                |
    ----------------                           ------------------
Traffic is flowing between Endpoint A and endpoint B, using enthernet ports P and R



    ----------------                           -----------------
    |               |                          |                | 
    |               |P|                      |R|                | 
    |Endpoint A     |                          |  Endpoint  B   |
    |               |Q| <==================> |S|                |
    |               |                          |                |
    ----------------                           ------------------
The enternet port P is lost and now the traffic flows using ethernet port Q and S.

--------------040804090008040100090808--




From tsvwg-bounces@ietf.org Fri Apr 07 07:28:53 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRp8x-0005lO-4f; Fri, 07 Apr 2006 07:28:47 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRp8v-0005dF-AQ
	for tsvwg@ietf.org; Fri, 07 Apr 2006 07:28:45 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRp8t-0000OF-SN
	for tsvwg@ietf.org; Fri, 07 Apr 2006 07:28:45 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:Rnizw9xpXXeK12ziEYL+98yB+BD0QwNw@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k37BSgq04448;
	Fri, 7 Apr 2006 05:28:42 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k37BScX04240;
	Fri, 7 Apr 2006 05:28:38 -0600
Date: Fri, 7 Apr 2006 05:28:36 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: devayya <devayya.bopaiah@alcatel.com>
Subject: Re: [Tsvwg] SCTP multi-homing association
Message-ID: <20060407052836.A4198@openss7.org>
Mail-Followup-To: devayya <devayya.bopaiah@alcatel.com>, tsvwg <tsvwg@ietf.org>
References: <44362D55.1040806@alcatel.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <44362D55.1040806@alcatel.com>;
	from devayya.bopaiah@alcatel.com on Fri, Apr 07, 2006 at
	02:43:57PM +0530
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 4b800b1eab964a31702fa68f1ff0e955
Cc: tsvwg <tsvwg@ietf.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

devayya,

Does R have a route to Q?

--brian

devayya wrote:                                                     (Fri, 07 Apr 2006 14:43:57)
> Hi,
> 
> I have a question on SCTP multi-homing,
> 
> My SCTP multi-homing test set up is as follows,
> There are two etherenet ports with different ip address are configured 
> for one endpoint(say endpont A). and similarly I have another endpoint 
> (say endpoint B) with two ethernet ports with different ip addresses. 
> The association between two endpoints is ESTABLISHED.
> 
> Using a test tool I transmit traffic between these two endpoints.
> With less amount of traffic flowing, if one of the ethernet port IP 
> address is lost, then the association between these two endpoints remain 
> ACTIVE and the traffic flows on the second ethernet port.(As per 
> multi-homing  requirement)
> With increase in traffic flowing, if one of the ethernet port IP address 
> is lost, then the association between these two endpoints goes to 
> INACTIVE. and there is traffic loss. Although the SCTP association is 
> still present, the ULP association goes down, with the reason "delay ACK 
> timer expiry". Can anybody help find out why?
> 
> As per SCTP implemtors guide, there is mention of  updation of cwnd when 
> user tries to transmit DATA, cwnd should be modified depending on the 
> flightsize and a newly introduced MAX.BURST.(not present in RFC)
>                if((flightsize + Max.Burst*MTU) < cwnd)
>                     cwnd = flightsize +  Max.Burst*MTU
> 
> Is it really needed to implement this MAX.BURST?
> Will addition of this MAX.BURST solve the earlier problem mentioned by me?
> 
> Thanks and regards,
> devayya
> 
> P.s: For your reference I have attached a text file of the set up.

> 
> 
> 
>     ----------------                           -----------------
>     |               |                          |                | 
>     |               |P| <==================> |R|                | 
>     |Endpoint A     |                          |  Endpoint  B   |
>     |               |Q|                      |S|                |
>     |               |                          |                |
>     ----------------                           ------------------
> Traffic is flowing between Endpoint A and endpoint B, using enthernet ports P and R
> 
> 
> 
>     ----------------                           -----------------
>     |               |                          |                | 
>     |               |P|                      |R|                | 
>     |Endpoint A     |                          |  Endpoint  B   |
>     |               |Q| <==================> |S|                |
>     |               |                          |                |
>     ----------------                           ------------------
> The enternet port P is lost and now the traffic flows using ethernet port Q and S.


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




From tsvwg-bounces@ietf.org Fri Apr 07 07:33:46 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRpDk-0007oA-8m; Fri, 07 Apr 2006 07:33:44 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRpDj-0007o5-Bh
	for tsvwg@ietf.org; Fri, 07 Apr 2006 07:33:43 -0400
Received: from hostreea2.alcatel.com ([143.209.238.162]
	helo=audl953.usa.alcatel.com)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRpDi-0000U6-UO
	for tsvwg@ietf.org; Fri, 07 Apr 2006 07:33:43 -0400
Received: from axes-mach01.usa.alcatel.com (axes-mach01.usa.alcatel.com
	[10.255.0.20])
	by audl953.usa.alcatel.com (ALCANET) with ESMTP id k37BXejt009355;
	Fri, 7 Apr 2006 06:33:41 -0500
Received: from [10.255.1.166] (ax-sp-pc114.usa.alcatel.com [10.255.1.166])
	by axes-mach01.usa.alcatel.com (8.12.10+Sun/8.12.2) with ESMTP id
	k37BbDtL019093; Fri, 7 Apr 2006 17:07:15 +0530 (IST)
Message-ID: <44364D0B.3030107@alcatel.com>
Date: Fri, 07 Apr 2006 16:59:15 +0530
From: devayya <devayya.bopaiah@alcatel.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: bidulock@openss7.org
Subject: Re: [Tsvwg] SCTP multi-homing association
References: <44362D55.1040806@alcatel.com> <20060407052836.A4198@openss7.org>
In-Reply-To: <20060407052836.A4198@openss7.org>
Content-Type: multipart/alternative;
	boundary="------------050004000908090107090404"
X-Scanned-By: MIMEDefang 2.51 on 143.209.238.34
X-Spam-Score: 0.1 (/)
X-Scan-Signature: 93df555cbdbcdae9621e5b95d44b301e
Cc: tsvwg <tsvwg@ietf.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

This is a multi-part message in MIME format.
--------------050004000908090107090404
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

There is route from P to R and Q to S. I have configured this using the 
system command
"route add -net x.x.x -iface y.y.y.y"
However I could not configure route from R to Q or S to P. Can you guide me?

thanks,
devayya

Brian F. G. Bidulock wrote:

>devayya,
>
>Does R have a route to Q?
>
>--brian
>
>devayya wrote:                                                     (Fri, 07 Apr 2006 14:43:57)
>  
>
>>Hi,
>>
>>I have a question on SCTP multi-homing,
>>
>>My SCTP multi-homing test set up is as follows,
>>There are two etherenet ports with different ip address are configured 
>>for one endpoint(say endpont A). and similarly I have another endpoint 
>>(say endpoint B) with two ethernet ports with different ip addresses. 
>>The association between two endpoints is ESTABLISHED.
>>
>>Using a test tool I transmit traffic between these two endpoints.
>>With less amount of traffic flowing, if one of the ethernet port IP 
>>address is lost, then the association between these two endpoints remain 
>>ACTIVE and the traffic flows on the second ethernet port.(As per 
>>multi-homing  requirement)
>>With increase in traffic flowing, if one of the ethernet port IP address 
>>is lost, then the association between these two endpoints goes to 
>>INACTIVE. and there is traffic loss. Although the SCTP association is 
>>still present, the ULP association goes down, with the reason "delay ACK 
>>timer expiry". Can anybody help find out why?
>>
>>As per SCTP implemtors guide, there is mention of  updation of cwnd when 
>>user tries to transmit DATA, cwnd should be modified depending on the 
>>flightsize and a newly introduced MAX.BURST.(not present in RFC)
>>               if((flightsize + Max.Burst*MTU) < cwnd)
>>                    cwnd = flightsize +  Max.Burst*MTU
>>
>>Is it really needed to implement this MAX.BURST?
>>Will addition of this MAX.BURST solve the earlier problem mentioned by me?
>>
>>Thanks and regards,
>>devayya
>>
>>P.s: For your reference I have attached a text file of the set up.
>>    
>>
>
>  
>
>>
>>    ----------------                           -----------------
>>    |               |                          |                | 
>>    |               |P| <==================> |R|                | 
>>    |Endpoint A     |                          |  Endpoint  B   |
>>    |               |Q|                      |S|                |
>>    |               |                          |                |
>>    ----------------                           ------------------
>>Traffic is flowing between Endpoint A and endpoint B, using enthernet ports P and R
>>
>>
>>
>>    ----------------                           -----------------
>>    |               |                          |                | 
>>    |               |P|                      |R|                | 
>>    |Endpoint A     |                          |  Endpoint  B   |
>>    |               |Q| <==================> |S|                |
>>    |               |                          |                |
>>    ----------------                           ------------------
>>The enternet port P is lost and now the traffic flows using ethernet port Q and S.
>>    
>>
>
>
>  
>

--------------050004000908090107090404
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
  <title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
There is route from P to R and Q to S. I have configured this using the
system command <br>
"route add -net x.x.x -iface y.y.y.y"<br>
However I could not configure route from R to Q or S to P. Can you
guide me?<br>
<br>
thanks,<br>
devayya<br>
<br>
Brian F. G. Bidulock wrote:<br>
<blockquote cite="mid20060407052836.A4198@openss7.org" type="cite">
  <pre wrap="">devayya,

Does R have a route to Q?

--brian

devayya wrote:                                                     (Fri, 07 Apr 2006 14:43:57)
  </pre>
  <blockquote type="cite">
    <pre wrap="">Hi,

I have a question on SCTP multi-homing,

My SCTP multi-homing test set up is as follows,
There are two etherenet ports with different ip address are configured 
for one endpoint(say endpont A). and similarly I have another endpoint 
(say endpoint B) with two ethernet ports with different ip addresses. 
The association between two endpoints is ESTABLISHED.

Using a test tool I transmit traffic between these two endpoints.
With less amount of traffic flowing, if one of the ethernet port IP 
address is lost, then the association between these two endpoints remain 
ACTIVE and the traffic flows on the second ethernet port.(As per 
multi-homing  requirement)
With increase in traffic flowing, if one of the ethernet port IP address 
is lost, then the association between these two endpoints goes to 
INACTIVE. and there is traffic loss. Although the SCTP association is 
still present, the ULP association goes down, with the reason "delay ACK 
timer expiry". Can anybody help find out why?

As per SCTP implemtors guide, there is mention of  updation of cwnd when 
user tries to transmit DATA, cwnd should be modified depending on the 
flightsize and a newly introduced MAX.BURST.(not present in RFC)
               if((flightsize + Max.Burst*MTU) &lt; cwnd)
                    cwnd = flightsize +  Max.Burst*MTU

Is it really needed to implement this MAX.BURST?
Will addition of this MAX.BURST solve the earlier problem mentioned by me?

Thanks and regards,
devayya

P.s: For your reference I have attached a text file of the set up.
    </pre>
  </blockquote>
  <pre wrap=""><!---->
  </pre>
  <blockquote type="cite">
    <pre wrap="">

    ----------------                           -----------------
    |               |                          |                | 
    |               |P| &lt;==================&gt; |R|                | 
    |Endpoint A     |                          |  Endpoint  B   |
    |               |Q|                      |S|                |
    |               |                          |                |
    ----------------                           ------------------
Traffic is flowing between Endpoint A and endpoint B, using enthernet ports P and R



    ----------------                           -----------------
    |               |                          |                | 
    |               |P|                      |R|                | 
    |Endpoint A     |                          |  Endpoint  B   |
    |               |Q| &lt;==================&gt; |S|                |
    |               |                          |                |
    ----------------                           ------------------
The enternet port P is lost and now the traffic flows using ethernet port Q and S.
    </pre>
  </blockquote>
  <pre wrap=""><!---->

  </pre>
</blockquote>
</body>
</html>

--------------050004000908090107090404--




From tsvwg-bounces@ietf.org Fri Apr 07 07:46:23 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRpPv-0005OC-LF; Fri, 07 Apr 2006 07:46:19 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRpPt-0005O7-V6
	for tsvwg@ietf.org; Fri, 07 Apr 2006 07:46:17 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRpPs-0000rs-G7
	for tsvwg@ietf.org; Fri, 07 Apr 2006 07:46:17 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:Akz8xTw4mpSww+7ZTcCxWafjQo7tI8cj@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k37BkFq04695;
	Fri, 7 Apr 2006 05:46:15 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k37BkFE04405;
	Fri, 7 Apr 2006 05:46:15 -0600
Date: Fri, 7 Apr 2006 05:46:15 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: devayya <devayya.bopaiah@alcatel.com>
Subject: Re: [Tsvwg] SCTP multi-homing association
Message-ID: <20060407054615.A4328@openss7.org>
Mail-Followup-To: devayya <devayya.bopaiah@alcatel.com>, tsvwg <tsvwg@ietf.org>
References: <44362D55.1040806@alcatel.com> <20060407052836.A4198@openss7.org>
	<44364D0B.3030107@alcatel.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <44364D0B.3030107@alcatel.com>;
	from devayya.bopaiah@alcatel.com on Fri, Apr 07, 2006 at
	04:59:15PM +0530
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 9a2be21919e71dc6faef12b370c4ecf5
Cc: tsvwg <tsvwg@ietf.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

devayya,

Yes, your problem is with your routing tables.  You have two disjoint
subnetworks.  SCTP is notorious for failing to handle this situation.
When P fails, Endpoint A loses its route to R, but some SCTP
implementations cannot detect that loss of a route to R and will attempt
to retransmit packets that time out to S to R, resulting in more
timeouts and rather quick association failure.

I cannot help you configure your routing tables on an unknown OS.  Read
the manual pages for the OS.  It might be that you have no hub/switch/
router between the subnets (e.g. direct turnover ethernet cable
connection).  In that case, install a router/switch/hub.

--brian

devayya wrote:                                                     (Fri, 07 Apr 2006 16:59:15)
> 
>    There  is  route  from P to R and Q to S. I have configured this using
>    the system command
>    "route add -net x.x.x -iface y.y.y.y"
>    However  I  could  not  configure route from R to Q or S to P. Can you
>    guide me?
>    thanks,
>    devayya
>    Brian F. G. Bidulock wrote:
> 
> devayya,
> 
> Does R have a route to Q?
> 
> --brian
> 
> devayya wrote:                                                     (Fri, 07 Apr
>  2006 14:43:57)
>   
> 
> Hi,
> 
> I have a question on SCTP multi-homing,
> 
> My SCTP multi-homing test set up is as follows,
> There are two etherenet ports with different ip address are configured 
> for one endpoint(say endpont A). and similarly I have another endpoint 
> (say endpoint B) with two ethernet ports with different ip addresses. 
> The association between two endpoints is ESTABLISHED.
> 
> Using a test tool I transmit traffic between these two endpoints.
> With less amount of traffic flowing, if one of the ethernet port IP 
> address is lost, then the association between these two endpoints remain 
> ACTIVE and the traffic flows on the second ethernet port.(As per 
> multi-homing  requirement)
> With increase in traffic flowing, if one of the ethernet port IP address 
> is lost, then the association between these two endpoints goes to 
> INACTIVE. and there is traffic loss. Although the SCTP association is 
> still present, the ULP association goes down, with the reason "delay ACK 
> timer expiry". Can anybody help find out why?
> 
> As per SCTP implemtors guide, there is mention of  updation of cwnd when 
> user tries to transmit DATA, cwnd should be modified depending on the 
> flightsize and a newly introduced MAX.BURST.(not present in RFC)
>                if((flightsize + Max.Burst*MTU) < cwnd)
>                     cwnd = flightsize +  Max.Burst*MTU
> 
> Is it really needed to implement this MAX.BURST?
> Will addition of this MAX.BURST solve the earlier problem mentioned by me?
> 
> Thanks and regards,
> devayya
> 
> P.s: For your reference I have attached a text file of the set up.
>     
> 
>   
> 
> 
>     ----------------                           -----------------
>     |               |                          |                | 
>     |               |P| <==================> |R|                | 
>     |Endpoint A     |                          |  Endpoint  B   |
>     |               |Q|                      |S|                |
>     |               |                          |                |
>     ----------------                           ------------------
> Traffic is flowing between Endpoint A and endpoint B, using enthernet ports P a
> nd R
> 
> 
> 
>     ----------------                           -----------------
>     |               |                          |                | 
>     |               |P|                      |R|                | 
>     |Endpoint A     |                          |  Endpoint  B   |
>     |               |Q| <==================> |S|                |
>     |               |                          |                |
>     ----------------                           ------------------
> The enternet port P is lost and now the traffic flows using ethernet port Q and
>  S.
>     
> 
> 
>   

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




From tsvwg-bounces@ietf.org Fri Apr 07 07:55:33 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRpYk-0002hz-1A; Fri, 07 Apr 2006 07:55:26 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRpYi-0002hu-Rm
	for tsvwg@ietf.org; Fri, 07 Apr 2006 07:55:24 -0400
Received: from mail-n.franken.de ([193.175.24.27] helo=ilsa.franken.de)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRpYh-0001CR-H7
	for tsvwg@ietf.org; Fri, 07 Apr 2006 07:55:24 -0400
Received: from [192.168.1.10] (p508F712C.dip.t-dialin.net [80.143.113.44])
	by ilsa.franken.de (Postfix) with ESMTP
	id 53F1B245C3; Fri,  7 Apr 2006 13:55:21 +0200 (CEST)
	(KNF account authenticated via SMTP-AUTH)
In-Reply-To: <44362D55.1040806@alcatel.com>
References: <44362D55.1040806@alcatel.com>
Mime-Version: 1.0 (Apple Message framework v746.2)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <24A3ABE8-1B4B-4127-8AF9-70BEAD16801C@lurchi.franken.de>
Content-Transfer-Encoding: 7bit
From: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>
Subject: Re: [Tsvwg] SCTP multi-homing association
Date: Fri, 7 Apr 2006 13:54:56 +0200
To: devayya <devayya.bopaiah@alcatel.com>
X-Mailer: Apple Mail (2.746.2)
X-Spam-Score: 0.0 (/)
X-Scan-Signature: fb6060cb60c0cea16e3f7219e40a0a81
Cc: tsvwg <tsvwg@ietf.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Hi Devayya,

'delay ACK timer expiry' is something which should result in the  
sending of a SACK, not
in the teardown of an association. This looks like a bug to me...

Best regards
Michael

On Apr 7, 2006, at 11:13 AM, devayya wrote:

> Hi,
>
> I have a question on SCTP multi-homing,
>
> My SCTP multi-homing test set up is as follows,
> There are two etherenet ports with different ip address are  
> configured for one endpoint(say endpont A). and similarly I have  
> another endpoint (say endpoint B) with two ethernet ports with  
> different ip addresses. The association between two endpoints is  
> ESTABLISHED.
>
> Using a test tool I transmit traffic between these two endpoints.
> With less amount of traffic flowing, if one of the ethernet port IP  
> address is lost, then the association between these two endpoints  
> remain ACTIVE and the traffic flows on the second ethernet port.(As  
> per multi-homing  requirement)
> With increase in traffic flowing, if one of the ethernet port IP  
> address is lost, then the association between these two endpoints  
> goes to INACTIVE. and there is traffic loss. Although the SCTP  
> association is still present, the ULP association goes down, with  
> the reason "delay ACK timer expiry". Can anybody help find out why?
>
> As per SCTP implemtors guide, there is mention of  updation of cwnd  
> when user tries to transmit DATA, cwnd should be modified depending  
> on the flightsize and a newly introduced MAX.BURST.(not present in  
> RFC)
>               if((flightsize + Max.Burst*MTU) < cwnd)
>                    cwnd = flightsize +  Max.Burst*MTU
>
> Is it really needed to implement this MAX.BURST?
> Will addition of this MAX.BURST solve the earlier problem mentioned  
> by me?
>
> Thanks and regards,
> devayya
>
> P.s: For your reference I have attached a text file of the set up.
>
>
>
>     ----------------                           -----------------
>     |               |                          |                |
>     |               |P| <==================> |R|                |
>     |Endpoint A     |                          |  Endpoint  B   |
>     |               |Q|                      |S|                |
>     |               |                          |                |
>     ----------------                           ------------------
> Traffic is flowing between Endpoint A and endpoint B, using  
> enthernet ports P and R
>
>
>
>     ----------------                           -----------------
>     |               |                          |                |
>     |               |P|                      |R|                |
>     |Endpoint A     |                          |  Endpoint  B   |
>     |               |Q| <==================> |S|                |
>     |               |                          |                |
>     ----------------                           ------------------
> The enternet port P is lost and now the traffic flows using  
> ethernet port Q and S.





From tsvwg-bounces@ietf.org Fri Apr 07 08:09:35 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRpmH-0003yK-SC; Fri, 07 Apr 2006 08:09:25 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRpmG-0003yA-4G
	for tsvwg@ietf.org; Fri, 07 Apr 2006 08:09:24 -0400
Received: from hostreea2.alcatel.com ([143.209.238.162]
	helo=audl953.usa.alcatel.com)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRpmF-0001uI-Qx
	for tsvwg@ietf.org; Fri, 07 Apr 2006 08:09:24 -0400
Received: from axes-mach01.usa.alcatel.com (axes-mach01.usa.alcatel.com
	[10.255.0.20])
	by audl953.usa.alcatel.com (ALCANET) with ESMTP id k37C9LQx025953;
	Fri, 7 Apr 2006 07:09:22 -0500
Received: from [10.255.1.166] (ax-sp-pc114.usa.alcatel.com [10.255.1.166])
	by axes-mach01.usa.alcatel.com (8.12.10+Sun/8.12.2) with ESMTP id
	k37CCstL022288; Fri, 7 Apr 2006 17:42:55 +0530 (IST)
Message-ID: <44365568.2010902@alcatel.com>
Date: Fri, 07 Apr 2006 17:34:56 +0530
From: devayya <devayya.bopaiah@alcatel.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>
Subject: Re: [Tsvwg] SCTP multi-homing association
References: <44362D55.1040806@alcatel.com>
	<24A3ABE8-1B4B-4127-8AF9-70BEAD16801C@lurchi.franken.de>
In-Reply-To: <24A3ABE8-1B4B-4127-8AF9-70BEAD16801C@lurchi.franken.de>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Scanned-By: MIMEDefang 2.51 on 143.209.238.34
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 32b73d73e8047ed17386f9799119ce43
Cc: tsvwg <tsvwg@ietf.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Michael,

Consider a case where Endpoint A has sent a data to endpoint B. and 
endpoint B has sent SACK for that received data. But before that SACK 
could reach endpoint A the ehernet port P of endpoint A has gone down. 
Now the traffic switches over to ethernet port Q of endpoint A. Within 
that fraction of second wher the switch over fro port P to port Q 
happens there is bound to be some message loss. There is no way that 
endpoint B will know that the SACK sent by it is not received by 
endpoint A. Or is it there?  and so the delay ack timer expiry happens, 
Correct me if I am wrong.

thanks,
devayya

Michael Tuexen wrote:

> Hi Devayya,
>
> 'delay ACK timer expiry' is something which should result in the  
> sending of a SACK, not
> in the teardown of an association. This looks like a bug to me...
>
> Best regards
> Michael
>
> On Apr 7, 2006, at 11:13 AM, devayya wrote:
>
>> Hi,
>>
>> I have a question on SCTP multi-homing,
>>
>> My SCTP multi-homing test set up is as follows,
>> There are two etherenet ports with different ip address are  
>> configured for one endpoint(say endpont A). and similarly I have  
>> another endpoint (say endpoint B) with two ethernet ports with  
>> different ip addresses. The association between two endpoints is  
>> ESTABLISHED.
>>
>> Using a test tool I transmit traffic between these two endpoints.
>> With less amount of traffic flowing, if one of the ethernet port IP  
>> address is lost, then the association between these two endpoints  
>> remain ACTIVE and the traffic flows on the second ethernet port.(As  
>> per multi-homing  requirement)
>> With increase in traffic flowing, if one of the ethernet port IP  
>> address is lost, then the association between these two endpoints  
>> goes to INACTIVE. and there is traffic loss. Although the SCTP  
>> association is still present, the ULP association goes down, with  
>> the reason "delay ACK timer expiry". Can anybody help find out why?
>>
>> As per SCTP implemtors guide, there is mention of  updation of cwnd  
>> when user tries to transmit DATA, cwnd should be modified depending  
>> on the flightsize and a newly introduced MAX.BURST.(not present in  RFC)
>>               if((flightsize + Max.Burst*MTU) < cwnd)
>>                    cwnd = flightsize +  Max.Burst*MTU
>>
>> Is it really needed to implement this MAX.BURST?
>> Will addition of this MAX.BURST solve the earlier problem mentioned  
>> by me?
>>
>> Thanks and regards,
>> devayya
>>
>> P.s: For your reference I have attached a text file of the set up.
>>
>>
>>
>>     ----------------                           -----------------
>>     |               |                          |                |
>>     |               |P| <==================> |R|                |
>>     |Endpoint A     |                          |  Endpoint  B   |
>>     |               |Q|                      |S|                |
>>     |               |                          |                |
>>     ----------------                           ------------------
>> Traffic is flowing between Endpoint A and endpoint B, using  
>> enthernet ports P and R
>>
>>
>>
>>     ----------------                           -----------------
>>     |               |                          |                |
>>     |               |P|                      |R|                |
>>     |Endpoint A     |                          |  Endpoint  B   |
>>     |               |Q| <==================> |S|                |
>>     |               |                          |                |
>>     ----------------                           ------------------
>> The enternet port P is lost and now the traffic flows using  ethernet 
>> port Q and S.
>
>
>




From tsvwg-bounces@ietf.org Fri Apr 07 08:13:56 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRpqa-0006RH-E0; Fri, 07 Apr 2006 08:13:52 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRpqY-0006LD-Rg
	for tsvwg@ietf.org; Fri, 07 Apr 2006 08:13:50 -0400
Received: from mail-n.franken.de ([193.175.24.27] helo=ilsa.franken.de)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRpqY-00020K-Gu
	for tsvwg@ietf.org; Fri, 07 Apr 2006 08:13:50 -0400
Received: from [192.168.1.10] (p508F712C.dip.t-dialin.net [80.143.113.44])
	by ilsa.franken.de (Postfix) with ESMTP
	id 11784245CE; Fri,  7 Apr 2006 14:13:49 +0200 (CEST)
	(KNF account authenticated via SMTP-AUTH)
In-Reply-To: <44365568.2010902@alcatel.com>
References: <44362D55.1040806@alcatel.com>
	<24A3ABE8-1B4B-4127-8AF9-70BEAD16801C@lurchi.franken.de>
	<44365568.2010902@alcatel.com>
Mime-Version: 1.0 (Apple Message framework v746.2)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <13768463-862B-4964-9FEB-E8AA4B306770@lurchi.franken.de>
Content-Transfer-Encoding: 7bit
From: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>
Subject: Re: [Tsvwg] SCTP multi-homing association
Date: Fri, 7 Apr 2006 14:13:47 +0200
To: devayya <devayya.bopaiah@alcatel.com>
X-Mailer: Apple Mail (2.746.2)
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 287c806b254c6353fcb09ee0e53bbc5e
Cc: tsvwg <tsvwg@ietf.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Devayya,

see my comments in-line.

Best regards
Michael

On Apr 7, 2006, at 2:04 PM, devayya wrote:

> Michael,
>
> Consider a case where Endpoint A has sent a data to endpoint B. and  
> endpoint B has sent SACK for that received data. But before that  
> SACK could reach endpoint A the ehernet port P of endpoint A has  
> gone down. Now the traffic switches over to
If the SACK does not arrive a Endpoint A, there will be a T3 timer  
running at Endpoint A for that DATA chunk and
it will be retransmitted when T3 expires.
> ethernet port Q of endpoint A. Within that fraction of second wher  
> the switch over fro port P to port Q happens there is bound to be  
> some message loss. There is no way that endpoint B will know that  
> the SACK sent by it is not received by endpoint A. Or is it there?   
> and so the delay ack timer expiry happens, Correct me if I am wrong.
The delay ack timer will only delay the sending of SACKs. That is  
all. When it expires, a SACK will be sent. It does
NOT supervise the successful transmission of SACKs.
>
> thanks,
> devayya
>
> Michael Tuexen wrote:
>
>> Hi Devayya,
>>
>> 'delay ACK timer expiry' is something which should result in the   
>> sending of a SACK, not
>> in the teardown of an association. This looks like a bug to me...
>>
>> Best regards
>> Michael
>>
>> On Apr 7, 2006, at 11:13 AM, devayya wrote:
>>
>>> Hi,
>>>
>>> I have a question on SCTP multi-homing,
>>>
>>> My SCTP multi-homing test set up is as follows,
>>> There are two etherenet ports with different ip address are   
>>> configured for one endpoint(say endpont A). and similarly I have   
>>> another endpoint (say endpoint B) with two ethernet ports with   
>>> different ip addresses. The association between two endpoints is   
>>> ESTABLISHED.
>>>
>>> Using a test tool I transmit traffic between these two endpoints.
>>> With less amount of traffic flowing, if one of the ethernet port  
>>> IP  address is lost, then the association between these two  
>>> endpoints  remain ACTIVE and the traffic flows on the second  
>>> ethernet port.(As  per multi-homing  requirement)
>>> With increase in traffic flowing, if one of the ethernet port IP   
>>> address is lost, then the association between these two  
>>> endpoints  goes to INACTIVE. and there is traffic loss. Although  
>>> the SCTP  association is still present, the ULP association goes  
>>> down, with  the reason "delay ACK timer expiry". Can anybody help  
>>> find out why?
>>>
>>> As per SCTP implemtors guide, there is mention of  updation of  
>>> cwnd  when user tries to transmit DATA, cwnd should be modified  
>>> depending  on the flightsize and a newly introduced MAX.BURST. 
>>> (not present in  RFC)
>>>               if((flightsize + Max.Burst*MTU) < cwnd)
>>>                    cwnd = flightsize +  Max.Burst*MTU
>>>
>>> Is it really needed to implement this MAX.BURST?
>>> Will addition of this MAX.BURST solve the earlier problem  
>>> mentioned  by me?
>>>
>>> Thanks and regards,
>>> devayya
>>>
>>> P.s: For your reference I have attached a text file of the set up.
>>>
>>>
>>>
>>>     ----------------                           -----------------
>>>     |               |                          |                |
>>>     |               |P| <==================> |R|                |
>>>     |Endpoint A     |                          |  Endpoint  B   |
>>>     |               |Q|                      |S|                |
>>>     |               |                          |                |
>>>     ----------------                           ------------------
>>> Traffic is flowing between Endpoint A and endpoint B, using   
>>> enthernet ports P and R
>>>
>>>
>>>
>>>     ----------------                           -----------------
>>>     |               |                          |                |
>>>     |               |P|                      |R|                |
>>>     |Endpoint A     |                          |  Endpoint  B   |
>>>     |               |Q| <==================> |S|                |
>>>     |               |                          |                |
>>>     ----------------                           ------------------
>>> The enternet port P is lost and now the traffic flows using   
>>> ethernet port Q and S.
>>
>>
>>
>





From tsvwg-bounces@ietf.org Fri Apr 07 08:15:32 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRpsC-000741-Ed; Fri, 07 Apr 2006 08:15:32 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRpsB-00073w-Lo
	for tsvwg@ietf.org; Fri, 07 Apr 2006 08:15:31 -0400
Received: from mgw-ext14.nokia.com ([131.228.20.173])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRpsB-00021g-4n
	for tsvwg@ietf.org; Fri, 07 Apr 2006 08:15:31 -0400
Received: from esebh108.NOE.Nokia.com (esebh108.ntc.nokia.com [172.21.143.145])
	by mgw-ext14.nokia.com (Switch-3.1.8/Switch-3.1.7) with ESMTP id
	k37CEWcr009422; Fri, 7 Apr 2006 15:14:32 +0300
Received: from esebh102.NOE.Nokia.com ([172.21.138.183]) by
	esebh108.NOE.Nokia.com with Microsoft SMTPSVC(6.0.3790.1830); 
	Fri, 7 Apr 2006 15:15:22 +0300
Received: from esebe104.NOE.Nokia.com ([172.21.143.44]) by
	esebh102.NOE.Nokia.com with Microsoft SMTPSVC(6.0.3790.1830); 
	Fri, 7 Apr 2006 15:15:22 +0300
x-mimeole: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Subject: RE: [Tsvwg] SCTP multi-homing association
Date: Fri, 7 Apr 2006 15:15:21 +0300
Message-ID: <921445D056CC0E4FAFA3F0566793E82701E915FE@esebe104.NOE.Nokia.com>
In-Reply-To: <44365568.2010902@alcatel.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [Tsvwg] SCTP multi-homing association
Thread-Index: AcZaPEAtm3Rw8glsRAm6ETTXDHNNrgAACq3A
From: <Ivan.Arias-Rodriguez@nokia.com>
To: <devayya.bopaiah@alcatel.com>, <Michael.Tuexen@lurchi.franken.de>
X-OriginalArrivalTime: 07 Apr 2006 12:15:22.0047 (UTC)
	FILETIME=[F15B54F0:01C65A3C]
X-Spam-Score: 0.2 (/)
X-Scan-Signature: 2086112c730e13d5955355df27e3074b
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

	Hi!

	I think you are mixing two things here: SACK delay timer and =
retransmission timer. When you get an SCTP datagram containing DATA =
chunks, you don't send the SACK chunk immediately. You either wait till =
you get a second datagram containing more DATA chunks or the expiration =
of the SACK delay timer, whatever happens earlier. And then you send =
back the SACK chunk.=20

	That timer should be running at the receiver of the DATA chunks, and by =
no means should cause any problem to the association. Simply, when the =
SACK delay timer expires you send the SACK chunk, and that's all.

	BR Iv=E1n Arias Rodr=EDguez

>-----Original Message-----
>From: ext devayya [mailto:devayya.bopaiah@alcatel.com]=20
>Sent: 07 April, 2006 15:05
>To: Michael Tuexen
>Cc: tsvwg
>Subject: Re: [Tsvwg] SCTP multi-homing association
>
>Michael,
>
>Consider a case where Endpoint A has sent a data to endpoint=20
>B. and endpoint B has sent SACK for that received data. But=20
>before that SACK could reach endpoint A the ehernet port P of=20
>endpoint A has gone down.=20
>Now the traffic switches over to ethernet port Q of endpoint=20
>A. Within that fraction of second wher the switch over fro=20
>port P to port Q happens there is bound to be some message=20
>loss. There is no way that endpoint B will know that the SACK=20
>sent by it is not received by endpoint A. Or is it there?  and=20
>so the delay ack timer expiry happens, Correct me if I am wrong.
>
>thanks,
>devayya
>
>Michael Tuexen wrote:
>
>> Hi Devayya,
>>
>> 'delay ACK timer expiry' is something which should result in the=20
>> sending of a SACK, not in the teardown of an association. This looks=20
>> like a bug to me...
>>
>> Best regards
>> Michael
>>
>> On Apr 7, 2006, at 11:13 AM, devayya wrote:
>>
>>> Hi,
>>>
>>> I have a question on SCTP multi-homing,
>>>
>>> My SCTP multi-homing test set up is as follows, There are two=20
>>> etherenet ports with different ip address are configured for one=20
>>> endpoint(say endpont A). and similarly I have another endpoint (say=20
>>> endpoint B) with two ethernet ports with different ip=20
>addresses. The=20
>>> association between two endpoints is ESTABLISHED.
>>>
>>> Using a test tool I transmit traffic between these two endpoints.
>>> With less amount of traffic flowing, if one of the ethernet port IP=20
>>> address is lost, then the association between these two endpoints=20
>>> remain ACTIVE and the traffic flows on the second ethernet port.(As=20
>>> per multi-homing  requirement) With increase in traffic flowing, if=20
>>> one of the ethernet port IP address is lost, then the association=20
>>> between these two endpoints goes to INACTIVE. and there is traffic=20
>>> loss. Although the SCTP association is still present, the ULP=20
>>> association goes down, with the reason "delay ACK timer=20
>expiry". Can=20
>>> anybody help find out why?
>>>
>>> As per SCTP implemtors guide, there is mention of  updation of cwnd=20
>>> when user tries to transmit DATA, cwnd should be modified depending=20
>>> on the flightsize and a newly introduced MAX.BURST.(not=20
>present in  RFC)
>>>               if((flightsize + Max.Burst*MTU) < cwnd)
>>>                    cwnd =3D flightsize +  Max.Burst*MTU
>>>
>>> Is it really needed to implement this MAX.BURST?
>>> Will addition of this MAX.BURST solve the earlier problem mentioned=20
>>> by me?
>>>
>>> Thanks and regards,
>>> devayya
>>>
>>> P.s: For your reference I have attached a text file of the set up.
>>>
>>>
>>>
>>>     ----------------                           -----------------
>>>     |               |                          |                |
>>>     |               |P| =
<=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D> |R|             =
   |
>>>     |Endpoint A     |                          |  Endpoint  B   |
>>>     |               |Q|                      |S|                |
>>>     |               |                          |                |
>>>     ----------------                           ------------------
>>> Traffic is flowing between Endpoint A and endpoint B, using=20
>enthernet=20
>>> ports P and R
>>>
>>>
>>>
>>>     ----------------                           -----------------
>>>     |               |                          |                |
>>>     |               |P|                      |R|                |
>>>     |Endpoint A     |                          |  Endpoint  B   |
>>>     |               |Q| =
<=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D> |S|             =
   |
>>>     |               |                          |                |
>>>     ----------------                           ------------------
>>> The enternet port P is lost and now the traffic flows using=20
> ethernet=20
>>> port Q and S.
>>
>>
>>
>
>




From tsvwg-bounces@ietf.org Fri Apr 07 09:58:24 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRrTb-0008Dw-Bc; Fri, 07 Apr 2006 09:58:15 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRrTa-0008Do-4S
	for tsvwg@ietf.org; Fri, 07 Apr 2006 09:58:14 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRrTZ-0005OD-N0
	for tsvwg@ietf.org; Fri, 07 Apr 2006 09:58:14 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FRrTZ-0000WT-7q; Fri, 07 Apr 2006 07:57:51 -0600
Message-ID: <44366FF9.7020801@middle.net>
Date: Fri, 07 Apr 2006 07:58:17 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: devayya <devayya.bopaiah@alcatel.com>
Subject: Re: [Tsvwg] SCTP multi-homing association
References: <44362D55.1040806@alcatel.com>	<24A3ABE8-1B4B-4127-8AF9-70BEAD16801C@lurchi.franken.de>
	<44365568.2010902@alcatel.com>
In-Reply-To: <44365568.2010902@alcatel.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 1676547e4f33b5e63227e9c02bd359e3
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

devayya wrote:

> Michael,
>
> Consider a case where Endpoint A has sent a data to endpoint B. and 
> endpoint B has sent SACK for that received data. But before that SACK 
> could reach endpoint A the ehernet port P of endpoint A has gone down. 
> Now the traffic switches over to ethernet port Q of endpoint A. Within 
> that fraction of second wher the switch over fro port P to port Q 
> happens there is bound to be some message loss. There is no way that 
> endpoint B will know that the SACK sent by it is not received by 
> endpoint A. Or is it there?  and so the delay ack timer expiry 
> happens, Correct me if I am wrong.
>
> thanks,
> devayya

First of all, can you tell us which OS and SCTP implementation you are 
using?  If not a common one, 'internal/proprietary' is fine.   A few 
comments:

1. Normally in SCTP, associations do not go active / inactive, paths do. 
2. Delayed ACK timer expiry is a normal event that is not an indication 
of a failure,
3. If a DATA chunk is lost, either a later SACK or the T3 timer will 
eventually trigger a retransmission, which is supposed to be on an 
alternate path if available.
4.  If a DATA chunk is received but a SACK is lost, and there are no 
follow on SACKs, the DATA chunk will also be retransmitted in the same 
manner, but ignored since it was already received.
5. Max.Burst probably doesn't have anything to do with your problem.

- Mark B.




>
>
> Michael Tuexen wrote:
>
>> Hi Devayya,
>>
>> 'delay ACK timer expiry' is something which should result in the  
>> sending of a SACK, not
>> in the teardown of an association. This looks like a bug to me...
>>
>> Best regards
>> Michael
>>
>> On Apr 7, 2006, at 11:13 AM, devayya wrote:
>>
>>> Hi,
>>>
>>> I have a question on SCTP multi-homing,
>>>
>>> My SCTP multi-homing test set up is as follows,
>>> There are two etherenet ports with different ip address are  
>>> configured for one endpoint(say endpont A). and similarly I have  
>>> another endpoint (say endpoint B) with two ethernet ports with  
>>> different ip addresses. The association between two endpoints is  
>>> ESTABLISHED.
>>>
>>> Using a test tool I transmit traffic between these two endpoints.
>>> With less amount of traffic flowing, if one of the ethernet port IP  
>>> address is lost, then the association between these two endpoints  
>>> remain ACTIVE and the traffic flows on the second ethernet port.(As  
>>> per multi-homing  requirement)
>>> With increase in traffic flowing, if one of the ethernet port IP  
>>> address is lost, then the association between these two endpoints  
>>> goes to INACTIVE. and there is traffic loss. Although the SCTP  
>>> association is still present, the ULP association goes down, with  
>>> the reason "delay ACK timer expiry". Can anybody help find out why?
>>>
>>> As per SCTP implemtors guide, there is mention of  updation of cwnd  
>>> when user tries to transmit DATA, cwnd should be modified depending  
>>> on the flightsize and a newly introduced MAX.BURST.(not present in  
>>> RFC)
>>>               if((flightsize + Max.Burst*MTU) < cwnd)
>>>                    cwnd = flightsize +  Max.Burst*MTU
>>>
>>> Is it really needed to implement this MAX.BURST?
>>> Will addition of this MAX.BURST solve the earlier problem mentioned  
>>> by me?
>>>
>>> Thanks and regards,
>>> devayya
>>>
>>> P.s: For your reference I have attached a text file of the set up.
>>>
>>>
>>>
>>>     ----------------                           -----------------
>>>     |               |                          |                |
>>>     |               |P| <==================> |R|                |
>>>     |Endpoint A     |                          |  Endpoint  B   |
>>>     |               |Q|                      |S|                |
>>>     |               |                          |                |
>>>     ----------------                           ------------------
>>> Traffic is flowing between Endpoint A and endpoint B, using  
>>> enthernet ports P and R
>>>
>>>
>>>
>>>     ----------------                           -----------------
>>>     |               |                          |                |
>>>     |               |P|                      |R|                |
>>>     |Endpoint A     |                          |  Endpoint  B   |
>>>     |               |Q| <==================> |S|                |
>>>     |               |                          |                |
>>>     ----------------                           ------------------
>>> The enternet port P is lost and now the traffic flows using  
>>> ethernet port Q and S.
>>
>>
>>
>>
>
>





From tsvwg-bounces@ietf.org Fri Apr 07 15:01:43 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRwDA-0000MB-E1; Fri, 07 Apr 2006 15:01:36 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRwD9-0000M1-79
	for tsvwg@ietf.org; Fri, 07 Apr 2006 15:01:35 -0400
Received: from sj-iport-2-in.cisco.com ([171.71.176.71]
	helo=sj-iport-2.cisco.com) by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FRwD8-0001l1-Ee
	for tsvwg@ietf.org; Fri, 07 Apr 2006 15:01:35 -0400
Received: from sj-core-5.cisco.com ([171.71.177.238])
	by sj-iport-2.cisco.com with ESMTP; 07 Apr 2006 12:01:34 -0700
X-IronPort-AV: i="4.04,98,1144047600"; 
	d="scan'208"; a="319384147:sNHT31258852"
Received: from xbh-sjc-211.amer.cisco.com (xbh-sjc-211.cisco.com
	[171.70.151.144])
	by sj-core-5.cisco.com (8.12.10/8.12.6) with ESMTP id k37J1X7b025940
	for <tsvwg@ietf.org>; Fri, 7 Apr 2006 12:01:33 -0700 (PDT)
Received: from xfe-sjc-212.amer.cisco.com ([171.70.151.187]) by
	xbh-sjc-211.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Fri, 7 Apr 2006 12:01:23 -0700
Received: from jmpolk-wxp.cisco.com ([10.89.16.24]) by
	xfe-sjc-212.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Fri, 7 Apr 2006 12:01:23 -0700
Message-Id: <4.3.2.7.2.20060407135247.0492d4a8@email.cisco.com>
X-Sender: jmpolk@email.cisco.com
X-Mailer: QUALCOMM Windows Eudora Version 4.3.2
Date: Fri, 07 Apr 2006 14:01:22 -0500
To: tsvwg <tsvwg@ietf.org>
From: "James M. Polk" <jmpolk@cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"; format=flowed
Content-Transfer-Encoding: quoted-printable
X-OriginalArrivalTime: 07 Apr 2006 19:01:23.0168 (UTC)
	FILETIME=[A9B76600:01C65A75]
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 9178bae9f85419fdc08e9f2c86e345d0
Subject: [Tsvwg] DRAFT TSVWG meeting notes - corrections requested
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

TSVWG

Here are the DRAFT notes for our meeting in Dallas. These are not short=20
notes, as there was a lot going on during the meeting, and we included the=
=20
Bar BoF notes for those that did not or could not attend.

I've added a new section, called a Hum report, to review what the room=20
agreed to.  Objections to any hum ought to be made asap, for the good of=20
the group's progress.

Please offer corrections, modifications, deletions, additions..... (in a=20
timely manner)

I'd like to upload these sometime later next week

+++++++++++++++++++++
TSVWG Session
Wednesday, March 22nd, 2006
0900-1130                                   Attendance 36+19=3D55 people

Chair's                                                 9:00-9:15
       Agenda Bashing                                    (15 min)
       NOTE WELL
       Document Status and Accomplishments
       New Charter?
       New Milestones?

No Agenda Bashing
- Matt said his TCP MIB Ext doc is ready for WGLC (see below)

- During Doc Status - to consider
   draft-ietf-tsvwg-vpn-signaled-preemption WGLC worthy (hum 100% for,
   0% against)

Bar BoF formal WG Announcement

- Gerald Ash was concerned of the scope of the bar BOF. Not prepared to
   present work there.
- Georgios balked at not having time to prepare slides
- Allison said to deal with it

Aaron's                                         9:15-9:25 (until 9:26)
Congestion Control Research Group               (10 min)

- Discussed congestion control RG
- Problem: High throughput requires unrealistic low packet loss rate,
       =3D=3D> low throughput in lossy environments
- Plan  4 Phases  Challenge, Solution Proposals, Discussion, Consensus;
- Rec to IETF.

Michael and Randy's                             9:25-9:35  (until 9:39)
draft-tuexen-tsvwg-sctp-padding-00              (10 min)
(might want to add SCTP auth and addip to here)

- Padding chuck is needed by path MTU discovery
- Need motivation in doc
- Lars suggested that the doc be split in two (padding and the other
   stuff)
- Matt said a lack of motivation is good to stay as is
- Michael said <something?>
- Hum to adopt as WG item (knowing sec sect needs a bit more)
- Hum 100% for adopting, 0% against

- ADDIP
       - Not intended for mobility, unlike DCCP
       - Need to slightly modify before progressing to WGLC

- SCTP-AUTH has known dependencies that are needed
       - AUTH Only want to secure ADDIP
       - Needs clarification of Null key, Key Identifier
       - Need to slightly modify before progressing to WGLC
       - Must go together with ADDIP through process

Kwok's                                          9:35-9:45  (until 9:51)
draft-chan-tsvwg-diffserv-class-aggr-03         (10 min)

- Fred presented
- Motivation to provide less granular treatments
- Editorial changes
- Changed name of default class
- Dave McDyson believe this should go to WG status and his network will
   use this
- Dave Blacks thinks this is a good idea and should be a WG item
- Hum for WG status (100% pass, 0% against)
- Dave Black will be a reviewer
- Magnus thinks this ought to have mail-list confirmation for WG status

Phil's                                          9:45-9:57  (until 9:57)
draft-briscoe-tsvwg-cl-phb-01.txt               (12 min)
draft-briscoe-tsvwg-cl-architecture-02.txt

- Split document into two (a deployment ID and a marking ID)
- CL-Arch has changes (on a slide)
- CL-PHB has changes (on a slide)
- Problem: they have 5 things to encode, and 4 possibilities, so
   creativity is necessary
- Asked that folks come to the Bar BoF and give comments and feedback

Bob's                                          9:57-10:15 (until 10:21)
draft-briscoe-tsvwg-re-ecn-tcp-01.txt           (18 min)
draft-briscoe-tsvwg-re-ecn-border-cheat-00.txt

- RE-ECN-TCP
       - About adding accountability to IP congestion
       - How to allow some networks to do policing.
       - Allow different networks to deploy different policies some
         conservative, some liberal.
       - Backwards compatible with the Evil Bit
       - Goal is also to motivate the use of ECN to allow congestion
         marks to be able to be taken advantage of by SPs
       - Question from someone - Problem finding the extra bit in a IPv6
         header, but there is the problem of by adding a bit - you need
         to add a whole 128 bit extension
       - Dave Black wants to open up discussion on using the last bit as
         a real discussion point
         - He's also concerned about the differences between this
           solution in v4 vs. v6 (which isn't public yet)
       - Allison offered that perhaps a creative use of the flow label
         in v6 could be the answer

- RE-ECN-Border
       - Chairs only allowed 1 slide of preso due to time
       - Gave short summary only

Matt's                                        10:15-10:25 (until 10:24)
draft-ietf-tsvwg-tcp-mib-extension-08.txt       (10 min)

- MIB Dr didn't give many comments
- MIB Dr said it's basically done
- There was a template problem that wasn't caught, but this can be
   corrected during the WGLC process
- Ready for WGLC

Francois's                                    10:25-10:40 (until 10:53)
draft-lefaucheur-emergency-rsvp-00              (20 min)
draft-ietf-tsvwg-rsvp-ipsec

- RSVP-IPSec
       - Removed SPI from what identifies the Reservations
       - This made everything here simpler
       - Added notion of extended virtual destination port, also used in
         MPLS
       - Hum to WGLC (100% pass)
       - Agenda had a mistake that this was a individual status (the
         chairs take the hit on that one)

- RSVP-Emergency
       - Discussed capacity handling in a number of ways based on a
         local configuration
       - Speaker asked to take this
       - Gerald Ash - how is this mech mapped to the RPH
       - James answered - that SIP is disconnected from Router
         behaviors, and that local policy determines the mapping from
         the RPH to RSVP Priority values and treatments
       - Janet commented - <missed it>
       - Georgios asked about if there are too many higher priority
         calls, at what point, at what capacity does this doc suggest to
         terminate calls
       - Francois answer that preemption is not discussed
       - Allison commented on reorganizing the doc but separating the BW
         model from the mechanism, as this ID is clearly a PS track, but
         the model is not - so it can be in an appendix
       - Francois to think about that

Sally's                                     10:40-10:50   (until 11:09)
draft-ietf-tsvwg-quickstart                     (10 min)

- Mitchell Airblix (sp?) comments off-line will cause a few more
   paragraphs to be added to the ID
- Tim Sheppard - Multi-Access link Ethernet scenario is clearly
   problematic (and Sally agreed)
- This won=92t work if TCP header is encrypted (military nets) but there
   are some nets where this will work.
- Assumption is that this mechanism will never ever be a default
   enabled feature
- Fred - asked who the audience to be then?
- Sally - expects to be used in very high BW little used intra-
   networks, ID clearly lists numerous cases in which this mechanism
   will not work - including anytime a packet in tunneled
- Bob Briscoe - believes this could work in a server farm
- Randy Stewart - asked about IPv6
- Sally - this is fine in IPv6
- Lou Berger - believes this is limited in scope

Jukka's and Markku's                          10:50-11:00 (until 11:19)
draft-manner-tsvwg-lrsvp-00                      (10 min)

- This is not a new ID
- Motivation: clearly e2e QoS is not here, and in times that the remote
   node is not RSVP enable, this mechanism allows RSVP as far as it can,
   and not mandate e2e or even a mechanism at the remote node.
- Reviewed the 2 extensions necessary to enable this mechanism
- Wants 2 bits from Session Object, 2 new message types.
- Francois - thinks there is value is documenting proxy behavior, not
   clear that the endsystem needs to be the starter of this mechanism
   (i.e. a L-RSVP proxy in a early router, more like edge-to-edge)
       - Will send comments to the list now
- Bob - interested how a pair of hosts will ever know if there are in
   this situation or not, perhaps adding a new protocol ID could help
   identify it
       - Will also send comments to the list now that he understand this
         (from the preso, as he didn't from the ID itself)
- No vote for WG status vs. the slide suggestion of moving to an EXP
   status
- Chairs asked for list comments before going forward in either
   direction

Jonathan's                                      11:00-11:10
draft-rosenberg-sipping-overload-reqs-00        (10 min)

- This preso was skipped

Transport Area Chairs Update                 11:10-11:25 (until 11:32)
DCCP Chair (Lars Eggert)                        (15 min)

- Gave overview of DCCP to TSVWG
       - specs pending.
       - WG being rechartered.
       - New chair Gorry

RMT rep (Allison stepping in)

- Reliable Multicast WG is moving their documents from experimental to
   proposed this spring. Flute/ACL with a NACK protocol. One called NORM
   with building blocks.. giant group of protocols going to PS.. must
   look at very closely...

IPPM rep (Matthew J Zekauskas)

- Gave overview of IPPM to TSVWG
- Chartered to do metrics for paths
- list of them on slides.
- Cooperate with ITU
- Did all the low hanging fruit first (slide 7)
- No bulk transfer capacity metric yet.
- Slide 8/9 more completed work.
- Still working .. stuff in page 10.
    reordering moving forward WGLC.

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
Hum Report

A review of Hums taken during the meeting:

- draft-ietf-tsvwg-vpn-signaled-preemption-00 considered WGLC worthy
   (hum 100% for, 0% against)
- draft-ietf-tsvwg-rsvp-ipsec-00 considered WGLC worthy (hum 100% for,
   0% against)
- draft-tuexen-tsvwg-sctp-padding-00 to be WG item (hum 100% for, 0%
   against)
- draft-chan-tsvwg-diffserv-class-aggr-03 to be WG item (hum 100% for,
   0% against)

no other hums were asked for.

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
CL+RMD Bar BoF                           attendance:  13+15+5=3D33

Chairing James Polk

Francois on the overall approaches.

Edge2Edge/Controlled Deployment Model

David Black: Assumption that all in PCN domain are in on the PCN. That
domain is then interacting with ECN routers.

PCN is associated with some changes in the ECN field treatment. What
are opportunities if there are mixed ECN and PCN end to end. The model
discussed is an Edge to Edge/controlled. Important to consider these
models when doing the discussion.

See first model slide for this model.

The end system does not need to be trusted.
There is the use of normal RSVP to the end system.

End to end/open deployment model is the other version. Consider mixed
usage model with PCN and ECN.

Questions:

A) Focus on edge to edge/controlled

B) Decide that we want to address both edge to edge and end-to-end.

c) Work out viability of end-to-end/open first?

Are there other models to cover?


Pratik Bose: Will ECN have the same meaning in end to end and
edge-to-edge?

?: Experimental solution for end to end. Use different marking system
and how would they be used? Need input how fast we should go forward.

David Black: Focus on A. Mechanism are used seem to be fine granularity
adjustments that would not be available in end-to-end. Figure out
co-existence with normal ECN (RFC 3168).

Bob Briscoe: Started thinking about the end-to-end case and could not
easily resolve it. That why we started with edge to edge. Requires
reasonable behavior from systems, more likely in network than end
hosts.

Encoding
--------
Presentation of 4 alternatives.

Bob Briscoe: Do people believe this should be done in ECN or in the
DSCP field?

Fran=E7ois: Why would you use DSCP in a controlled environment?
What is the motivation for this?

John Anders: Is okay to use another DSCP codepoint, like alt 1?

David Black: It is no obvious that you need to use another code point
as you could filter them into another traffic class and not handle that
with PCN end to end.

Fran=E7ois: Filter voice into its own DSCP. Have the PCN behavior depend
on the DSCP.

James: Is this a local decision on what to use?

David Black: No, especially for alt 1. One thing that is happening. Not
have the ECN field to depend on the DSCP.

Bob: Use AM control to traffic that handle ECN. Non ECN is static
assignments.

Fran=E7ois: No problem in the controlled environment to use one DSCP.
Down path it might be a bigger problem. Like MPLS. EXP is scarce and
using EXP for this may not be available

Kwok: Pretty open on this issue.

Divide it on DSCP and have that control the ECN behavior:

David Black: As DSCP are operator dependent this can be a problem.
Complexity problem. Would be good to standardize one way to do this.

Fran=E7ois: Good to do this in one way for Hard ware support.

Kwok: Not have ECT encoding within a the field. Question on what is
better to use.

David Black: This is not obvious. Needs to go to the list.

Fran=E7ois: Goal was to get people thinking.

Initial simulation Study
Anna Charny: Limited applicability. It works in the simulation we have
done. Not done a complete verification. This is a proof of concept. And
a demonstration that there is ongoing simulation work.


PCN/RMD
Fran=E7ois: Pre congestion marking: STD or EXP?

John Loughney: Good bring to that into NSIS is supported.
There is a individual draft for integrated service control load:
Will post the draft ref to the TSVWG list.

Bob: Do people think it is enough that PHB and ECN marking is
specified. Do we need to a top level document declaring that
combination that is required to get the service? This would be a one
page.

Restatement

David Black: From process point of view separating ECN from PHB is a
better question. Would provide better clarity.

Anna: A PDB document would need to define a lot of behavior and group
things together.

David Black: It will probably be needed to define how the border
handling is done between domains.

Jozef: Do we need a PDB that defines both the DSCP and ECN.

David black: Sloppy use of PDB. Document needs to define the domain
behavior, not necessarily a Diffserv behavior.


Bob: Are people here for interest, or to actually contribute.

How many people would contribute to this work and review:
14


James
Lars
Magnus
	IETF TSVWG co-chairs




From tsvwg-bounces@ietf.org Fri Apr 07 16:28:58 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRxZe-0006XP-1W; Fri, 07 Apr 2006 16:28:54 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRxZc-0006V7-8Q
	for tsvwg@ietf.org; Fri, 07 Apr 2006 16:28:52 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRxZb-00050f-0E
	for tsvwg@ietf.org; Fri, 07 Apr 2006 16:28:52 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FRxZb-0001K8-1r; Fri, 07 Apr 2006 14:28:29 -0600
Message-ID: <4436CB81.40000@middle.net>
Date: Fri, 07 Apr 2006 14:28:49 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: tsvwg@ietf.org
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: d17f825e43c9aed4fd65b7edddddec89
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>,
	Randall Stewart <rrs@cisco.com>
Subject: [Tsvwg] SCTP PAD chunk
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

I have a comment on draft-tuexen-tsvwg-sctp-padding-00.  Shouldn't the 
draft for this state that this chunk is an exception to the rule that 
control chunks always precede data chunks?   Reordering chunks to pad 
seems rather pointless.

This chunk could be used to advantage in some environments (local 
clusters mostly) to align DATA payloads on 64 bit or potentially larger 
boundaries, in which case it would have to be able to be placed between 
DATA chunks as well.

 - Mark B.




From tsvwg-bounces@ietf.org Fri Apr 07 17:07:00 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRyAP-0007LZ-EN; Fri, 07 Apr 2006 17:06:53 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRyAM-0007JG-DD
	for tsvwg@ietf.org; Fri, 07 Apr 2006 17:06:50 -0400
Received: from sj-iport-4.cisco.com ([171.68.10.86])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRy3Z-0005wa-GX
	for tsvwg@ietf.org; Fri, 07 Apr 2006 16:59:50 -0400
Received: from sj-core-3.cisco.com ([171.68.223.137])
	by sj-iport-4.cisco.com with ESMTP; 07 Apr 2006 13:59:50 -0700
X-IronPort-AV: i="4.04,98,1144047600"; 
	d="scan'208"; a="1792966114:sNHT29998438"
Received: from [IPv6:::1] (ssh-sjc-1.cisco.com [171.68.225.134])
	by sj-core-3.cisco.com (8.12.10/8.12.6) with ESMTP id k37Kxl1j024659;
	Fri, 7 Apr 2006 13:59:48 -0700 (PDT)
Message-ID: <4436D2C3.3090509@cisco.com>
Date: Fri, 07 Apr 2006 16:59:47 -0400
From: Randall Stewart <rrs@cisco.com>
User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060223
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Mark Butler <butlerm@middle.net>
References: <4436CB81.40000@middle.net>
In-Reply-To: <4436CB81.40000@middle.net>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: ea4ac80f790299f943f0a53be7e1a21a
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
Subject: [Tsvwg] Re: SCTP PAD chunk
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark Butler wrote:
> I have a comment on draft-tuexen-tsvwg-sctp-padding-00.  Shouldn't the 
> draft for this state that this chunk is an exception to the rule that 
> control chunks always precede data chunks?   Reordering chunks to pad 
> seems rather pointless.

Well.. its actually a control chunk.. alibeit a fairly useless one :-D

The main purpose for it has nothing to do with data.. but instead
its for P-MTU discovery using (see the PMTU work of Matt Mathis)...

> 
> This chunk could be used to advantage in some environments (local 
> clusters mostly) to align DATA payloads on 64 bit or potentially larger 
> boundaries, in which case it would have to be able to be placed between 
> DATA chunks as well.

Well.. it was not really thought of for that...

R

> 
> - Mark B.
> 


-- 
Randall Stewart
ITD - Cisco Systems Inc.
803-345-0369 <or> 815-342-5222 (cell)




From tsvwg-bounces@ietf.org Fri Apr 07 17:24:34 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRyRT-0004E7-7U; Fri, 07 Apr 2006 17:24:31 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRyRS-0004E2-Gc
	for tsvwg@ietf.org; Fri, 07 Apr 2006 17:24:30 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRyRR-0006eo-6g
	for tsvwg@ietf.org; Fri, 07 Apr 2006 17:24:30 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FRyRR-0001TG-W0; Fri, 07 Apr 2006 15:24:08 -0600
Message-ID: <4436D893.5050804@middle.net>
Date: Fri, 07 Apr 2006 15:24:35 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Randall Stewart <rrs@cisco.com>
References: <4436CB81.40000@middle.net> <4436D2C3.3090509@cisco.com>
In-Reply-To: <4436D2C3.3090509@cisco.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 3e15cc4fdc61d7bce84032741d11c8e5
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
Subject: [Tsvwg] Re: SCTP PAD chunk
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Randall Stewart wrote:

> Mark Butler wrote:
>
>> I have a comment on draft-tuexen-tsvwg-sctp-padding-00.  Shouldn't 
>> the draft for this state that this chunk is an exception to the rule 
>> that control chunks always precede data chunks?   Reordering chunks 
>> to pad seems rather pointless.
>
>
> Well.. its actually a control chunk.. alibeit a fairly useless one :-D
>
> The main purpose for it has nothing to do with data.. but instead
> its for P-MTU discovery using (see the PMTU work of Matt Mathis)...

I am familiar with its application to PMTU discovery.  It just seems 
much more natural to put padding at the end of a packet instead of at 
the beginning.

There are incidental benefits to putting it at the end.  Padding at the 
end improves L2 cache performance by minimizing the amount of useless 
data read.  It saves memory when chunk groups are pre-prepared in memory 
prior to transmission.  It can save memory in some forms of receive 
processing as well.


>>
>> This chunk could be used to advantage in some environments (local 
>> clusters mostly) to align DATA payloads on 64 bit or potentially 
>> larger boundaries, in which case it would have to be able to be 
>> placed between DATA chunks as well.
>
>
> Well.. it was not really thought of for that...

I often hear the argument - "that is not what it was intended for".  Who 
cares?  The proper question should be "what is it good for?". 

Speaking of what things are good for - what is the point of the control 
chunk before data chunk rule anyway?  It has little or no advantages - 
the primary effect of the rule is to increase implementation complexity 
and the number of packets that have to be transmitted.  If I the data 
and control chunks are transmitted into different packets, there is no 
way to guarantee their ordering.  So why care about their ordering if 
they happen to be in the same packet?

- Mark B.












From tsvwg-bounces@ietf.org Fri Apr 07 17:37:40 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRye8-0002xh-0N; Fri, 07 Apr 2006 17:37:36 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRye7-0002xc-0U
	for tsvwg@ietf.org; Fri, 07 Apr 2006 17:37:35 -0400
Received: from sj-iport-4.cisco.com ([171.68.10.86])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRye6-00079T-N4
	for tsvwg@ietf.org; Fri, 07 Apr 2006 17:37:34 -0400
Received: from sj-core-4.cisco.com ([171.68.223.138])
	by sj-iport-4.cisco.com with ESMTP; 07 Apr 2006 14:37:35 -0700
X-IronPort-AV: i="4.04,98,1144047600"; 
	d="scan'208"; a="1792974970:sNHT33003128"
Received: from [IPv6:::1] (ssh-sjc-1.cisco.com [171.68.225.134])
	by sj-core-4.cisco.com (8.12.10/8.12.6) with ESMTP id k37LbWYg009462;
	Fri, 7 Apr 2006 14:37:33 -0700 (PDT)
Message-ID: <4436DB9C.2060503@cisco.com>
Date: Fri, 07 Apr 2006 17:37:32 -0400
From: Randall Stewart <rrs@cisco.com>
User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060223
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Mark Butler <butlerm@middle.net>
References: <4436CB81.40000@middle.net> <4436D2C3.3090509@cisco.com>
	<4436D893.5050804@middle.net>
In-Reply-To: <4436D893.5050804@middle.net>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: a2c12dacc0736f14d6b540e805505a86
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
Subject: [Tsvwg] Re: SCTP PAD chunk
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark Butler wrote:
> Randall Stewart wrote:
> 
>> Mark Butler wrote:
>>
>>> I have a comment on draft-tuexen-tsvwg-sctp-padding-00.  Shouldn't 
>>> the draft for this state that this chunk is an exception to the rule 
>>> that control chunks always precede data chunks?   Reordering chunks 
>>> to pad seems rather pointless.
>>
>>
>>
>> Well.. its actually a control chunk.. alibeit a fairly useless one :-D
>>
>> The main purpose for it has nothing to do with data.. but instead
>> its for P-MTU discovery using (see the PMTU work of Matt Mathis)...
> 
> 
> I am familiar with its application to PMTU discovery.  It just seems 
> much more natural to put padding at the end of a packet instead of at 
> the beginning.

And nothing stops this from occuring if its considered control
--------
HB
--------
PAD-CHK
--------

No data present..

We could go "crazy" and define two chunks.. one considered
"control" and one considered "data"

Or one could put an exclusion in the document saying that it
is neither DATA nor CONTROL so it can appear anywhere
including near other DATA chunks..

But this would probably break a lot of implementations...

I prefer to consider it control..

> 
> There are incidental benefits to putting it at the end.  Padding at the 
> end improves L2 cache performance by minimizing the amount of useless 
> data read.  It saves memory when chunk groups are pre-prepared in memory 
> prior to transmission.  It can save memory in some forms of receive 
> processing as well.

And what stops it from being put at the head or the tail when
sending a HB?



> 
> 
>>>
>>> This chunk could be used to advantage in some environments (local 
>>> clusters mostly) to align DATA payloads on 64 bit or potentially 
>>> larger boundaries, in which case it would have to be able to be 
>>> placed between DATA chunks as well.
>>
>>
>>
>> Well.. it was not really thought of for that...
> 
> 
> I often hear the argument - "that is not what it was intended for".  Who 
> cares?  The proper question should be "what is it good for?".
What it was designed for :-D

PMTU :-D



> Speaking of what things are good for - what is the point of the control 
> chunk before data chunk rule anyway?  It has little or no advantages - 
> the primary effect of the rule is to increase implementation complexity 
> and the number of packets that have to be transmitted.  If I the data 
> and control chunks are transmitted into different packets, there is no 
> way to guarantee their ordering.  So why care about their ordering if 
> they happen to be in the same packet?
>
Good question.. and I can't say as I remember why we did it
that way.. my memory fails me.. I guess its been to long ago :-0


R


-- 
Randall Stewart
ITD - Cisco Systems Inc.
803-345-0369 <or> 815-342-5222 (cell)




From tsvwg-bounces@ietf.org Fri Apr 07 17:41:54 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRyiF-0006SO-9M; Fri, 07 Apr 2006 17:41:51 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRyiD-0006SJ-KW
	for tsvwg@ietf.org; Fri, 07 Apr 2006 17:41:49 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRyi9-0007KF-6X
	for tsvwg@ietf.org; Fri, 07 Apr 2006 17:41:49 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:NmwhEY6BtLlbEK/ON3nECzyTXi7aUdZL@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k37Lfiq21939;
	Fri, 7 Apr 2006 15:41:44 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k37Lfh011137;
	Fri, 7 Apr 2006 15:41:43 -0600
Date: Fri, 7 Apr 2006 15:41:43 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Mark Butler <butlerm@middle.net>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Message-ID: <20060407154143.A11048@openss7.org>
Mail-Followup-To: Mark Butler <butlerm@middle.net>,
	Randall Stewart <rrs@cisco.com>,
	Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
References: <4436CB81.40000@middle.net> <4436D2C3.3090509@cisco.com>
	<4436D893.5050804@middle.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <4436D893.5050804@middle.net>;
	from butlerm@middle.net on Fri, Apr 07, 2006 at 03:24:35PM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: e5ba305d0e64821bf3d8bc5d3bb07228
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>,
	Randall Stewart <rrs@cisco.com>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark,

Mark Butler wrote:                         (Fri, 07 Apr 2006 15:24:35)

-X--snip--X-
> 
> Speaking of what things are good for - what is the point of the control 
> chunk before data chunk rule anyway?  It has little or no advantages - 
> the primary effect of the rule is to increase implementation complexity 
> and the number of packets that have to be transmitted.

Well, I found the rule afforded the reverse: that it decreases implementation
complexity and allows for processing optimizations.  Control chunks other than
SACK are rare during data transfer phase, whereas DATA chunks are not.  Only
one SACK normally occurs in a packet.  Placing the SACK in the first chunks
locates it nicely and a hard loop on DATA chunks can be performed after that.
Placing the SACK ahead of all the DATA chunks allows data buffers in the
retransmission buffer to be cleared before allocating any for the received
data chunks.

Ask yourself the reverse question: what's the good in burying a control chunk
in the middle of a bunch of data chunks?

Why not just make the padding chunk a "data" chunk.

--brian

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




From tsvwg-bounces@ietf.org Fri Apr 07 17:47:49 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRynz-0000BL-2z; Fri, 07 Apr 2006 17:47:47 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRyny-0000BG-H6
	for tsvwg@ietf.org; Fri, 07 Apr 2006 17:47:46 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRyny-0007VO-47
	for tsvwg@ietf.org; Fri, 07 Apr 2006 17:47:46 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:ZrvU7Bp2T0jBzBYWuQtYNwX8OHUwFjYH@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k37Lljq22099;
	Fri, 7 Apr 2006 15:47:45 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k37Llju11183;
	Fri, 7 Apr 2006 15:47:45 -0600
Date: Fri, 7 Apr 2006 15:47:44 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <rrs@cisco.com>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Message-ID: <20060407154744.B11048@openss7.org>
Mail-Followup-To: Randall Stewart <rrs@cisco.com>,
	Mark Butler <butlerm@middle.net>,
	Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
References: <4436CB81.40000@middle.net> <4436D2C3.3090509@cisco.com>
	<4436D893.5050804@middle.net> <4436DB9C.2060503@cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <4436DB9C.2060503@cisco.com>;
	from rrs@cisco.com on Fri, Apr 07, 2006 at 05:37:32PM -0400
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 856eb5f76e7a34990d1d457d8e8e5b7f
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Randall,

Randall Stewart wrote:                                             (Fri, 07 Apr 2006 17:37:32)

-X--snip--X-
> 
> We could go "crazy" and define two chunks.. one considered
> "control" and one considered "data"
> 

Why not just one considered "data"?

--brian

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




From tsvwg-bounces@ietf.org Fri Apr 07 17:57:45 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRyxb-0007si-75; Fri, 07 Apr 2006 17:57:43 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRyxZ-0007sd-T2
	for tsvwg@ietf.org; Fri, 07 Apr 2006 17:57:41 -0400
Received: from test-iport-2.cisco.com ([171.71.176.105])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRyxY-0007ex-Jy
	for tsvwg@ietf.org; Fri, 07 Apr 2006 17:57:41 -0400
Received: from sj-core-1.cisco.com ([171.71.177.237])
	by test-iport-2.cisco.com with ESMTP; 07 Apr 2006 14:57:40 -0700
Received: from [IPv6:::1] (ssh-sjc-1.cisco.com [171.68.225.134])
	by sj-core-1.cisco.com (8.12.10/8.12.6) with ESMTP id k37Lvdw1021924;
	Fri, 7 Apr 2006 14:57:39 -0700 (PDT)
Message-ID: <4436E052.8080005@cisco.com>
Date: Fri, 07 Apr 2006 17:57:38 -0400
From: Randall Stewart <rrs@cisco.com>
User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060223
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: bidulock@openss7.org
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
References: <4436CB81.40000@middle.net> <4436D2C3.3090509@cisco.com>
	<4436D893.5050804@middle.net> <4436DB9C.2060503@cisco.com>
	<20060407154744.B11048@openss7.org>
In-Reply-To: <20060407154744.B11048@openss7.org>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: f607d15ccc2bc4eaf3ade8ffa8af02a0
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Brian F. G. Bidulock wrote:
> Randall,
> 
> Randall Stewart wrote:                                             (Fri, 07 Apr 2006 17:37:32)
> 
> -X--snip--X-
> 
>>We could go "crazy" and define two chunks.. one considered
>>"control" and one considered "data"
>>
> 
> 
> Why not just one considered "data"?

Fine by me.. but the only question I would have is this going
to break any implementations?

I know from inter-op testing that we have throughly tested
the upper bits by placing different chunks in a packet
and sending it off... but I cannot say that we have every
placed:

|control| data| bogus-chunk| data|

In a packet..

And I have a feeling that most implementations, like yours optimize
looking at the control at the beginning... aka handle the SACK etc
and then go into a data-processing loop...  which I think you are
correct.. makes it more efficeient...

The question is, if your data processing loop hits something that is
not CHUNKTYPE == 00

What will you do?

Looking at the BSD implementation real quick... I see that it
would stop processing at the first non-data chunk... its in a
while( ch == DATA) {
}

Optimized much as you stated in your other email..

It does NOT even consider other chunk types... which may not
be optimal.. but it fits the old rules... :-0

That can be fixed of course.. but if other implementations are
like this it imply's a non-backward compatability.. or at least
data chunks following the "pad-data-chunk" would be lost.

R
> 
> --brian
> 


-- 
Randall Stewart
ITD - Cisco Systems Inc.
803-345-0369 <or> 815-342-5222 (cell)




From tsvwg-bounces@ietf.org Fri Apr 07 18:05:12 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRz4o-0003pp-Gc; Fri, 07 Apr 2006 18:05:10 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRz4m-0003nl-Gd
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:05:08 -0400
Received: from mail-n.franken.de ([193.175.24.27] helo=ilsa.franken.de)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRz4k-0007zs-1s
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:05:08 -0400
Received: from [192.168.1.10] (p508F76C5.dip.t-dialin.net [80.143.118.197])
	by ilsa.franken.de (Postfix) with ESMTP
	id 78E0E245CE; Sat,  8 Apr 2006 00:05:03 +0200 (CEST)
	(KNF account authenticated via SMTP-AUTH)
In-Reply-To: <20060407154143.A11048@openss7.org>
References: <4436CB81.40000@middle.net> <4436D2C3.3090509@cisco.com>
	<4436D893.5050804@middle.net> <20060407154143.A11048@openss7.org>
Mime-Version: 1.0 (Apple Message framework v746.2)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <A2242C50-EB96-4E11-BFEB-3A617465B37E@lurchi.franken.de>
Content-Transfer-Encoding: 7bit
From: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Date: Sat, 8 Apr 2006 00:05:10 +0200
To: bidulock@openss7.org
X-Mailer: Apple Mail (2.746.2)
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 41c17b4b16d1eedaa8395c26e9a251c4
Cc: Randall Stewart <rrs@cisco.com>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Dear all,

all SCTP implementations consider only DATA chunks (ie chunks with  
chunk type == 0)
as data chunks and all others as control chunks. This is also what  
RFC 2960 states and
I think we should not change that.

Looking at the point Mark raises of aligning data not only to 32 bits  
but possibly also to
64 or even 128 bits, I think this is not only relevant inside the  
kernel but also on
the kernel/user land. Having in mind that the DATA chunk header is 16  
bytes (128 bits)
the padding can be done at the application layer layer. At the  
transport layer only
an initial padding before the DATA chunks may be required, which is  
possible with the
PAD chunk.

Best regards
Michael

On Apr 7, 2006, at 11:41 PM, Brian F. G. Bidulock wrote:

> Mark,
>
> Mark Butler wrote:                         (Fri, 07 Apr 2006 15:24:35)
>
> -X--snip--X-
>>
>> Speaking of what things are good for - what is the point of the  
>> control
>> chunk before data chunk rule anyway?  It has little or no  
>> advantages -
>> the primary effect of the rule is to increase implementation  
>> complexity
>> and the number of packets that have to be transmitted.
>
> Well, I found the rule afforded the reverse: that it decreases  
> implementation
> complexity and allows for processing optimizations.  Control chunks  
> other than
> SACK are rare during data transfer phase, whereas DATA chunks are  
> not.  Only
> one SACK normally occurs in a packet.  Placing the SACK in the  
> first chunks
> locates it nicely and a hard loop on DATA chunks can be performed  
> after that.
> Placing the SACK ahead of all the DATA chunks allows data buffers  
> in the
> retransmission buffer to be cleared before allocating any for the  
> received
> data chunks.
>
> Ask yourself the reverse question: what's the good in burying a  
> control chunk
> in the middle of a bunch of data chunks?
>
> Why not just make the padding chunk a "data" chunk.
>
> --brian
>
> -- 
> Brian F. G. Bidulock
> bidulock@openss7.org
> http://www.openss7.org/
>





From tsvwg-bounces@ietf.org Fri Apr 07 18:18:34 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRzHj-0004js-QT; Fri, 07 Apr 2006 18:18:31 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRzHi-0004jn-0x
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:18:30 -0400
Received: from mail-n.franken.de ([193.175.24.27] helo=ilsa.franken.de)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRzHh-00008O-Ms
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:18:30 -0400
Received: from [192.168.1.10] (p508F76C5.dip.t-dialin.net [80.143.118.197])
	by ilsa.franken.de (Postfix) with ESMTP
	id 5DFD2245C3; Sat,  8 Apr 2006 00:18:28 +0200 (CEST)
	(KNF account authenticated via SMTP-AUTH)
In-Reply-To: <A2242C50-EB96-4E11-BFEB-3A617465B37E@lurchi.franken.de>
References: <4436CB81.40000@middle.net> <4436D2C3.3090509@cisco.com>
	<4436D893.5050804@middle.net> <20060407154143.A11048@openss7.org>
	<A2242C50-EB96-4E11-BFEB-3A617465B37E@lurchi.franken.de>
Mime-Version: 1.0 (Apple Message framework v746.2)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <D822CC04-D6F1-4C0A-B108-B1878B974D2E@lurchi.franken.de>
Content-Transfer-Encoding: 7bit
From: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Date: Sat, 8 Apr 2006 00:18:35 +0200
To: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>
X-Mailer: Apple Mail (2.746.2)
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 386e0819b1192672467565a524848168
Cc: bidulock@openss7.org, Randall Stewart <rrs@cisco.com>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark,

and if you do not want to do it at the application layer you could  
add parameters
to the INIT/INIT-ACK which both endpoints can use to "negotiate" the  
alignment.
The default is 32 bits. You could negotiate for longer values and  
then just use
a longer chunk padding. This
- gives you the padding you want
- is an extension which is interoperable with older implementations  
(since they
   would report the new parameter as unsupported)

Best regards
Michael

On Apr 8, 2006, at 12:05 AM, Michael Tuexen wrote:

> Dear all,
>
> all SCTP implementations consider only DATA chunks (ie chunks with  
> chunk type == 0)
> as data chunks and all others as control chunks. This is also what  
> RFC 2960 states and
> I think we should not change that.
>
> Looking at the point Mark raises of aligning data not only to 32  
> bits but possibly also to
> 64 or even 128 bits, I think this is not only relevant inside the  
> kernel but also on
> the kernel/user land. Having in mind that the DATA chunk header is  
> 16 bytes (128 bits)
> the padding can be done at the application layer layer. At the  
> transport layer only
> an initial padding before the DATA chunks may be required, which is  
> possible with the
> PAD chunk.
>
> Best regards
> Michael
>
> On Apr 7, 2006, at 11:41 PM, Brian F. G. Bidulock wrote:
>
>> Mark,
>>
>> Mark Butler wrote:                         (Fri, 07 Apr 2006  
>> 15:24:35)
>>
>> -X--snip--X-
>>>
>>> Speaking of what things are good for - what is the point of the  
>>> control
>>> chunk before data chunk rule anyway?  It has little or no  
>>> advantages -
>>> the primary effect of the rule is to increase implementation  
>>> complexity
>>> and the number of packets that have to be transmitted.
>>
>> Well, I found the rule afforded the reverse: that it decreases  
>> implementation
>> complexity and allows for processing optimizations.  Control  
>> chunks other than
>> SACK are rare during data transfer phase, whereas DATA chunks are  
>> not.  Only
>> one SACK normally occurs in a packet.  Placing the SACK in the  
>> first chunks
>> locates it nicely and a hard loop on DATA chunks can be performed  
>> after that.
>> Placing the SACK ahead of all the DATA chunks allows data buffers  
>> in the
>> retransmission buffer to be cleared before allocating any for the  
>> received
>> data chunks.
>>
>> Ask yourself the reverse question: what's the good in burying a  
>> control chunk
>> in the middle of a bunch of data chunks?
>>
>> Why not just make the padding chunk a "data" chunk.
>>
>> --brian
>>
>> -- 
>> Brian F. G. Bidulock
>> bidulock@openss7.org
>> http://www.openss7.org/
>>
>
>
>





From tsvwg-bounces@ietf.org Fri Apr 07 18:23:19 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRzML-0008Mp-59; Fri, 07 Apr 2006 18:23:17 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRzMJ-0008Fq-80
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:23:15 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRzMI-0000CJ-IA
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:23:15 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FRzMJ-0001WH-Fn; Fri, 07 Apr 2006 16:22:53 -0600
Message-ID: <4436E659.90705@middle.net>
Date: Fri, 07 Apr 2006 16:23:21 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Randall Stewart <rrs@cisco.com>
References: <4436CB81.40000@middle.net> <4436D2C3.3090509@cisco.com>
	<4436D893.5050804@middle.net> <4436DB9C.2060503@cisco.com>
In-Reply-To: <4436DB9C.2060503@cisco.com>
Content-Type: multipart/alternative;
	boundary="------------040102090000020508070000"
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.6 (/)
X-Scan-Signature: 155726d2f5fe5eb5c40a9f079fd9e841
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
Subject: [Tsvwg] Re: SCTP PAD chunk
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

This is a multi-part message in MIME format.
--------------040102090000020508070000
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Randall Stewart wrote:

> Mark Butler wrote:
>
>> Randall Stewart wrote:
>>
>>> Mark Butler wrote:
>>>
>>>> I have a comment on draft-tuexen-tsvwg-sctp-padding-00.  Shouldn't 
>>>> the draft for this state that this chunk is an exception to the 
>>>> rule that control chunks always precede data chunks?   Reordering 
>>>> chunks to pad seems rather pointless.
>>>
>>>
>>> Well.. its actually a control chunk.. alibeit a fairly useless one :-D
>>>
>>> The main purpose for it has nothing to do with data.. but instead
>>> its for P-MTU discovery using (see the PMTU work of Matt Mathis)...
>>
>>
>> I am familiar with its application to PMTU discovery.  It just seems 
>> much more natural to put padding at the end of a packet instead of at 
>> the beginning.
>
>
> And nothing stops this from occuring if its considered control
> --------
> HB
> --------
> PAD-CHK
> --------
>
> No data present..
>
> We could go "crazy" and define two chunks.. one considered
> "control" and one considered "data"
>
> Or one could put an exclusion in the document saying that it
> is neither DATA nor CONTROL so it can appear anywhere
> including near other DATA chunks..
>
> But this would probably break a lot of implementations...

Is the "control" or "data" nature of an unknown chunk cast in stone - 
i.e. is it a foregone conclusion that all unknown chunks are control 
chunks?  If not, then any forward compatible implementation must allow 
arbitrary ordering for unknown chunks, and any implementation that does 
not is broken.

>
> I prefer to consider it control..

Why not have a third category: "other"?

>> There are incidental benefits to putting it at the end.  Padding at 
>> the end improves L2 cache performance by minimizing the amount of 
>> useless data read.  It saves memory when chunk groups are 
>> pre-prepared in memory prior to transmission.  It can save memory in 
>> some forms of receive processing as well.
>
>
> And what stops it from being put at the head or the tail when
> sending a HB?

Nothing.  The issue is position relative to DATA chunks, where L2 cache 
trashing is lessened when it is placed after the last DATA chunk, 
instead of prior to the first one.   Contemporary L2 caches typically 
read 128 bytes at a time, so placing the pad chunk at the end reduces 
the cache footprint demanded, because on average the pad chunk cache 
line overlap will be ~1/2 cache line, or 64 bytes of the L2 cache, 
instead of a full 128 if it is in the middle of the packet.

If the purpose of the PAD chunk is only to pad HEARTBEAT containing 
packets, that should be in the draft. Right now it says "This chunk is 
used to pad an SCTP packet to an arbitrary size".  My inclination would 
be to get some mileage out of the overhead, by tacking on data chunks 
more than 1 RTT away from regular transmission where reasonable.

Speaking of which, isn't sending a large PAD mostly packet without 
congestion control a potential problem?  Perhaps there should be some 
kind of rate limit.

>>>>
>>>> This chunk could be used to advantage in some environments (local 
>>>> clusters mostly) to align DATA payloads on 64 bit or potentially 
>>>> larger boundaries, in which case it would have to be able to be 
>>>> placed between DATA chunks as well.
>>>
>>> Well.. it was not really thought of for that...
>>
>> I often hear the argument - "that is not what it was intended for".  
>> Who cares?  The proper question should be "what is it good for?".
>
> What it was designed for :-D
>
> PMTU :-D
>
I had an essentially identical PAD chunk extension intended primarily 
for data alignment purposes on high bandwidth networks.  I think it 
would be silly to have to create a PAD2 chunk.

As far as as PMTU discovery is concerned, I realize why this technique 
can be useful on broken / misconfigured networks, but shouldn't there 
also be a new ICMP packet defined that routers modify along the way with 
the MTU of the path it traverses, and then is echoed back to the sender 
when it reaches the destination?

- Mark B.


--------------040102090000020508070000
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
  <title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
Randall Stewart wrote:<br>
<blockquote cite="mid4436DB9C.2060503@cisco.com" type="cite">Mark
Butler wrote:
  <br>
  <blockquote type="cite">Randall Stewart wrote:
    <br>
    <blockquote type="cite">Mark Butler wrote:
      <br>
      <blockquote type="cite">I have a comment on
draft-tuexen-tsvwg-sctp-padding-00.&nbsp; Shouldn't the draft for this state
that this chunk is an exception to the rule that control chunks always
precede data chunks?&nbsp;&nbsp; Reordering chunks to pad seems rather pointless.
        <br>
      </blockquote>
      <br>
Well.. its actually a control chunk.. alibeit a fairly useless one :-D
      <br>
      <br>
The main purpose for it has nothing to do with data.. but instead
      <br>
its for P-MTU discovery using (see the PMTU work of Matt Mathis)...
      <br>
    </blockquote>
    <br>
I am familiar with its application to PMTU discovery.&nbsp; It just seems
much more natural to put padding at the end of a packet instead of at
the beginning.
    <br>
  </blockquote>
  <br>
And nothing stops this from occuring if its considered control
  <br>
--------
  <br>
HB
  <br>
--------
  <br>
PAD-CHK
  <br>
--------
  <br>
  <br>
No data present..
  <br>
  <br>
We could go "crazy" and define two chunks.. one considered
  <br>
"control" and one considered "data"
  <br>
  <br>
Or one could put an exclusion in the document saying that it
  <br>
is neither DATA nor CONTROL so it can appear anywhere
  <br>
including near other DATA chunks..
  <br>
  <br>
But this would probably break a lot of implementations...
  <br>
</blockquote>
Is the "control" or "data" nature of an unknown chunk cast in stone -
i.e. is it a foregone conclusion that all unknown chunks are control
chunks?&nbsp; If not, then any forward compatible implementation must allow
arbitrary ordering for unknown chunks, and any implementation that does
not is broken.<br>
<br>
<blockquote cite="mid4436DB9C.2060503@cisco.com" type="cite"><br>
I prefer to consider it control..
  <br>
</blockquote>
Why not have a third category: "other"?<br>
<br>
<blockquote cite="mid4436DB9C.2060503@cisco.com" type="cite">
  <blockquote type="cite">There are incidental benefits to putting it
at the end.&nbsp; Padding at the end improves L2 cache performance by
minimizing the amount of useless data read.&nbsp; It saves memory when chunk
groups are pre-prepared in memory prior to transmission.&nbsp; It can save
memory in some forms of receive processing as well.
    <br>
  </blockquote>
  <br>
And what stops it from being put at the head or the tail when
  <br>
sending a HB?
  <br>
</blockquote>
Nothing.&nbsp; The issue is position relative to DATA chunks, where L2 cache
trashing is lessened when it is placed after the last DATA chunk,
instead of prior to the first one.&nbsp;&nbsp; Contemporary L2 caches typically
read 128 bytes at a time, so placing the pad chunk at the end reduces
the cache footprint demanded, because on average the pad chunk cache
line overlap will be ~1/2 cache line, or 64 bytes of the L2 cache,
instead of a full 128 if it is in the middle of the packet.<br>
<br>
If the purpose of the PAD chunk is only to pad HEARTBEAT containing
packets, that should be in the draft. Right now it says "This chunk is
used to <b>pad</b> an <b>SCTP</b> packet to an arbitrary size".&nbsp; My
inclination would be to get some mileage out of the overhead, by
tacking on data chunks more than 1 RTT away from regular transmission
where reasonable.<br>
<br>
Speaking of which, isn't sending a large PAD mostly packet without
congestion control a potential problem?&nbsp; Perhaps there should be some
kind of rate limit.<br>
<blockquote cite="mid4436DB9C.2060503@cisco.com" type="cite">
  <blockquote type="cite">
    <blockquote type="cite">
      <blockquote type="cite"><br>
This chunk could be used to advantage in some environments (local
clusters mostly) to align DATA payloads on 64 bit or potentially larger
boundaries, in which case it would have to be able to be placed between
DATA chunks as well.
        <br>
      </blockquote>
Well.. it was not really thought of for that...
      <br>
    </blockquote>
I often hear the argument - "that is not what it was intended for".&nbsp;
Who cares?&nbsp; The proper question should be "what is it good for?".
    <br>
  </blockquote>
What it was designed for :-D
  <br>
  <br>
PMTU :-D
  <br>
  <br>
</blockquote>
I had an essentially identical PAD chunk extension intended primarily
for data alignment purposes on high bandwidth networks.&nbsp; I think it
would be silly to have to create a PAD2 chunk.<br>
<br>
As far as as PMTU discovery is concerned, I realize why this technique
can be useful on broken / misconfigured networks, but shouldn't there
also be a new ICMP packet defined that routers modify along the way
with the MTU of the path it traverses, and then is echoed back to the
sender when it reaches the destination?<br>
<br>
- Mark B.<br>
<br>
</body>
</html>

--------------040102090000020508070000--




From tsvwg-bounces@ietf.org Fri Apr 07 18:37:48 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRzaF-0000mH-0A; Fri, 07 Apr 2006 18:37:39 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRzaD-0000mC-Sn
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:37:37 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRzaC-0000gD-F9
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:37:37 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:AbGvnMk/QfCFDs7+cs7njhWh3XKKpLXY@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k37MbYq24834;
	Fri, 7 Apr 2006 16:37:34 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k37MbYP12619;
	Fri, 7 Apr 2006 16:37:34 -0600
Date: Fri, 7 Apr 2006 16:37:34 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <rrs@cisco.com>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Message-ID: <20060407163734.B12178@openss7.org>
Mail-Followup-To: Randall Stewart <rrs@cisco.com>,
	Mark Butler <butlerm@middle.net>,
	Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
References: <4436CB81.40000@middle.net> <4436D2C3.3090509@cisco.com>
	<4436D893.5050804@middle.net> <4436DB9C.2060503@cisco.com>
	<20060407154744.B11048@openss7.org> <4436E052.8080005@cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <4436E052.8080005@cisco.com>;
	from rrs@cisco.com on Fri, Apr 07, 2006 at 05:57:38PM -0400
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 386e0819b1192672467565a524848168
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Randall,

Randall Stewart wrote:                                             (Fri, 07 Apr 2006 17:57:38)
> Brian F. G. Bidulock wrote:
> > Randall,
> > 
> > Randall Stewart wrote:                                             (Fri, 07 Apr 2006 17:37:32)
> > 
> > -X--snip--X-
> > 
> >>We could go "crazy" and define two chunks.. one considered
> >>"control" and one considered "data"
> >>
> > 
> > 
> > Why not just one considered "data"?
> 
> Fine by me.. but the only question I would have is this going
> to break any implementations?
> 
> I know from inter-op testing that we have throughly tested
> the upper bits by placing different chunks in a packet
> and sending it off... but I cannot say that we have every
> placed:
> 
> |control| data| bogus-chunk| data|
> 
> In a packet..
> 
> And I have a feeling that most implementations, like yours optimize
> looking at the control at the beginning... aka handle the SACK etc
> and then go into a data-processing loop...  which I think you are
> correct.. makes it more efficeient...
> 
> The question is, if your data processing loop hits something that is
> not CHUNKTYPE == 00
> 
> What will you do?

At the moment?  Report and continue based on the flags in the chunk type.

> 
> Looking at the BSD implementation real quick... I see that it
> would stop processing at the first non-data chunk... its in a
> while( ch == DATA) {
> }

I currently fast-path data in the main loop as:

  while (more chunks) {
	  if (ch == DATA) {
		  if ((rtn = sctp_recv_data()) >= 0)
			  continue;
	  } else switch (ch) {
		  case SACK:
			rtn = sctp_recv_sack();
			break;
		  ...
		  etc.
		  default:
			rtn = sctp_unrec_chunk_type();
			break;
	  }
	  if (rtn < 0)
		  return (rtn);
	  ..blah..
  }

So I automatically continue after control chunks that are marked
continue.  I decided to be forgiving in receiving control chunks
that are in the wrong place.  I always send them before data in
a packet.

> 
> Optimized much as you stated in your other email..
> 
> It does NOT even consider other chunk types... which may not
> be optimal.. but it fits the old rules... :-0
> 
> That can be fixed of course.. but if other implementations are
> like this it imply's a non-backward compatability.. or at least
> data chunks following the "pad-data-chunk" would be lost.

Why not set the report flag so that the sender can know the
behaviour of the receiver?

--brian

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




From tsvwg-bounces@ietf.org Fri Apr 07 18:40:17 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRzcm-0002b5-10; Fri, 07 Apr 2006 18:40:16 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRzck-0002Vs-59
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:40:14 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRzci-0000lO-NU
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:40:14 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:bq5wzmxy+vVcCzbGRg9PYo+scQ7dx5wt@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k37MeBq24904;
	Fri, 7 Apr 2006 16:40:11 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k37MeAl12684;
	Fri, 7 Apr 2006 16:40:10 -0600
Date: Fri, 7 Apr 2006 16:40:10 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Mark Butler <butlerm@middle.net>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Message-ID: <20060407164010.C12178@openss7.org>
Mail-Followup-To: Mark Butler <butlerm@middle.net>,
	Randall Stewart <rrs@cisco.com>,
	Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
References: <4436CB81.40000@middle.net> <4436D2C3.3090509@cisco.com>
	<4436D893.5050804@middle.net> <4436DB9C.2060503@cisco.com>
	<4436E659.90705@middle.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <4436E659.90705@middle.net>;
	from butlerm@middle.net on Fri, Apr 07, 2006 at 04:23:21PM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 93238566e09e6e262849b4f805833007
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>,
	Randall Stewart <rrs@cisco.com>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark,

Mark Butler wrote:                                                    (Fri, 07 Apr 2006 16:23:21)

-X--snip--X-
> 
>    Is  the "control" or "data" nature of an unknown chunk cast in stone -
>    i.e.  is  it a foregone conclusion that all unknown chunks are control
>    chunks?  If not, then any forward compatible implementation must allow
>    arbitrary  ordering  for  unknown  chunks, and any implementation that
>    does not is broken.
> 

I went a little bit farther: following the principle in being strict in what
you send and forgiving in what you receive, I currently process the chunk
wheter it is unrecognized or not, even if it is nested in data.  It is not
efficient (because I expect and fast-path data), but it works.

--brian

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




From tsvwg-bounces@ietf.org Fri Apr 07 18:46:54 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRzj7-0004Za-Th; Fri, 07 Apr 2006 18:46:49 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRzj6-0004ZS-2m
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:46:48 -0400
Received: from sj-iport-2-in.cisco.com ([171.71.176.71]
	helo=sj-iport-2.cisco.com) by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FRzj4-0000rH-Qh
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:46:48 -0400
Received: from sj-core-1.cisco.com ([171.71.177.237])
	by sj-iport-2.cisco.com with ESMTP; 07 Apr 2006 15:46:46 -0700
X-IronPort-AV: i="4.04,100,1144047600"; 
	d="scan'208"; a="319417451:sNHT25531750"
Received: from [IPv6:::1] (ssh-sjc-1.cisco.com [171.68.225.134])
	by sj-core-1.cisco.com (8.12.10/8.12.6) with ESMTP id k37Mkjw1015457;
	Fri, 7 Apr 2006 15:46:45 -0700 (PDT)
Message-ID: <4436EBD5.50804@cisco.com>
Date: Fri, 07 Apr 2006 18:46:45 -0400
From: Randall Stewart <rrs@cisco.com>
User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060223
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: bidulock@openss7.org
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
References: <4436CB81.40000@middle.net> <4436D2C3.3090509@cisco.com>
	<4436D893.5050804@middle.net> <4436DB9C.2060503@cisco.com>
	<4436E659.90705@middle.net> <20060407164010.C12178@openss7.org>
In-Reply-To: <20060407164010.C12178@openss7.org>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: ea4ac80f790299f943f0a53be7e1a21a
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Brian F. G. Bidulock wrote:
> Mark,
> 
> Mark Butler wrote:                                                    (Fri, 07 Apr 2006 16:23:21)
> 
> -X--snip--X-
> 
>>   Is  the "control" or "data" nature of an unknown chunk cast in stone -
>>   i.e.  is  it a foregone conclusion that all unknown chunks are control
>>   chunks?  If not, then any forward compatible implementation must allow
>>   arbitrary  ordering  for  unknown  chunks, and any implementation that
>>   does not is broken.
>>
> 
> 
> I went a little bit farther: following the principle in being strict in what
> you send and forgiving in what you receive, I currently process the chunk
> wheter it is unrecognized or not, even if it is nested in data.  It is not
> efficient (because I expect and fast-path data), but it works.
> 
> --brian
> 
This will be a good set of tests to add for the interop :-D

R

-- 
Randall Stewart
ITD - Cisco Systems Inc.
803-345-0369 <or> 815-342-5222 (cell)




From tsvwg-bounces@ietf.org Fri Apr 07 18:47:38 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRzju-0004pe-EO; Fri, 07 Apr 2006 18:47:38 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRzjs-0004pZ-Tk
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:47:36 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRzjs-0000ro-Dp
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:47:36 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FRzjt-0001YH-8N; Fri, 07 Apr 2006 16:47:15 -0600
Message-ID: <4436EC0F.3020707@middle.net>
Date: Fri, 07 Apr 2006 16:47:43 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: tsvwg@ietf.org
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
References: <4436CB81.40000@middle.net> <4436D2C3.3090509@cisco.com>
	<4436D893.5050804@middle.net> <20060407154143.A11048@openss7.org>
In-Reply-To: <20060407154143.A11048@openss7.org>
Content-Type: multipart/alternative;
	boundary="------------020206020201040800060302"
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.6 (/)
X-Scan-Signature: 68ba2b07ef271dba6ee42a93832cfa4c
Cc: bidulock@openss7.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

This is a multi-part message in MIME format.
--------------020206020201040800060302
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Brian F. G. Bidulock wrote:

>Mark Butler wrote:                         (Fri, 07 Apr 2006 15:24:35)
>
>-X--snip--X-
>  
>
>>Speaking of what things are good for - what is the point of the control 
>>chunk before data chunk rule anyway?  It has little or no advantages - 
>>the primary effect of the rule is to increase implementation complexity 
>>and the number of packets that have to be transmitted.
>>    
>>
>
>Well, I found the rule afforded the reverse: that it decreases implementation
>complexity and allows for processing optimizations.  Control chunks other than
>SACK are rare during data transfer phase, whereas DATA chunks are not.  Only
>one SACK normally occurs in a packet.  Placing the SACK in the first chunks
>locates it nicely and a hard loop on DATA chunks can be performed after that.
>Placing the SACK ahead of all the DATA chunks allows data buffers in the
>retransmission buffer to be cleared before allocating any for the received
>data chunks.
>  
>
You can do it that way (on the send side) without the rule.  What 
benefit is there on the receive side?  Hard coding the definition of 
"control" and "data" chunks on the receive side is likely to break.

>Ask yourself the reverse question: what's the good in burying a control chunk
>in the middle of a bunch of data chunks?
>  
>
Well, if you want 64 bit aligned data payloads, placing 4 byte PAD 
chunks between DATA chunks is one way to achieve it.  It is nice that we 
have 32 bit alignment now, but a decade from now we might look on 32 bit 
alignment the same way we look on 16 bit alignment today.  In some 
contexts (notably IPoIB, 10G Ethernet on 64 bit CPUs / DMA engines) that 
situation already exists.

Other than that I cannot think of any current situations where one would 
normally benefit from having an embedded control chunk, because one 
usually wants control chunks to be sent immediately.  However, tacking 
control chunks on the end of a series of one or more data chunks is 
another story entirely.   If one pre-assembles ~MTU sized series of DATA 
chunks, it is generally much more efficient to tack a SACK or other 
control chunk on the end of the series rather than at the beginning or 
in a separate packet.

>Why not just make the padding chunk a "data" chunk.
>  
>
That would definitely be better than making it a "control" chunk.  I 
prefer "other" though.

 - Mark B.


--------------020206020201040800060302
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
  <title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
Brian F. G. Bidulock wrote:<br>
<blockquote cite="mid20060407154143.A11048@openss7.org" type="cite">
  <pre wrap="">Mark Butler wrote:                         (Fri, 07 Apr 2006 15:24:35)

-X--snip--X-
  </pre>
  <blockquote type="cite">
    <pre wrap="">Speaking of what things are good for - what is the point of the control 
chunk before data chunk rule anyway?  It has little or no advantages - 
the primary effect of the rule is to increase implementation complexity 
and the number of packets that have to be transmitted.
    </pre>
  </blockquote>
  <pre wrap=""><!---->
Well, I found the rule afforded the reverse: that it decreases implementation
complexity and allows for processing optimizations.  Control chunks other than
SACK are rare during data transfer phase, whereas DATA chunks are not.  Only
one SACK normally occurs in a packet.  Placing the SACK in the first chunks
locates it nicely and a hard loop on DATA chunks can be performed after that.
Placing the SACK ahead of all the DATA chunks allows data buffers in the
retransmission buffer to be cleared before allocating any for the received
data chunks.
  </pre>
</blockquote>
You can do it that way (on the send side) without the rule.&nbsp; What
benefit is there on the receive side?&nbsp; Hard coding the definition of
"control" and "data" chunks on the receive side is likely to break.<br>
<blockquote cite="mid20060407154143.A11048@openss7.org" type="cite">
  <pre wrap="">Ask yourself the reverse question: what's the good in burying a control chunk
in the middle of a bunch of data chunks?
  </pre>
</blockquote>
Well, if you want 64 bit aligned data payloads, placing 4 byte PAD
chunks between DATA chunks is one way to achieve it.&nbsp; It is nice that
we have 32 bit alignment now, but a decade from now we might look on 32
bit alignment the same way we look on 16 bit alignment today.&nbsp; In some
contexts (notably IPoIB, 10G Ethernet on 64 bit CPUs / DMA engines)
that situation already exists.<br>
<br>
Other than that I cannot think of any current situations where one
would normally benefit from having an embedded control chunk, because
one usually wants control chunks to be sent immediately.&nbsp; However,
tacking control chunks on the end of a series of one or more data
chunks is another story entirely.&nbsp;&nbsp; If one pre-assembles ~MTU sized
series of DATA chunks, it is generally much more efficient to tack a
SACK or other control chunk on the end of the series rather than at the
beginning or in a separate packet.<br>
<br>
<blockquote cite="mid20060407154143.A11048@openss7.org" type="cite">
  <pre wrap="">Why not just make the padding chunk a "data" chunk.
  </pre>
</blockquote>
That would definitely be better than making it a "control" chunk.&nbsp; I
prefer "other" though.<br>
<br>
&nbsp;- Mark B.<br>
<br>
</body>
</html>

--------------020206020201040800060302--




From tsvwg-bounces@ietf.org Fri Apr 07 18:50:51 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRzmy-0000tl-8Y; Fri, 07 Apr 2006 18:50:48 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRzmw-0000tg-Qd
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:50:46 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRzmv-0000uV-Cn
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:50:46 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:4D0g7dHjMqEp6hE+mV/4CDhaJaucqQUv@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k37Mogq25318;
	Fri, 7 Apr 2006 16:50:42 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k37MofU12910;
	Fri, 7 Apr 2006 16:50:41 -0600
Date: Fri, 7 Apr 2006 16:50:41 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Mark Butler <butlerm@middle.net>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Message-ID: <20060407165041.A12872@openss7.org>
Mail-Followup-To: Mark Butler <butlerm@middle.net>, tsvwg@ietf.org
References: <4436CB81.40000@middle.net> <4436D2C3.3090509@cisco.com>
	<4436D893.5050804@middle.net> <20060407154143.A11048@openss7.org>
	<4436EC0F.3020707@middle.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <4436EC0F.3020707@middle.net>;
	from butlerm@middle.net on Fri, Apr 07, 2006 at 04:47:43PM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: de4f315c9369b71d7dd5909b42224370
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark,

Mark Butler wrote:                                                    (Fri, 07 Apr 2006 16:47:43)
> 
>   
> 
>    Well,  if  you  want  64 bit aligned data payloads, placing 4 byte PAD
>    chunks  between DATA chunks is one way to achieve it.

There are better ways.

--brian

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




From tsvwg-bounces@ietf.org Fri Apr 07 18:52:26 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FRzoX-0002Nb-OG; Fri, 07 Apr 2006 18:52:25 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FRzoW-0002NW-Vs
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:52:24 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRzoV-0000va-HU
	for tsvwg@ietf.org; Fri, 07 Apr 2006 18:52:24 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:3i+c0abNGReV1pauF8scP6Ik/y8a6Yl3@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k37MqMq25356;
	Fri, 7 Apr 2006 16:52:22 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k37MqM812943;
	Fri, 7 Apr 2006 16:52:22 -0600
Date: Fri, 7 Apr 2006 16:52:22 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <rrs@cisco.com>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Message-ID: <20060407165222.A12928@openss7.org>
Mail-Followup-To: Randall Stewart <rrs@cisco.com>,
	Mark Butler <butlerm@middle.net>,
	Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
References: <4436CB81.40000@middle.net> <4436D2C3.3090509@cisco.com>
	<4436D893.5050804@middle.net> <4436DB9C.2060503@cisco.com>
	<4436E659.90705@middle.net> <20060407164010.C12178@openss7.org>
	<4436EBD5.50804@cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <4436EBD5.50804@cisco.com>;
	from rrs@cisco.com on Fri, Apr 07, 2006 at 06:46:45PM -0400
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 21c69d3cfc2dd19218717dbe1d974352
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Randall,

Yes, it would be interesting to see how other implementations respond
and determine the extent of the issue.

--brian


Randall Stewart wrote:                                             (Fri, 07 Apr 2006 18:46:45)
> Brian F. G. Bidulock wrote:
> > Mark,
> > 
> > Mark Butler wrote:                                                    (Fri, 07 Apr 2006 16:23:21)
> > 
> > -X--snip--X-
> > 
> >>   Is  the "control" or "data" nature of an unknown chunk cast in stone -
> >>   i.e.  is  it a foregone conclusion that all unknown chunks are control
> >>   chunks?  If not, then any forward compatible implementation must allow
> >>   arbitrary  ordering  for  unknown  chunks, and any implementation that
> >>   does not is broken.
> >>
> > 
> > 
> > I went a little bit farther: following the principle in being strict in what
> > you send and forgiving in what you receive, I currently process the chunk
> > wheter it is unrecognized or not, even if it is nested in data.  It is not
> > efficient (because I expect and fast-path data), but it works.
> > 
> > --brian
> > 
> This will be a good set of tests to add for the interop :-D
> 
> R
> 
> -- 
> Randall Stewart
> ITD - Cisco Systems Inc.
> 803-345-0369 <or> 815-342-5222 (cell)

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




From tsvwg-bounces@ietf.org Fri Apr 07 19:04:22 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FS002-0001Ep-HR; Fri, 07 Apr 2006 19:04:18 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FS001-0001Ek-8m
	for tsvwg@ietf.org; Fri, 07 Apr 2006 19:04:17 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FRzzz-00019Y-VU
	for tsvwg@ietf.org; Fri, 07 Apr 2006 19:04:17 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FS000-0001ZM-WC; Fri, 07 Apr 2006 17:03:55 -0600
Message-ID: <4436EFF7.7090104@middle.net>
Date: Fri, 07 Apr 2006 17:04:23 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: tsvwg@ietf.org
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
References: <4436CB81.40000@middle.net> <4436D2C3.3090509@cisco.com>
	<4436D893.5050804@middle.net> <20060407154143.A11048@openss7.org>
	<A2242C50-EB96-4E11-BFEB-3A617465B37E@lurchi.franken.de>
In-Reply-To: <A2242C50-EB96-4E11-BFEB-3A617465B37E@lurchi.franken.de>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 9ed51c9d1356100bce94f1ae4ec616a9
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Michael Tuexen wrote:

> Dear all,
>
> all SCTP implementations consider only DATA chunks (ie chunks with  
> chunk type == 0)
> as data chunks and all others as control chunks. This is also what  
> RFC 2960 states and
> I think we should not change that.

It has been recommended here in the past that the proper way to solve 
the the inter-stream TSN dependency problem is to create a new data 
chunk.  This is the type of thing that has to be negotiated at startup 
time, but it is a good example of where the definition of what a data 
chunk is would be extended.

Given that the ordering rule is borderline pointless, it would be much 
easier to just drop it completely, and treat each chunk as if it arrived 
in its own packet, in order of course.

>
> Looking at the point Mark raises of aligning data not only to 32 bits  
> but possibly also to
> 64 or even 128 bits, I think this is not only relevant inside the  
> kernel but also on
> the kernel/user land. Having in mind that the DATA chunk header is 16  
> bytes (128 bits)
> the padding can be done at the application layer layer. At the  
> transport layer only
> an initial padding before the DATA chunks may be required, which is  
> possible with the
> PAD chunk.

The total DATA chunk header is actually 20 bytes, which with a 12 byte 
header size, normally aligns the payload of the first DATA chunk on a 64 
bit boundary.  However, if the first data chunk payload is a multiple of 
eight bytes, the next data chunk payload will not be 64 bit aligned.  So 
a performance oriented ULP with relatively small messages and 64 bit 
aligned message sizes will end up with half of its messages unaligned, 
as the alignment alternates from chunk to chunk.  Interleaved 4 byte PAD 
chunks could be used to fix this problem.

- Mark B.




From tsvwg-bounces@ietf.org Sat Apr 08 00:11:39 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FS4nM-000630-Od; Sat, 08 Apr 2006 00:11:32 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FS4nL-00062k-Qu
	for tsvwg@ietf.org; Sat, 08 Apr 2006 00:11:31 -0400
Received: from hostree9f.alcatel.com ([143.209.238.159]
	helo=audl952.usa.alcatel.com)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FS4nL-0002tI-Gt
	for tsvwg@ietf.org; Sat, 08 Apr 2006 00:11:31 -0400
Received: from axes-mach01.usa.alcatel.com (axes-mach01.usa.alcatel.com
	[10.255.0.20])
	by audl952.usa.alcatel.com (ALCANET) with ESMTP id k384BTGg027004;
	Fri, 7 Apr 2006 23:11:29 -0500
Received: from [10.255.1.166] (ax-sp-pc114.usa.alcatel.com [10.255.1.166])
	by axes-mach01.usa.alcatel.com (8.12.10+Sun/8.12.2) with ESMTP id
	k384F2tL023391; Sat, 8 Apr 2006 09:45:03 +0530 (IST)
Message-ID: <443736E6.6050004@alcatel.com>
Date: Sat, 08 Apr 2006 09:37:02 +0530
From: devayya <devayya.bopaiah@alcatel.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Mark Butler <butlerm@middle.net>
Subject: Re: [Tsvwg] SCTP multi-homing association
References: <44362D55.1040806@alcatel.com>
	<24A3ABE8-1B4B-4127-8AF9-70BEAD16801C@lurchi.franken.de>
	<44365568.2010902@alcatel.com> <44366FF9.7020801@middle.net>
In-Reply-To: <44366FF9.7020801@middle.net>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Scanned-By: MIMEDefang 2.51 on 143.209.238.34
X-Spam-Score: 0.0 (/)
X-Scan-Signature: b5d20af10c334b36874c0264b10f59f1
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Hi Mark,

Please see inline comments.


Mark Butler wrote:

> First of all, can you tell us which OS and SCTP implementation you are 
> using?  If not a common one, 'internal/proprietary' is fine.

I am using Lynx RTOS(version 3.0.0) and the SCTP implementation is done 
inhouse according to the  rfc2960, with some changes from SCTP 
implementors guide version 14.

>    A few comments:
>
> 1. Normally in SCTP, associations do not go active / inactive, paths 
> do. 2. Delayed ACK timer expiry is a normal event that is not an 
> indication of a failure,

Here I was talking about the Upper Layer links(M2PA) going ACTIVE and 
INACTIVE. and also the delay ACK timer is the Upper Layer delay ack 
timer(T7 timer). I had metioned in my first mail that the SCTP 
association remains in ESTABLISHED state, but the Upper Layer link(M2PA) 
goes INACTIVE, probably due to no-delivery of the message from SCTP.
I am sorry, that everybody misunderstood this as SCTP delay ack timer 
expiry. My mistake, I should have been more clear.

> 3. If a DATA chunk is lost, either a later SACK or the T3 timer will 
> eventually trigger a retransmission, which is supposed to be on an 
> alternate path if available.

I agree.

>
> 4.  If a DATA chunk is received but a SACK is lost, and there are no 
> follow on SACKs, the DATA chunk will also be retransmitted in the same 
> manner, but ignored since it was already received.

This could be one of the reason.

> 5. Max.Burst probably doesn't have anything to do with your problem.

I agree.

>
> - Mark B.
>
>
Thanks and regards,
devayya

>>> On Apr 7, 2006, at 11:13 AM, devayya wrote:
>>>
>>>> Hi,
>>>>
>>>> I have a question on SCTP multi-homing,
>>>>
>>>> My SCTP multi-homing test set up is as follows,
>>>> There are two etherenet ports with different ip address are  
>>>> configured for one endpoint(say endpont A). and similarly I have  
>>>> another endpoint (say endpoint B) with two ethernet ports with  
>>>> different ip addresses. The association between two endpoints is  
>>>> ESTABLISHED.
>>>>
>>>> Using a test tool I transmit traffic between these two endpoints.
>>>> With less amount of traffic flowing, if one of the ethernet port 
>>>> IP  address is lost, then the association between these two 
>>>> endpoints  remain ACTIVE and the traffic flows on the second 
>>>> ethernet port.(As  per multi-homing  requirement)
>>>> With increase in traffic flowing, if one of the ethernet port IP  
>>>> address is lost, then the association between these two endpoints  
>>>> goes to INACTIVE. and there is traffic loss. Although the SCTP  
>>>> association is still present, the ULP association goes down, with  
>>>> the reason "delay ACK timer expiry". Can anybody help find out why?
>>>>
>>>> As per SCTP implemtors guide, there is mention of  updation of 
>>>> cwnd  when user tries to transmit DATA, cwnd should be modified 
>>>> depending  on the flightsize and a newly introduced MAX.BURST.(not 
>>>> present in  RFC)
>>>>               if((flightsize + Max.Burst*MTU) < cwnd)
>>>>                    cwnd = flightsize +  Max.Burst*MTU
>>>>
>>>> Is it really needed to implement this MAX.BURST?
>>>> Will addition of this MAX.BURST solve the earlier problem 
>>>> mentioned  by me?
>>>>
>>>> Thanks and regards,
>>>> devayya
>>>>
>>>> P.s: For your reference I have attached a text file of the set up.
>>>>
>




From tsvwg-bounces@ietf.org Mon Apr 10 05:38:35 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FSsqt-0003dU-Co; Mon, 10 Apr 2006 05:38:31 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FSsqs-0003dP-Q2
	for tsvwg@ietf.org; Mon, 10 Apr 2006 05:38:30 -0400
Received: from lizzard.sbs.de ([194.138.37.39])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FSsqr-0000np-6v
	for tsvwg@ietf.org; Mon, 10 Apr 2006 05:38:30 -0400
Received: from mail1.sbs.de (localhost [127.0.0.1])
	by lizzard.sbs.de (8.12.6/8.12.6) with ESMTP id k3A9cOJi028543;
	Mon, 10 Apr 2006 11:38:26 +0200
Received: from fthw9xpa.ww002.siemens.net (fthw9xpa.ww002.siemens.net
	[157.163.133.222])
	by mail1.sbs.de (8.12.6/8.12.6) with ESMTP id k3A9cLW5032568;
	Mon, 10 Apr 2006 11:38:21 +0200
Received: from MCHP7I6A.ww002.siemens.net ([139.25.131.137]) by
	fthw9xpa.ww002.siemens.net with Microsoft SMTPSVC(6.0.3790.1830); 
	Mon, 10 Apr 2006 11:38:21 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Subject: AW: [Tsvwg] Re: SCTP PAD chunk
Date: Mon, 10 Apr 2006 11:38:20 +0200
Message-ID: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [Tsvwg] Re: SCTP PAD chunk
Thread-Index: AcZaj18BMF9a6es2QuWFN3/snwplOQB8XqmQ
From: "Meyknecht, Bernward" <bernward.meyknecht@siemens.com>
To: "Michael Tuexen" <Michael.Tuexen@lurchi.franken.de>,
	"Randall Stewart" <rrs@cisco.com>
X-OriginalArrivalTime: 10 Apr 2006 09:38:21.0032 (UTC)
	FILETIME=[813BC680:01C65C82]
X-Spam-Score: 0.0 (/)
X-Scan-Signature: a2c12dacc0736f14d6b540e805505a86
Cc: bidulock@openss7.org, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org


Hi all,

only one remark: please ensure that the introduction/usage of PAD=20
is compatible with all existing implementations,
thinking e.g. about the thing with control chunks following a
DATA chunk.
Even a test at an interop will only give you a hint, what could
happen, since not all implementations may take part or=20
there are 'older' code versions deployed, not being=20
tested at the interop.

Kind regards,

Bernward


-----Urspr=FCngliche Nachricht-----
Von: Michael Tuexen [mailto:Michael.Tuexen@lurchi.franken.de]=20
Gesendet: Samstag, 8. April 2006 00:05
An: bidulock@openss7.org
Cc: Randall Stewart; tsvwg@ietf.org
Betreff: Re: [Tsvwg] Re: SCTP PAD chunk

Dear all,

all SCTP implementations consider only DATA chunks (ie chunks with =20
chunk type =3D=3D 0)
as data chunks and all others as control chunks. This is also what =20
RFC 2960 states and
I think we should not change that.

Looking at the point Mark raises of aligning data not only to 32 bits =20
but possibly also to
64 or even 128 bits, I think this is not only relevant inside the =20
kernel but also on
the kernel/user land. Having in mind that the DATA chunk header is 16 =20
bytes (128 bits)
the padding can be done at the application layer layer. At the =20
transport layer only
an initial padding before the DATA chunks may be required, which is =20
possible with the
PAD chunk.

Best regards
Michael

On Apr 7, 2006, at 11:41 PM, Brian F. G. Bidulock wrote:

> Mark,
>
> Mark Butler wrote:                         (Fri, 07 Apr 2006 15:24:35)
>
> -X--snip--X-
>>
>> Speaking of what things are good for - what is the point of the =20
>> control
>> chunk before data chunk rule anyway?  It has little or no =20
>> advantages -
>> the primary effect of the rule is to increase implementation =20
>> complexity
>> and the number of packets that have to be transmitted.
>
> Well, I found the rule afforded the reverse: that it decreases =20
> implementation
> complexity and allows for processing optimizations.  Control chunks =20
> other than
> SACK are rare during data transfer phase, whereas DATA chunks are =20
> not.  Only
> one SACK normally occurs in a packet.  Placing the SACK in the =20
> first chunks
> locates it nicely and a hard loop on DATA chunks can be performed =20
> after that.
> Placing the SACK ahead of all the DATA chunks allows data buffers =20
> in the
> retransmission buffer to be cleared before allocating any for the =20
> received
> data chunks.
>
> Ask yourself the reverse question: what's the good in burying a =20
> control chunk
> in the middle of a bunch of data chunks?
>
> Why not just make the padding chunk a "data" chunk.
>
> --brian
>
> --=20
> Brian F. G. Bidulock
> bidulock@openss7.org
> http://www.openss7.org/
>






From tsvwg-bounces@ietf.org Mon Apr 10 05:55:57 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FSt7k-0001YA-5e; Mon, 10 Apr 2006 05:55:56 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FSt7j-0001Y5-14
	for tsvwg@ietf.org; Mon, 10 Apr 2006 05:55:55 -0400
Received: from sj-iport-4.cisco.com ([171.68.10.86])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FSt7h-00013l-MV
	for tsvwg@ietf.org; Mon, 10 Apr 2006 05:55:54 -0400
Received: from sj-core-3.cisco.com ([171.68.223.137])
	by sj-iport-4.cisco.com with ESMTP; 10 Apr 2006 02:55:53 -0700
X-IronPort-AV: i="4.04,107,1144047600"; 
	d="scan'208"; a="1793341563:sNHT33490672"
Received: from [IPv6:::1] (ssh-sjc-1.cisco.com [171.68.225.134])
	by sj-core-3.cisco.com (8.12.10/8.12.6) with ESMTP id k3A9tn1j022216;
	Mon, 10 Apr 2006 02:55:49 -0700 (PDT)
Message-ID: <443A2BA5.4000100@cisco.com>
Date: Mon, 10 Apr 2006 05:55:49 -0400
From: Randall Stewart <rrs@cisco.com>
User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060223
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: "Meyknecht, Bernward" <bernward.meyknecht@siemens.com>
Subject: Re: AW: [Tsvwg] Re: SCTP PAD chunk
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
In-Reply-To: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 1676547e4f33b5e63227e9c02bd359e3
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, bidulock@openss7.org,
	tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Bernward:

Yes, this is a real concern I have about putting a
pad chunk between two data chunks... I am not
to sure that implementations will support this... not
that they should not.. but it is something that we must
use care if we define this new chunk...

If we make it able to go anywhere.. then maybe a restriction
on not sending it unless your peer announces support
for it... with the supported chunk types parameters... this
would then make it so an implemenation would say.. yes I
support this.. and in so doing .. say it may fall anywhere
in the packet.

Michael, what do you think?

R

Meyknecht, Bernward wrote:
> Hi all,
> 
> only one remark: please ensure that the introduction/usage of PAD 
> is compatible with all existing implementations,
> thinking e.g. about the thing with control chunks following a
> DATA chunk.
> Even a test at an interop will only give you a hint, what could
> happen, since not all implementations may take part or 
> there are 'older' code versions deployed, not being 
> tested at the interop.
> 
> Kind regards,
> 
> Bernward
> 
> 
> -----Ursprüngliche Nachricht-----
> Von: Michael Tuexen [mailto:Michael.Tuexen@lurchi.franken.de] 
> Gesendet: Samstag, 8. April 2006 00:05
> An: bidulock@openss7.org
> Cc: Randall Stewart; tsvwg@ietf.org
> Betreff: Re: [Tsvwg] Re: SCTP PAD chunk
> 
> Dear all,
> 
> all SCTP implementations consider only DATA chunks (ie chunks with  
> chunk type == 0)
> as data chunks and all others as control chunks. This is also what  
> RFC 2960 states and
> I think we should not change that.
> 
> Looking at the point Mark raises of aligning data not only to 32 bits  
> but possibly also to
> 64 or even 128 bits, I think this is not only relevant inside the  
> kernel but also on
> the kernel/user land. Having in mind that the DATA chunk header is 16  
> bytes (128 bits)
> the padding can be done at the application layer layer. At the  
> transport layer only
> an initial padding before the DATA chunks may be required, which is  
> possible with the
> PAD chunk.
> 
> Best regards
> Michael
> 
> On Apr 7, 2006, at 11:41 PM, Brian F. G. Bidulock wrote:
> 
> 
>>Mark,
>>
>>Mark Butler wrote:                         (Fri, 07 Apr 2006 15:24:35)
>>
>>-X--snip--X-
>>
>>>Speaking of what things are good for - what is the point of the  
>>>control
>>>chunk before data chunk rule anyway?  It has little or no  
>>>advantages -
>>>the primary effect of the rule is to increase implementation  
>>>complexity
>>>and the number of packets that have to be transmitted.
>>
>>Well, I found the rule afforded the reverse: that it decreases  
>>implementation
>>complexity and allows for processing optimizations.  Control chunks  
>>other than
>>SACK are rare during data transfer phase, whereas DATA chunks are  
>>not.  Only
>>one SACK normally occurs in a packet.  Placing the SACK in the  
>>first chunks
>>locates it nicely and a hard loop on DATA chunks can be performed  
>>after that.
>>Placing the SACK ahead of all the DATA chunks allows data buffers  
>>in the
>>retransmission buffer to be cleared before allocating any for the  
>>received
>>data chunks.
>>
>>Ask yourself the reverse question: what's the good in burying a  
>>control chunk
>>in the middle of a bunch of data chunks?
>>
>>Why not just make the padding chunk a "data" chunk.
>>
>>--brian
>>
>>-- 
>>Brian F. G. Bidulock
>>bidulock@openss7.org
>>http://www.openss7.org/
> 
>>
> 


-- 
Randall Stewart
ITD - Cisco Systems Inc.
803-345-0369 <or> 815-342-5222 (cell)




From tsvwg-bounces@ietf.org Mon Apr 10 06:30:26 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FStf1-0004Xl-Qs; Mon, 10 Apr 2006 06:30:19 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FStf0-0004Xg-R6
	for tsvwg@ietf.org; Mon, 10 Apr 2006 06:30:18 -0400
Received: from gecko.sbs.de ([194.138.37.40])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FStey-00024I-AF
	for tsvwg@ietf.org; Mon, 10 Apr 2006 06:30:18 -0400
Received: from mail1.sbs.de (localhost [127.0.0.1])
	by gecko.sbs.de (8.12.6/8.12.6) with ESMTP id k3AAUDYd007063;
	Mon, 10 Apr 2006 12:30:14 +0200
Received: from fthw9xoa.ww002.siemens.net (fthw9xoa.ww002.siemens.net
	[157.163.133.201])
	by mail1.sbs.de (8.12.6/8.12.6) with ESMTP id k3AAUDLq025997;
	Mon, 10 Apr 2006 12:30:13 +0200
Received: from MCHP7I6A.ww002.siemens.net ([139.25.131.137]) by
	fthw9xoa.ww002.siemens.net with Microsoft SMTPSVC(6.0.3790.1830); 
	Mon, 10 Apr 2006 12:30:13 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Subject: AW: AW: [Tsvwg] Re: SCTP PAD chunk
Date: Mon, 10 Apr 2006 12:30:12 +0200
Message-ID: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC6@MCHP7I6A.ww002.siemens.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: AW: [Tsvwg] Re: SCTP PAD chunk
Thread-Index: AcZchRAD35VuHPWpQzyvbODMm5czQAABD0lQ
From: "Meyknecht, Bernward" <bernward.meyknecht@siemens.com>
To: "Randall Stewart" <rrs@cisco.com>
X-OriginalArrivalTime: 10 Apr 2006 10:30:13.0557 (UTC)
	FILETIME=[C0716650:01C65C89]
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 932cba6e0228cc603da43d861a7e09d8
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, bidulock@openss7.org,
	tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Hi Randy,

usage of the supported chunk types parameter sounds=20
perfect to me and allows for nearly anything=20
with the new PAD...
also for that what Mark likes to have.

Bernie

=20

-----Urspr=FCngliche Nachricht-----
Von: Randall Stewart [mailto:rrs@cisco.com]=20
Gesendet: Montag, 10. April 2006 11:56
An: Meyknecht, Bernward
Cc: Michael Tuexen; bidulock@openss7.org; tsvwg@ietf.org
Betreff: Re: AW: [Tsvwg] Re: SCTP PAD chunk

Bernward:

Yes, this is a real concern I have about putting a
pad chunk between two data chunks... I am not
to sure that implementations will support this... not
that they should not.. but it is something that we must
use care if we define this new chunk...

If we make it able to go anywhere.. then maybe a restriction
on not sending it unless your peer announces support
for it... with the supported chunk types parameters... this
would then make it so an implemenation would say.. yes I
support this.. and in so doing .. say it may fall anywhere
in the packet.

Michael, what do you think?

R

Meyknecht, Bernward wrote:
> Hi all,
>=20
> only one remark: please ensure that the introduction/usage of PAD=20
> is compatible with all existing implementations,
> thinking e.g. about the thing with control chunks following a
> DATA chunk.
> Even a test at an interop will only give you a hint, what could
> happen, since not all implementations may take part or=20
> there are 'older' code versions deployed, not being=20
> tested at the interop.
>=20
> Kind regards,
>=20
> Bernward
>=20
>=20
> -----Urspr=FCngliche Nachricht-----
> Von: Michael Tuexen [mailto:Michael.Tuexen@lurchi.franken.de]=20
> Gesendet: Samstag, 8. April 2006 00:05
> An: bidulock@openss7.org
> Cc: Randall Stewart; tsvwg@ietf.org
> Betreff: Re: [Tsvwg] Re: SCTP PAD chunk
>=20
> Dear all,
>=20
> all SCTP implementations consider only DATA chunks (ie chunks with =20
> chunk type =3D=3D 0)
> as data chunks and all others as control chunks. This is also what =20
> RFC 2960 states and
> I think we should not change that.
>=20
> Looking at the point Mark raises of aligning data not only to 32 bits  =

> but possibly also to
> 64 or even 128 bits, I think this is not only relevant inside the =20
> kernel but also on
> the kernel/user land. Having in mind that the DATA chunk header is 16  =

> bytes (128 bits)
> the padding can be done at the application layer layer. At the =20
> transport layer only
> an initial padding before the DATA chunks may be required, which is =20
> possible with the
> PAD chunk.
>=20
> Best regards
> Michael
>=20
> On Apr 7, 2006, at 11:41 PM, Brian F. G. Bidulock wrote:
>=20
>=20
>>Mark,
>>
>>Mark Butler wrote:                         (Fri, 07 Apr 2006 15:24:35)
>>
>>-X--snip--X-
>>
>>>Speaking of what things are good for - what is the point of the =20
>>>control
>>>chunk before data chunk rule anyway?  It has little or no =20
>>>advantages -
>>>the primary effect of the rule is to increase implementation =20
>>>complexity
>>>and the number of packets that have to be transmitted.
>>
>>Well, I found the rule afforded the reverse: that it decreases =20
>>implementation
>>complexity and allows for processing optimizations.  Control chunks =20
>>other than
>>SACK are rare during data transfer phase, whereas DATA chunks are =20
>>not.  Only
>>one SACK normally occurs in a packet.  Placing the SACK in the =20
>>first chunks
>>locates it nicely and a hard loop on DATA chunks can be performed =20
>>after that.
>>Placing the SACK ahead of all the DATA chunks allows data buffers =20
>>in the
>>retransmission buffer to be cleared before allocating any for the =20
>>received
>>data chunks.
>>
>>Ask yourself the reverse question: what's the good in burying a =20
>>control chunk
>>in the middle of a bunch of data chunks?
>>
>>Why not just make the padding chunk a "data" chunk.
>>
>>--brian
>>
>>--=20
>>Brian F. G. Bidulock
>>bidulock@openss7.org
>>http://www.openss7.org/
>=20
>>
>=20


--=20
Randall Stewart
ITD - Cisco Systems Inc.
803-345-0369 <or> 815-342-5222 (cell)





From tsvwg-bounces@ietf.org Mon Apr 10 10:53:15 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FSxlK-0005qz-Ds; Mon, 10 Apr 2006 10:53:06 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FSxlJ-0005ps-62
	for tsvwg@ietf.org; Mon, 10 Apr 2006 10:53:05 -0400
Received: from athena.hosts.co.uk ([212.84.175.19])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FSxlH-0004gs-SD
	for tsvwg@ietf.org; Mon, 10 Apr 2006 10:53:05 -0400
Received: from [69.138.71.89] (helo=[192.168.1.3])
	by athena.hosts.co.uk with esmtpsa (TLSv1:RC4-SHA:128) (Exim 4.52)
	id 1FSxqs-0006Ak-Uc; Mon, 10 Apr 2006 15:58:51 +0100
In-Reply-To: <4.3.2.7.2.20060407135247.0492d4a8@email.cisco.com>
References: <4.3.2.7.2.20060407135247.0492d4a8@email.cisco.com>
Mime-Version: 1.0 (Apple Message framework v749.3)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <7A6FC5C6-5FD2-40FA-B4E9-FF8F68324078@g11.org.uk>
Content-Transfer-Encoding: 7bit
From: ken carlberg <carlberg@g11.org.uk>
Subject: Re: [Tsvwg] DRAFT TSVWG meeting notes - corrections requested
Date: Mon, 10 Apr 2006 10:52:59 -0400
To: James M. Polk <jmpolk@cisco.com>
X-Mailer: Apple Mail (2.749.3)
X-Spam-Score: -1.8 (-)
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 69a74e02bbee44ab4f8eafdbcedd94a1
Cc: tsvwg <tsvwg@ietf.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

 > Kwok's                                          9:35-9:45  (until  
9:51)
 > draft-chan-tsvwg-diffserv-class-aggr-03         (10 min)
 >
 > - Fred presented
 > - Motivation to provide less granular treatments
 > - Editorial changes
 > - Changed name of default class
 > - Dave McDyson believe this should go to WG status and his network  
will
 >   use this
 > - Dave Blacks thinks this is a good idea and should be a WG item
 > - Hum for WG status (100% pass, 0% against)
 > - Dave Black will be a reviewer
 > - Magnus thinks this ought to have mail-list confirmation for WG  
status

I missed part of the meeting, but in reviewing the minutes, I was  
struck by two for the above points: the hum for WG status followed by  
the thought that we need mail-list confirmation.

why do we need mail-list confirmation for WG status?  it doesn't  
sound like there was any contention at the meeting.  I don't recall  
any contention on the list going back to the Paris-ietf.  Is there an  
informal change of policy within TSVWG in deciding how a draft is  
determined to be a WG item?

Note:  I'll acknowledge that there isn't anything special about  
asking the list for WG status.  And I'd actually prefer to see more  
decisions made on the list rather than waiting for meetings.  But on  
the other hand, over the years I've seen a disturbing trend in  
deferring lots of work/decisions with no strong reasons as to why.   
And it would be nice to move away from this trend.

-ken

ps, I have no vested interest in the aforementioned draft.  I'm just  
concerned about the process.




From tsvwg-bounces@ietf.org Mon Apr 10 11:24:31 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FSyFi-0002eK-ET; Mon, 10 Apr 2006 11:24:30 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FSyFh-0002eF-NI
	for tsvwg@ietf.org; Mon, 10 Apr 2006 11:24:29 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FSyFg-0006Uh-Eg
	for tsvwg@ietf.org; Mon, 10 Apr 2006 11:24:29 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FSyFV-0002eU-GU; Mon, 10 Apr 2006 09:23:55 -0600
Message-ID: <443A78A3.5010508@middle.net>
Date: Mon, 10 Apr 2006 09:24:19 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Randall Stewart <rrs@cisco.com>
Subject: Re: AW: [Tsvwg] Re: SCTP PAD chunk
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
	<443A2BA5.4000100@cisco.com>
In-Reply-To: <443A2BA5.4000100@cisco.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 7655788c23eb79e336f5f8ba8bce7906
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, bidulock@openss7.org,
	tsvwg@ietf.org, "Meyknecht, Bernward" <bernward.meyknecht@siemens.com>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Randall Stewart wrote:

> Bernward:
>
> Yes, this is a real concern I have about putting a
> pad chunk between two data chunks... I am not
> to sure that implementations will support this... not
> that they should not.. but it is something that we must
> use care if we define this new chunk...
>
> If we make it able to go anywhere.. then maybe a restriction
> on not sending it unless your peer announces support
> for it... with the supported chunk types parameters... this
> would then make it so an implemenation would say.. yes I
> support this.. and in so doing .. say it may fall anywhere
> in the packet.

I would prefer a parameter that indicates the sender allows control and 
data chunks to be bundled in any reasonable order, as if they were in 
separate packets.  Any compliant implementation implicitly supports a 
PAD chunk that occurs before any DATA chunks.

 - Mark B.




From tsvwg-bounces@ietf.org Mon Apr 10 13:26:13 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FT09T-0000mB-1E; Mon, 10 Apr 2006 13:26:11 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FT09R-0000m6-SP
	for tsvwg@ietf.org; Mon, 10 Apr 2006 13:26:10 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FT09Q-0002nZ-FQ
	for tsvwg@ietf.org; Mon, 10 Apr 2006 13:26:09 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:vPTxmAnddA7MzlDodya0Giso1416Qf5E@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k3AHQ6q17493;
	Mon, 10 Apr 2006 11:26:06 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k3AHQ5N29994;
	Mon, 10 Apr 2006 11:26:05 -0600
Date: Mon, 10 Apr 2006 11:26:05 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Mark Butler <butlerm@middle.net>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Message-ID: <20060410112605.A29853@openss7.org>
Mail-Followup-To: Mark Butler <butlerm@middle.net>,
	Randall Stewart <rrs@cisco.com>,
	"Meyknecht, Bernward" <bernward.meyknecht@siemens.com>,
	Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
	<443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <443A78A3.5010508@middle.net>;
	from butlerm@middle.net on Mon, Apr 10, 2006 at 09:24:19AM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: a7d6aff76b15f3f56fcb94490e1052e4
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>,
	Randall Stewart <rrs@cisco.com>, tsvwg@ietf.org, "Meyknecht,
	Bernward" <bernward.meyknecht@siemens.com>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark,

While your defining the cool new data chunks that includes fragment
numbers, why not have it specify the length and offset of the data
within the chunk to allow for internal padding.

I really question the use of this ISO like feature.  Everything of
interest (including data) is already padded to a 32-bit alignment
boundary.  Did you have some 128-bit architecture for which that was
insufficient?

--brian

Mark Butler wrote:                             (Mon, 10 Apr 2006 09:24:19)
> Randall Stewart wrote:
> 
> > Bernward:
> >
> > Yes, this is a real concern I have about putting a
> > pad chunk between two data chunks... I am not
> > to sure that implementations will support this... not
> > that they should not.. but it is something that we must
> > use care if we define this new chunk...
> >
> > If we make it able to go anywhere.. then maybe a restriction
> > on not sending it unless your peer announces support
> > for it... with the supported chunk types parameters... this
> > would then make it so an implemenation would say.. yes I
> > support this.. and in so doing .. say it may fall anywhere
> > in the packet.
> 
> I would prefer a parameter that indicates the sender allows control and 
> data chunks to be bundled in any reasonable order, as if they were in 
> separate packets.  Any compliant implementation implicitly supports a 
> PAD chunk that occurs before any DATA chunks.
> 
>  - Mark B.

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




From tsvwg-bounces@ietf.org Mon Apr 10 14:03:28 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FT0jW-0004uw-Ee; Mon, 10 Apr 2006 14:03:26 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FT0jV-0004uL-04
	for tsvwg@ietf.org; Mon, 10 Apr 2006 14:03:25 -0400
Received: from vapor.isi.edu ([128.9.64.64])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FT0ge-0003vj-Ab
	for tsvwg@ietf.org; Mon, 10 Apr 2006 14:00:30 -0400
Received: from [192.168.1.47] (pool-71-106-130-244.lsanca.dsl-w.verizon.net
	[71.106.130.244])
	by vapor.isi.edu (8.11.6p2+0917/8.11.2) with ESMTP id k3AHxhR24250;
	Mon, 10 Apr 2006 10:59:43 -0700 (PDT)
Message-ID: <443A9D08.8000200@isi.edu>
Date: Mon, 10 Apr 2006 10:59:36 -0700
From: Joe Touch <touch@ISI.EDU>
User-Agent: Thunderbird 1.5 (Windows/20051201)
MIME-Version: 1.0
To: ken carlberg <carlberg@g11.org.uk>
Subject: Re: [Tsvwg] DRAFT TSVWG meeting notes - corrections requested
References: <4.3.2.7.2.20060407135247.0492d4a8@email.cisco.com>
	<7A6FC5C6-5FD2-40FA-B4E9-FF8F68324078@g11.org.uk>
In-Reply-To: <7A6FC5C6-5FD2-40FA-B4E9-FF8F68324078@g11.org.uk>
X-Enigmail-Version: 0.94.0.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
X-ISI-4-43-8-MailScanner: Found to be clean
X-MailScanner-From: touch@isi.edu
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 52e1467c2184c31006318542db5614d5
Cc: "James M. Polk" <jmpolk@cisco.com>, tsvwg <tsvwg@ietf.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org



ken carlberg wrote:
>> Kwok's                                          9:35-9:45  (until 9:51)
>> draft-chan-tsvwg-diffserv-class-aggr-03         (10 min)
>>
>> - Fred presented
>> - Motivation to provide less granular treatments
>> - Editorial changes
>> - Changed name of default class
>> - Dave McDyson believe this should go to WG status and his network will
>>   use this
>> - Dave Blacks thinks this is a good idea and should be a WG item
>> - Hum for WG status (100% pass, 0% against)
>> - Dave Black will be a reviewer
>> - Magnus thinks this ought to have mail-list confirmation for WG status
> 
> I missed part of the meeting, but in reviewing the minutes, I was struck
> by two for the above points: the hum for WG status followed by the
> thought that we need mail-list confirmation.
> 
> why do we need mail-list confirmation for WG status?  it doesn't sound
> like there was any contention at the meeting.  I don't recall any
> contention on the list going back to the Paris-ietf.  Is there an
> informal change of policy within TSVWG in deciding how a draft is
> determined to be a WG item?

I thought that no official actions happened as a result of queries at a
meeting; officially it's the list on which WG business is conducted.
This allows those who didn't attend to participate in votes - most of
which are not announced in advance.

(that correspondingly begs the current practice of delaying email list
decisions for WG meetings, though)

Joe




From tsvwg-bounces@ietf.org Mon Apr 10 14:22:41 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FT124-0003uH-BE; Mon, 10 Apr 2006 14:22:36 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FT123-0003uC-0l
	for tsvwg@ietf.org; Mon, 10 Apr 2006 14:22:35 -0400
Received: from athena.hosts.co.uk ([212.84.175.19])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FT121-00053y-OR
	for tsvwg@ietf.org; Mon, 10 Apr 2006 14:22:34 -0400
Received: from [69.138.71.89] (helo=[192.168.1.3])
	by athena.hosts.co.uk with esmtpsa (TLSv1:RC4-SHA:128) (Exim 4.52)
	id 1FT17e-0003yD-4l; Mon, 10 Apr 2006 19:28:22 +0100
In-Reply-To: <443A9D08.8000200@isi.edu>
References: <4.3.2.7.2.20060407135247.0492d4a8@email.cisco.com>
	<7A6FC5C6-5FD2-40FA-B4E9-FF8F68324078@g11.org.uk>
	<443A9D08.8000200@isi.edu>
Mime-Version: 1.0 (Apple Message framework v749.3)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <822BE73B-2BE4-4225-A1A3-370FD1A2AC39@g11.org.uk>
Content-Transfer-Encoding: 7bit
From: ken carlberg <carlberg@g11.org.uk>
Subject: Re: [Tsvwg] DRAFT TSVWG meeting notes - corrections requested
Date: Mon, 10 Apr 2006 14:22:29 -0400
To: Joe Touch <touch@ISI.EDU>
X-Mailer: Apple Mail (2.749.3)
X-Spam-Score: -1.8 (-)
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 93238566e09e6e262849b4f805833007
Cc: "James M. Polk" <jmpolk@cisco.com>, tsvwg <tsvwg@ietf.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

> I thought that no official actions happened as a result of queries  
> at a
> meeting; officially it's the list on which WG business is conducted.
> This allows those who didn't attend to participate in votes - most of
> which are not announced in advance.

yes, in theory, that's the way its supposed to be.  and depending on  
the WG, this course of action was adhered to well in the past.  but a  
snapshot view of WGs today shows that a significant amount of  
official business is done in the meetings -- one can argue that  
"humming" came about because of this change.

> (that correspondingly begs the current practice of delaying email list
> decisions for WG meetings, though)

I have absolutely no qualms about this, and as I mentioned before,  
I'd like to see more official work done on the list.  but if the  
consideration of a draft's status is to be decided/re-affirmed on a  
list, then let's politely nudge the co-chairs to ping the group  
within one week after the meetings so as to take out the delay and  
the last minute work done right before a meeting.

-ken





From tsvwg-bounces@ietf.org Mon Apr 10 15:33:13 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FT28L-0007J5-1X; Mon, 10 Apr 2006 15:33:09 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FT28J-0007J0-RY
	for tsvwg@ietf.org; Mon, 10 Apr 2006 15:33:07 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FT28I-0008CM-IN
	for tsvwg@ietf.org; Mon, 10 Apr 2006 15:33:07 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FT28J-0003OH-0h; Mon, 10 Apr 2006 13:32:45 -0600
Message-ID: <443AB2F8.6000701@middle.net>
Date: Mon, 10 Apr 2006 13:33:12 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: tsvwg@ietf.org
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 9466e0365fc95844abaf7c3f15a05c7d
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>
Subject: [Tsvwg] SCTP UDP encapsulation ports
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org


On the topic of SCTP UDP encapsulation ports, I suggest that for 
practical reasons adopting a convention of using the same port numbers 
for both the inner (SCTP) and outer (UDP) ports would be a good thing 
for the following reasons:

1. Allows multiple userspace SCTP/UDP applications (including server 
applications) to operate simulataneously.
2. Allows the local UDP implementation to handle port allocation and 
security
3. Does not require a centralized proxy process or device.
4. Does not require DNS extensions
5. Allows independent port forwarding on common NAT devices.

The downside is that the SCTP/UDP port space necessarily becomes a 
subset of the UDP port space, which would require different well known 
ports to be adopted for SCTP/UDP applications in some cases. 

A typical use case for userspace SCTP/UDP would be a web browser running 
in an environment without kernel space SCTP support (e.g. Windows) and 
located behind a common NAT device accessing a web server that accepts 
TCP, SCTP/UDP, and SCTP/IP connections, on TCP port 80, UDP port 80, and 
SCTP port 80 respectively.  If there were actually a constituency for 
using UDP port 80 for something else (its hardly useful for HTTP), a 
different well known port would have to be adopted for that case. 

Comments?

 - Mark B.




From tsvwg-bounces@ietf.org Mon Apr 10 18:17:52 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FT4hh-0002pi-AD; Mon, 10 Apr 2006 18:17:49 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FT4hg-0002kf-2V
	for tsvwg@ietf.org; Mon, 10 Apr 2006 18:17:48 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FT4hc-0007xS-PK
	for tsvwg@ietf.org; Mon, 10 Apr 2006 18:17:48 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FT4hd-0003wU-Fu; Mon, 10 Apr 2006 16:17:23 -0600
Message-ID: <443AD990.30106@middle.net>
Date: Mon, 10 Apr 2006 16:17:52 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: tsvwg@ietf.org
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
	<443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
	<20060410112605.A29853@openss7.org>
In-Reply-To: <20060410112605.A29853@openss7.org>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 0bc60ec82efc80c84b8d02f4b0e4de22
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, bidulock@openss7.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Brian F. G. Bidulock wrote:

>Mark,
>
>While your defining the cool new data chunks that includes fragment
>numbers, why not have it specify the length and offset of the data
>within the chunk to allow for internal padding.
>  
>
That would be unnecessarily wasteful for every application, not just the
rare ones that prefer a larger grained alignment.

>I really question the use of this ISO like feature.  Everything of
>interest (including data) is already padded to a 32-bit alignment
>boundary.  Did you have some 128-bit architecture for which that was
>insufficient?
>  
>
Communication bandwidth on 64-bit machines is improved when the data is
64 bit aligned because memcpy and checksum functions are often optimized
for that case.  I doubt there are going to be any true 128-bit
architectures (in the contemporary sense) for several decades, although
some machines might still benefit from larger alignments.

- Mark B.







From tsvwg-bounces@ietf.org Mon Apr 10 19:47:51 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FT66l-0006sy-66; Mon, 10 Apr 2006 19:47:47 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FT66j-0006mv-0i
	for tsvwg@ietf.org; Mon, 10 Apr 2006 19:47:45 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FT66i-0002mz-JI
	for tsvwg@ietf.org; Mon, 10 Apr 2006 19:47:44 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24) id 1FT66j-0004Bj-F4
	for tsvwg@ietf.org; Mon, 10 Apr 2006 17:47:23 -0600
Message-ID: <443AEEA9.9080400@middle.net>
Date: Mon, 10 Apr 2006 17:47:53 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: tsvwg@ietf.org
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
	<443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
	<20060410112605.A29853@openss7.org> <443AD990.30106@middle.net>
	<20060410170359.A11262@openss7.org>
In-Reply-To: <20060410170359.A11262@openss7.org>
Content-Type: multipart/alternative;
	boundary="------------040204060701060504010203"
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.6 (/)
X-Scan-Signature: a2c12dacc0736f14d6b540e805505a86
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

This is a multi-part message in MIME format.
--------------040204060701060504010203
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Brian F. G. Bidulock wrote:

>>Communication bandwidth on 64-bit machines is improved when the data is
>>64 bit aligned because memcpy and checksum functions are often optimized
>>for that case.  I doubt there are going to be any true 128-bit
>>architectures (in the contemporary sense) for several decades, although
>>some machines might still benefit from larger alignments.
>>    
>>
>
>Oh, go on.  memcpy is optimized to transfer the first unaligned bytes and
>then long word copy from there.  The little stagger at the beginning is
>far less complex than adding padding chunks to a protocol (and calculating
>the alignment at the sender).
>  
>
memcpy cannot use 64 bit operations on many 64 bit architectures (say 
IA-64 or Power) unless both source and destination have the same 
relative alignment.  An initial adjustment can 64 bit align either the 
source or the destination pointers, but in general not both.

In any case, this behavior is strictly optional.  Broken network PMTU 
discovery does not need the receiver to recognize a PAD chunk at all.  
It just needs to know packets of a specified size arrive intact, (a 
purpose which would be aided by placing some random data in the PAD 
chunk, or putting the PAD chunk at the start of the packet, and the DATA 
or HEARTBEAT at the end).

I am just saying that if one is going to create a PAD chunk, it ought to 
be general rather than special purpose, and that an option to relax a 
unnecessary ordering requirement is the simplest way to that end.  
Unless someone else expresses interest in such an option, it is not 
likely to get standardized anyway. 

- Mark


--------------040204060701060504010203
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
  <title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
Brian F. G. Bidulock wrote:<br>
<blockquote cite="mid20060410170359.A11262@openss7.org" type="cite">
  <blockquote type="cite">
    <pre wrap="">Communication bandwidth on 64-bit machines is improved when the data is
64 bit aligned because memcpy and checksum functions are often optimized
for that case.  I doubt there are going to be any true 128-bit
architectures (in the contemporary sense) for several decades, although
some machines might still benefit from larger alignments.
    </pre>
  </blockquote>
  <pre wrap=""><!---->
Oh, go on.  memcpy is optimized to transfer the first unaligned bytes and
then long word copy from there.  The little stagger at the beginning is
far less complex than adding padding chunks to a protocol (and calculating
the alignment at the sender).
  </pre>
</blockquote>
memcpy cannot use 64 bit operations on many 64 bit architectures (say
IA-64 or Power) unless both source and destination have the same
relative alignment.&nbsp; An initial adjustment can 64 bit align either the
source or the destination pointers, but in general not both.<br>
<br>
In any case, this behavior is strictly optional.&nbsp; Broken network PMTU
discovery does not need the receiver to recognize a PAD chunk at all.&nbsp;
It just needs to know packets of a specified size arrive intact, (a
purpose which would be aided by placing some random data in the PAD
chunk, or putting the PAD chunk at the start of the packet, and the
DATA or HEARTBEAT at the end).<br>
<br>
I am just saying that if one is going to create a PAD chunk, it ought
to be general rather than special purpose, and that an option to relax
a unnecessary ordering requirement is the simplest way to that end.&nbsp;
Unless someone else expresses interest in such an option, it is not
likely to get standardized anyway.&nbsp; <br>
<br>
- Mark<br>
<br>
</body>
</html>

--------------040204060701060504010203--




From tsvwg-bounces@ietf.org Mon Apr 10 20:16:44 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FT6Yl-0002Vw-KB; Mon, 10 Apr 2006 20:16:43 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FT6Yj-0002Vr-UX
	for tsvwg@ietf.org; Mon, 10 Apr 2006 20:16:41 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FT6Yi-00042t-H7
	for tsvwg@ietf.org; Mon, 10 Apr 2006 20:16:41 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:D88oHMbrNqI+RAdCp+1Ehv2Rl/SaYAX3@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k3B0Gdq26868;
	Mon, 10 Apr 2006 18:16:39 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k3B0Gdp12846;
	Mon, 10 Apr 2006 18:16:39 -0600
Date: Mon, 10 Apr 2006 18:16:39 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Mark Butler <butlerm@middle.net>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Message-ID: <20060410181639.A12819@openss7.org>
Mail-Followup-To: Mark Butler <butlerm@middle.net>, tsvwg@ietf.org
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
	<443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
	<20060410112605.A29853@openss7.org> <443AD990.30106@middle.net>
	<20060410170359.A11262@openss7.org> <443AEEA9.9080400@middle.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <443AEEA9.9080400@middle.net>;
	from butlerm@middle.net on Mon, Apr 10, 2006 at 05:47:53PM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 50a516d93fd399dc60588708fd9a3002
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark,

IA-64 has instructions more suited for copying than reading 64-bit
words and writing them out.  I originally thought you meant a RISC
architecture.  OS's must deal with non-aligned copies all day long:
we don't need to make the protocol worry about it too.  Placing
architecture specific optimizations in the protocol is certainly
not a priority.

--brian

Mark Butler wrote:                                                    (Mon, 10 Apr 2006 17:47:53)
> 
>    Brian F. G. Bidulock wrote:
> 
> Communication bandwidth on 64-bit machines is improved when the data is
> 64 bit aligned because memcpy and checksum functions are often optimized
> for that case.  I doubt there are going to be any true 128-bit
> architectures (in the contemporary sense) for several decades, although
> some machines might still benefit from larger alignments.
>     
> 
> Oh, go on.  memcpy is optimized to transfer the first unaligned bytes and
> then long word copy from there.  The little stagger at the beginning is
> far less complex than adding padding chunks to a protocol (and calculating
> the alignment at the sender).
>   
> 
>    memcpy  cannot use 64 bit operations on many 64 bit architectures (say
>    IA-64  or  Power)  unless  both  source  and destination have the same
>    relative alignment.  An initial adjustment can 64 bit align either the
>    source or the destination pointers, but in general not both.
>    In  any case, this behavior is strictly optional.  Broken network PMTU
>    discovery  does not need the receiver to recognize a PAD chunk at all.
>    It  just  needs  to know packets of a specified size arrive intact, (a
>    purpose  which  would  be aided by placing some random data in the PAD
>    chunk,  or  putting  the PAD chunk at the start of the packet, and the
>    DATA or HEARTBEAT at the end).
>    I  am just saying that if one is going to create a PAD chunk, it ought
>    to be general rather than special purpose, and that an option to relax
>    a  unnecessary  ordering  requirement is the simplest way to that end.
>    Unless  someone  else  expresses interest in such an option, it is not
>    likely to get standardized anyway.
>    - Mark

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




From tsvwg-bounces@ietf.org Mon Apr 10 20:50:06 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FT752-0008Lj-DU; Mon, 10 Apr 2006 20:50:04 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FT752-0008Le-0k
	for tsvwg@ietf.org; Mon, 10 Apr 2006 20:50:04 -0400
Received: from stsc1260-eth-s1-s1p1-vip.va.neustar.com ([156.154.16.129]
	helo=chiedprmail1.ietf.org)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FT5fm-0001p5-5y
	for tsvwg@ietf.org; Mon, 10 Apr 2006 19:19:54 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by chiedprmail1.ietf.org with esmtp (Exim 4.43) id 1FT5QU-00008I-GG
	for tsvwg@ietf.org; Mon, 10 Apr 2006 19:04:08 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:V7+eZ5AQUPUzU+MbnoxJ/YzjxbO/nPUa@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k3AN40q25494;
	Mon, 10 Apr 2006 17:04:00 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k3AN40p11633;
	Mon, 10 Apr 2006 17:04:00 -0600
Date: Mon, 10 Apr 2006 17:03:59 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Mark Butler <butlerm@middle.net>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Message-ID: <20060410170359.A11262@openss7.org>
Mail-Followup-To: Mark Butler <butlerm@middle.net>, tsvwg@ietf.org,
	Michael Tuexen <Michael.Tuexen@lurchi.franken.de>
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
	<443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
	<20060410112605.A29853@openss7.org> <443AD990.30106@middle.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <443AD990.30106@middle.net>;
	from butlerm@middle.net on Mon, Apr 10, 2006 at 04:17:52PM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: -2.6 (--)
X-Scan-Signature: 7baded97d9887f7a0c7e8a33c2e3ea1b
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark,

Mark Butler wrote:                                                    (Mon, 10 Apr 2006 16:17:52)
> Brian F. G. Bidulock wrote:
> 
> >Mark,
> >
> >While your defining the cool new data chunks that includes fragment
> >numbers, why not have it specify the length and offset of the data
> >within the chunk to allow for internal padding.
> >  
> >
> That would be unnecessarily wasteful for every application, not just the
> rare ones that prefer a larger grained alignment.
> 
> >I really question the use of this ISO like feature.  Everything of
> >interest (including data) is already padded to a 32-bit alignment
> >boundary.  Did you have some 128-bit architecture for which that was
> >insufficient?
> >  
> >
> Communication bandwidth on 64-bit machines is improved when the data is
> 64 bit aligned because memcpy and checksum functions are often optimized
> for that case.  I doubt there are going to be any true 128-bit
> architectures (in the contemporary sense) for several decades, although
> some machines might still benefit from larger alignments.

Oh, go on.  memcpy is optimized to transfer the first unaligned bytes and
then long word copy from there.  The little stagger at the beginning is
far less complex than adding padding chunks to a protocol (and calculating
the alignment at the sender).

I say dump the requirement.

--brian

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




From tsvwg-bounces@ietf.org Mon Apr 10 22:23:06 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FT8X2-0007vK-Is; Mon, 10 Apr 2006 22:23:04 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FT8X1-0007t2-CM
	for tsvwg@ietf.org; Mon, 10 Apr 2006 22:23:03 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FT8Wz-0001jS-35
	for tsvwg@ietf.org; Mon, 10 Apr 2006 22:23:03 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FT8Wz-0004Mh-Ap; Mon, 10 Apr 2006 20:22:39 -0600
Message-ID: <443B130D.9020005@middle.net>
Date: Mon, 10 Apr 2006 20:23:09 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: bidulock@openss7.org
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
	<443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
	<20060410112605.A29853@openss7.org> <443AD990.30106@middle.net>
	<20060410170359.A11262@openss7.org> <443AEEA9.9080400@middle.net>
	<20060410181639.A12819@openss7.org>
In-Reply-To: <20060410181639.A12819@openss7.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: ffa9dfbbe7cc58b3fa6b8ae3e57b0aa3
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Brian F. G. Bidulock wrote:

>Mark,
>
>IA-64 has instructions more suited for copying than reading 64-bit
>words and writing them out.  I originally thought you meant a RISC
>architecture.  OS's must deal with non-aligned copies all day long:
>we don't need to make the protocol worry about it too.  Placing
>architecture specific optimizations in the protocol is certainly
>not a priority.
>  
>
Sure they do - and they do it *slower*.  But that is beside the point,  
I am suggesting an option to simplify the protocol.  Perhaps someday it 
might even be the default.  And rather than coming up with a germane 
objection, you are objecting to an obscure implementation possibility 
enabled by, but certainly not required by the option.  It is as classic 
a case of "worse is better" as I have ever seen.  Opposition to all 
changes no matter how inconsequential just because they are changes.  
i.e. we want SCTP to stay Specially Contrained Telecom Protocol, and not 
lift a finger to make it useful for anything else.

 - Mark B.


  ...




From tsvwg-bounces@ietf.org Tue Apr 11 01:22:07 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTBKA-0000UJ-Ou; Tue, 11 Apr 2006 01:21:58 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTBK9-0000TU-7x
	for tsvwg@ietf.org; Tue, 11 Apr 2006 01:21:57 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTBK2-00085w-Q6
	for tsvwg@ietf.org; Tue, 11 Apr 2006 01:21:57 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:/qJ6XZ1j8WIAF6V11s807z5Jj3mPAhp/@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k3B5Lnq00613;
	Mon, 10 Apr 2006 23:21:49 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k3B5LmT17215;
	Mon, 10 Apr 2006 23:21:48 -0600
Date: Mon, 10 Apr 2006 23:21:48 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Mark Butler <butlerm@middle.net>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Message-ID: <20060410232148.A17106@openss7.org>
Mail-Followup-To: Mark Butler <butlerm@middle.net>, tsvwg@ietf.org
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
	<443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
	<20060410112605.A29853@openss7.org> <443AD990.30106@middle.net>
	<20060410170359.A11262@openss7.org> <443AEEA9.9080400@middle.net>
	<20060410181639.A12819@openss7.org> <443B130D.9020005@middle.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <443B130D.9020005@middle.net>;
	from butlerm@middle.net on Mon, Apr 10, 2006 at 08:23:09PM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 50a516d93fd399dc60588708fd9a3002
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark,

Since when does avoiding in a general purpose protocol architecture
specific implementation issues (for specific processors yet) make it
specially constrained for telecom???

If your specific processor doesn't process the protocol fast enough, buy
one that does.  Don't ask for the protocol to change to optimize your
specific processor at the cost of most others.

Perhaps you could consider a network processor that is better suited to
the task.

If it is so important to you, I respectfully suggest you write a I-D.

--brian

Mark Butler wrote:                           (Mon, 10 Apr 2006 20:23:09)
> Brian F. G. Bidulock wrote:
> 
> >Mark,
> >
> >IA-64 has instructions more suited for copying than reading 64-bit
> >words and writing them out.  I originally thought you meant a RISC
> >architecture.  OS's must deal with non-aligned copies all day long:
> >we don't need to make the protocol worry about it too.  Placing
> >architecture specific optimizations in the protocol is certainly
> >not a priority.
> >  
> >
> Sure they do - and they do it *slower*.  But that is beside the point,  
> I am suggesting an option to simplify the protocol.  Perhaps someday it 
> might even be the default.  And rather than coming up with a germane 
> objection, you are objecting to an obscure implementation possibility 
> enabled by, but certainly not required by the option.  It is as classic 
> a case of "worse is better" as I have ever seen.  Opposition to all 
> changes no matter how inconsequential just because they are changes.  
> i.e. we want SCTP to stay Specially Contrained Telecom Protocol, and not 
> lift a finger to make it useful for anything else.
> 
>  - Mark B.
> 
> 
>   ...

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




From tsvwg-bounces@ietf.org Tue Apr 11 03:10:47 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTD11-0008SO-Bd; Tue, 11 Apr 2006 03:10:19 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTD0y-0008Qb-W1
	for tsvwg@ietf.org; Tue, 11 Apr 2006 03:10:16 -0400
Received: from mailgw3.ericsson.se ([193.180.251.60])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTD0x-0003lY-5O
	for tsvwg@ietf.org; Tue, 11 Apr 2006 03:10:16 -0400
Received: from esealmw126.eemea.ericsson.se (unknown [153.88.254.123])
	by mailgw3.ericsson.se (Symantec Mail Security) with ESMTP id 70D7E7CA; 
	Tue, 11 Apr 2006 09:10:14 +0200 (CEST)
Received: from esealmw126.eemea.ericsson.se ([153.88.254.174]) by
	esealmw126.eemea.ericsson.se with Microsoft SMTPSVC(6.0.3790.1830); 
	Tue, 11 Apr 2006 09:10:13 +0200
Received: from [147.214.30.119] ([147.214.30.119]) by
	esealmw126.eemea.ericsson.se with Microsoft SMTPSVC(6.0.3790.1830); 
	Tue, 11 Apr 2006 09:10:13 +0200
Message-ID: <443B5656.9010306@ericsson.com>
Date: Tue, 11 Apr 2006 09:10:14 +0200
From: Magnus Westerlund <magnus.westerlund@ericsson.com>
User-Agent: Thunderbird 1.5 (Windows/20051201)
MIME-Version: 1.0
To: ken carlberg <carlberg@g11.org.uk>
Subject: Re: [Tsvwg] DRAFT TSVWG meeting notes - corrections requested
References: <4.3.2.7.2.20060407135247.0492d4a8@email.cisco.com>	<7A6FC5C6-5FD2-40FA-B4E9-FF8F68324078@g11.org.uk>	<443A9D08.8000200@isi.edu>
	<822BE73B-2BE4-4225-A1A3-370FD1A2AC39@g11.org.uk>
In-Reply-To: <822BE73B-2BE4-4225-A1A3-370FD1A2AC39@g11.org.uk>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-OriginalArrivalTime: 11 Apr 2006 07:10:13.0742 (UTC)
	FILETIME=[FA68B0E0:01C65D36]
X-Brightmail-Tracker: AAAAAA==
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 9ed51c9d1356100bce94f1ae4ec616a9
Cc: "James M. Polk" <jmpolk@cisco.com>, tsvwg <tsvwg@ietf.org>,
	Joe Touch <touch@ISI.EDU>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Hi Ken,

ken carlberg wrote:
>> I thought that no official actions happened as a result of queries at a
>> meeting; officially it's the list on which WG business is conducted.
>> This allows those who didn't attend to participate in votes - most of
>> which are not announced in advance.
> 
> yes, in theory, that's the way its supposed to be.  and depending on the 
> WG, this course of action was adhered to well in the past.  but a 
> snapshot view of WGs today shows that a significant amount of official 
> business is done in the meetings -- one can argue that "humming" came 
> about because of this change.

I think it is important that the decision stays on the mailing list. 
There is a number of WG participants that aren't able to come to the 
meetings and I like to ensure that their possibility to influence things 
aren't reduced.

> 
>> (that correspondingly begs the current practice of delaying email list
>> decisions for WG meetings, though)
> 
> I have absolutely no qualms about this, and as I mentioned before, I'd 
> like to see more official work done on the list.  but if the 
> consideration of a draft's status is to be decided/re-affirmed on a 
> list, then let's politely nudge the co-chairs to ping the group within 
> one week after the meetings so as to take out the delay and the last 
> minute work done right before a meeting.

Yes, I agree. In general this should happen a lot faster. But please 
excuse us this time as we have a new WG chairing team that needs to 
learn and adapt our practices to each other.

Cheers

Magnus Westerlund

Multimedia Technologies, Ericsson Research EAB/TVA/A
----------------------------------------------------------------------
Ericsson AB                | Phone +46 8 4048287
Torshamsgatan 23           | Fax   +46 8 7575550
S-164 80 Stockholm, Sweden | mailto: magnus.westerlund@ericsson.com




From tsvwg-bounces@ietf.org Tue Apr 11 05:06:56 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTEpk-0002Uo-0h; Tue, 11 Apr 2006 05:06:48 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTEpi-0002Ug-7U
	for tsvwg@ietf.org; Tue, 11 Apr 2006 05:06:46 -0400
Received: from ams-iport-1.cisco.com ([144.254.224.140])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTEph-0008Ui-HB
	for tsvwg@ietf.org; Tue, 11 Apr 2006 05:06:46 -0400
Received: from ams-core-1.cisco.com ([144.254.224.150])
	by ams-iport-1.cisco.com with ESMTP; 11 Apr 2006 11:06:46 +0200
Received: from xbh-ams-331.emea.cisco.com (xbh-ams-331.cisco.com
	[144.254.231.71])
	by ams-core-1.cisco.com (8.12.10/8.12.6) with ESMTP id k3B96i36024286
	for <tsvwg@ietf.org>; Tue, 11 Apr 2006 11:06:44 +0200 (MEST)
Received: from xmb-ams-333.cisco.com ([144.254.231.78]) by
	xbh-ams-331.emea.cisco.com with Microsoft SMTPSVC(6.0.3790.0); 
	Tue, 11 Apr 2006 11:06:44 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Subject: RE: [Tsvwg] DRAFT TSVWG meeting notes - corrections requested
Date: Tue, 11 Apr 2006 11:06:42 +0200
Message-ID: <A05118C6DF9320488C77F3D5459B17B702AB8128@xmb-ams-333.emea.cisco.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [Tsvwg] DRAFT TSVWG meeting notes - corrections requested
Thread-Index: AcZadcalLV9k8nE1TpqN2nEMRhEsgACy9kkA
From: "Francois Le Faucheur \(flefauch\)" <flefauch@cisco.com>
To: "James Polk \(jmpolk\)" <jmpolk@cisco.com>, "tsvwg" <tsvwg@ietf.org>
X-OriginalArrivalTime: 11 Apr 2006 09:06:44.0438 (UTC)
	FILETIME=[41303F60:01C65D47]
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 8d89ee9312a95de8ee48d1c94511f1bb
Cc: 
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

James and all,
Below are suggested edits on minutes.
Francois


>>=20
>> Francois's                                    10:25-10:40=20
>> (until 10:53)
>> draft-lefaucheur-emergency-rsvp-00              (20 min)

Above : s/(20 min)/s(15 min)/



>> draft-ietf-tsvwg-rsvp-ipsec
>>=20
>> - RSVP-IPSec
>>        - Removed SPI from what identifies the Reservations
>>        - This made everything here simpler
>>        - Added notion of extended virtual destination port,=20
>> also used in
>>          MPLS

Above: s/MPLS/MPLS TE/


>> - RSVP-Emergency

Add:
"
	    - Draft presents RSVP extensions to identify Emergency sessions and =
give them elevated probability of session establishment
"

>>        - Speaker asked to take this
>>        - Gerald Ash - how is this mech mapped to the RPH
>>        - James answered - that SIP is disconnected from Router
>>          behaviors, and that local policy determines the mapping from
>>          the RPH to RSVP Priority values and treatments

Above:s/and treatments/and treatments (Note that the draft now proposes =
an extension to also convey the RPH inside RSVP signaling)/


>>        - Janet commented - <missed it>
>>        - Georgios asked about if there are too many higher priority
>>          calls, at what point, at what capacity does this=20
>> doc suggest to
>>          terminate calls

Replace previous bullet by:
"
	- Georgios asked, in the case of the new Priority Bypass Model, about =
if there are too many higher priority calls, at what point, at what =
capacity does this doc suggest to stop admitting new emergency calls or =
terminate non-emergency calls
"


>>        - Francois answer that preemption is not discussed

Replace previous bullet by:
"
	- Francois answered that, with the Priority Bypass Model, new emergency =
calls are always accepted and non-emergency calls do not get preempted. =
It is a simple model that works under the assumption that emergency load =
remains reasonable. The other models described in the draft can be used =
to deal with other cases (eg. stop admitting emergency calls at some =
stage).
"

>>        - Allison commented on reorganizing the doc but=20
>> separating the BW
>>          model from the mechanism, as this ID is clearly a=20
>> PS track, but
>>          the model is not - so it can be in an appendix
>>        - Francois to think about that

Replace previous bullet by:
"
       - Francois agreed. This will be reflected in next rev.
"

>>=20
>> Jukka's and Markku's                          10:50-11:00=20
>> (until 11:19)
>> draft-manner-tsvwg-lrsvp-00                      (10 min)
=20
>> - Francois - thinks there is value is documenting proxy behavior, not
>>    clear that the endsystem needs to be the starter of this mechanism
>>    (i.e. a L-RSVP proxy in a early router, more like edge-to-edge)
>>        - Will send comments to the list now

Above: remove "now".


>> =
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>> CL+RMD Bar BoF                           attendance:  13+15+5=3D33
>>=20
>> Chairing James Polk
>>=20
>> Francois on the overall approaches.
>>=20
>> Edge2Edge/Controlled Deployment Model
>>=20
>> David Black: Assumption that all in PCN domain are in on the=20
>> PCN. That
>> domain is then interacting with ECN routers.
>>=20

I think David's point was more to the effect that in =
Edege2Edge/Controlled we can assume that there is no interaction between =
PCN and ECN.
So I would propose something like:
s/That domain is then interacting with ECN routers./Inside the domain =
there is no interaction between PCN and ECN/



>> PCN is associated with some changes in the ECN field treatment. What
>> are opportunities if there are mixed ECN and PCN end to end.=20
>> The model
>> discussed is an Edge to Edge/controlled. Important to consider these
>> models when doing the discussion.
>>=20
>> See first model slide for this model.
>>=20
>> The end system does not need to be trusted.
>> There is the use of normal RSVP to the end system.
>>=20
>> End to end/open deployment model is the other version. Consider mixed
>> usage model with PCN and ECN.
>>=20
>> Questions:
>>=20
>> A) Focus on edge to edge/controlled

s/edge to edge/Edge2Edge/

>>=20
>> B) Decide that we want to address both edge to edge and end-to-end.

Replace by:
"
	B) Decide that we want to address both Edge2Edge/Controlled and =
End2End/Open.
"

>>=20
>> c) Work out viability of end-to-end/open first?
>>=20

s/end-to-end/End2End


>> Are there other models to cover?
>>=20
>>=20
>> Pratik Bose: Will ECN have the same meaning in end to end and
>> edge-to-edge?
>>=20
>> ?: Experimental solution for end to end. Use different marking system
>> and how would they be used? Need input how fast we should go forward.
>>=20
>> David Black: Focus on A. Mechanism are used seem to be fine=20
>> granularity
>> adjustments that would not be available in end-to-end. Figure out
>> co-existence with normal ECN (RFC 3168).
>>=20
>> Bob Briscoe: Started thinking about the end-to-end case and could not
>> easily resolve it. That why we started with edge to edge. Requires
>> reasonable behavior from systems, more likely in network than end
>> hosts.
>>=20
>> Encoding
>> --------
>> Presentation of 4 alternatives.
>>=20
>> Bob Briscoe: Do people believe this should be done in ECN or in the
>> DSCP field?

s/ECN/ECN field/

>>=20
>> Fran=E7ois: Why would you use DSCP in a controlled environment?

s/Why would you use/Why would you not use/

>> What is the motivation for this?
>>=20
>> John Anders: Is okay to use another DSCP codepoint, like alt 1?
>>=20
>> David Black: It is no obvious that you need to use another code point
>> as you could filter them into another traffic class and not=20
>> handle that
>> with PCN end to end.
>>=20
>> Fran=E7ois: Filter voice into its own DSCP. Have the PCN=20
>> behavior depend
>> on the DSCP.
>>=20
>> James: Is this a local decision on what to use?
>>=20
>> David Black: No, especially for alt 1. One thing that is=20
>> happening. Not
>> have the ECN field to depend on the DSCP.
>>=20
>> Bob: Use AM control to traffic that handle ECN. Non ECN is static
>> assignments.
>>=20
>> Fran=E7ois: No problem in the controlled environment to use one DSCP.
>> Down path it might be a bigger problem. Like MPLS. EXP is scarce and
>> using EXP for this may not be available
>>=20
>> Kwok: Pretty open on this issue.
>>=20
>> Divide it on DSCP and have that control the ECN behavior:

I think the above refered to Pratik's question:
I would replace it with something like:
"
Pratik Bose: How about we allow multiple alternatives for encoding and =
operator decides what they do in their network?
"

>>=20
>> David Black: As DSCP are operator dependent this can be a problem.
>> Complexity problem. Would be good to standardize one way to do this.
>>=20
>> Fran=E7ois: Good to do this in one way for Hard ware support.
>>=20
>> Kwok: Not have ECT encoding within a the field. Question on what is
>> better to use.
>>=20
>> David Black: This is not obvious. Needs to go to the list.
>>=20
>> Fran=E7ois: Goal was to get people thinking.
>>=20
>> Initial simulation Study
>> Anna Charny: Limited applicability. It works in the=20
>> simulation we have
>> done. Not done a complete verification. This is a proof of=20
>> concept. And
>> a demonstration that there is ongoing simulation work.
>>=20
>>=20
>> PCN/RMD
>> Fran=E7ois: Pre congestion marking: STD or EXP?
>>=20
>> John Loughney: Good bring to that into NSIS is supported.
>> There is a individual draft for integrated service control load:
>> Will post the draft ref to the TSVWG list.
>>=20
>> Bob: Do people think it is enough that PHB and ECN marking is
>> specified. Do we need to a top level document declaring that
>> combination that is required to get the service? This would be a one
>> page.

Above: s/that PHB and ECN marking is specified/that PHB and PCN marking =
are specified independently/


>>=20
>> Restatement
>>=20
>> David Black: From process point of view separating ECN from PHB is a
>> better question. Would provide better clarity.

s/ECN/PCN/

>>=20
>> Anna: A PDB document would need to define a lot of behavior and group
>> things together.

I think above was a question form Anna (not a statement)




From tsvwg-bounces@ietf.org Tue Apr 11 05:41:46 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTFNU-0008P9-W6; Tue, 11 Apr 2006 05:41:40 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTFNT-0008P4-O8
	for tsvwg@ietf.org; Tue, 11 Apr 2006 05:41:39 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTFNS-0001VU-Eu
	for tsvwg@ietf.org; Tue, 11 Apr 2006 05:41:39 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24) id 1FTFNR-0005Ov-Jz
	for tsvwg@ietf.org; Tue, 11 Apr 2006 03:41:15 -0600
Message-ID: <443B79DE.1020104@middle.net>
Date: Tue, 11 Apr 2006 03:41:50 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: tsvwg@ietf.org
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
	<443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
	<20060410112605.A29853@openss7.org> <443AD990.30106@middle.net>
	<20060410170359.A11262@openss7.org> <443AEEA9.9080400@middle.net>
	<20060410181639.A12819@openss7.org> <443B130D.9020005@middle.net>
	<20060410232148.A17106@openss7.org>
In-Reply-To: <20060410232148.A17106@openss7.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: ea4ac80f790299f943f0a53be7e1a21a
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Brian F. G. Bidulock wrote:

>Mark,
>
>Since when does avoiding in a general purpose protocol architecture
>specific implementation issues (for specific processors yet) make it
>specially constrained for telecom???
>  
>
It doesn't.  It is a the general practice of avoiding all protocol / API 
generalizations, harmonizations, and normalizations - even strictly 
optional ones - for trivial and petty reasons that I find disturbing.  
The contrast to the environment in many working groups is striking.  
With DCCP for example, one gets the impression that they are almost too 
flexible.  It is more like - "that could be useful, lets add an option" 
vs. the "over my dead body" attitude that dominates discussions here, 
amid a background of general indifference - indifference to weaknesses 
both large and small.  The general  impression one gets is that SCTP is 
a victim of premature and zealously guarded fossilization.

>If it is so important to you, I respectfully suggest you write a I-D.
>  
>
I am not inclined to spend time writing I-Ds that are likely to be 
rejected for reasons having little or nothing to do with actual merit.  
The only purpose it might serve would be to reserve a parameter slot or 
two.

 - Mark B.





From tsvwg-bounces@ietf.org Tue Apr 11 05:46:08 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTFRk-0002F5-JC; Tue, 11 Apr 2006 05:46:04 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTFRi-0002F0-SR
	for tsvwg@ietf.org; Tue, 11 Apr 2006 05:46:02 -0400
Received: from ams-iport-1.cisco.com ([144.254.224.140])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTFRi-0001dT-I6
	for tsvwg@ietf.org; Tue, 11 Apr 2006 05:46:02 -0400
Received: from ams-core-1.cisco.com ([144.254.224.150])
	by ams-iport-1.cisco.com with ESMTP; 11 Apr 2006 11:46:02 +0200
Received: from xbh-ams-332.emea.cisco.com (xbh-ams-332.cisco.com
	[144.254.231.87])
	by ams-core-1.cisco.com (8.12.10/8.12.6) with ESMTP id k3B9k138006651; 
	Tue, 11 Apr 2006 11:46:01 +0200 (MEST)
Received: from xmb-ams-333.cisco.com ([144.254.231.78]) by
	xbh-ams-332.emea.cisco.com with Microsoft SMTPSVC(6.0.3790.0); 
	Tue, 11 Apr 2006 11:46:01 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="US-ASCII"
Content-Transfer-Encoding: quoted-printable
Subject: Re: [Tsvwg] WGLC for draft-ietf-tsvwg-vpn-signaled-preemption-00
Date: Tue, 11 Apr 2006 11:45:59 +0200
Message-ID: <A05118C6DF9320488C77F3D5459B17B702AB8173@xmb-ams-333.emea.cisco.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Re: [Tsvwg] WGLC for draft-ietf-tsvwg-vpn-signaled-preemption-00
Thread-Index: AcZdTL0RyZ7A3L/iRWi1wCUrOnSu4g==
From: "Francois Le Faucheur \(flefauch\)" <flefauch@cisco.com>
To: "tsvwg" <tsvwg@ietf.org>, "Fred Baker \(fred\)" <fred@cisco.com>,
	"Bose, Pratik" <pratik.bose@lmco.com>
X-OriginalArrivalTime: 11 Apr 2006 09:46:01.0567 (UTC)
	FILETIME=[BE2596F0:01C65D4C]
X-Spam-Score: 1.1 (+)
X-Scan-Signature: 0fa76816851382eb71b0a882ccdc29ac
Cc: 
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Hello Fred, Pratik,

I think it would be worth doing a few edits to bring this document in
line with the changes in "Generic Aggregate RSVP reservations" (eg.
removal of SPI).=20

For example, section "3.1.1.  Plaintext to Ciphertext Data Flows",
shows:
"
   +-----------------------+    +----------------------+
   | +--------------------+|    |+--------------------+|
   | |RSVP                ||    ||Aggregate RSVP      ||
   | |                    ||    ||                    ||
   | |Per session:        || ID ||Agg. Session        ||
   | |  Destination       ||--->||  Agg. Destination  ||
   | |  Source            ||    ||  Agg. Source=3D self ||
   | |  potential SPI     ||    ||  Agg. SPI generated||
   | |  DSCP             ---------> DSCP              ||
   | |  vPort or protocol---------> vPort             ||
   | |           and port ||    ||                    ||
   | |  Mean rate        ---------> Sum of mean rates ||
   | |  Peak rate        ---------> f(Peak rates)     ||
   | |  Burst Size       ---------> Sum of Burst sizes||
"
I think the "Agg. SPI generated is no longer applicable".



Also, in section 1.3, current text says:
"
Definitions were therefore developed for IPsec [RFC2207] and
[I-D.lefaucheur-rsvp-ipsec] and for more
   generic forms of tunnels [RFC2746].  With the later development of
   the Differentiated Services Architecture [RFC2475], definitions were
   added to specify the DSCP [RFC2474]to be used by a standard RSVP data
   flow in [RFC2996] and to use a pair of IP addresses and a DSCP as the
   identifying information for a data flow [RFC3175].
"
I think this should be replaced by something like:
"
Definitions were therefore developed for IPsec [RFC2207] and for more
   generic forms of tunnels [RFC2746].  With the later development of
   the Differentiated Services Architecture [RFC2475], definitions were
   added to specify the DSCP [RFC2474]to be used by a standard RSVP data
   flow in [RFC2996] and to use a pair of IP addresses and a DSCP as the
   identifying information for a data flow
[RFC3175][I-D.lefaucheur-rsvp-ipsec].
"

I suspect similar edits may be needed in a few other places.


Hope that is useful.


Francois


>> -----Original Message-----
>> From: James Polk (jmpolk)=20
>> Sent: jeudi 6 avril 2006 02:02
>> To: tsvwg
>> Cc: Lars Eggert; magnus.westerlund@ericsson.com
>> Subject: [Tsvwg] WGLC for draft-ietf-tsvwg-vpn-signaled-preemption-00
>>=20
>> TSVWG
>>=20
>> Due to an oversight on my part, I did not send out a=20
>> official notice about=20
>> the WGLC for this document when I announced it in/during the=20
>> TSVWG session=20
>> in Dallas.
>>=20
>> I will therefore let the LC go a week longer, until April 12th, 2006.
>>=20
>> To the authors of this ID, I apologize for the delay.
>>=20




From tsvwg-bounces@ietf.org Tue Apr 11 09:29:57 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTIwK-0006er-Fs; Tue, 11 Apr 2006 09:29:52 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTIwI-0006eX-Tx
	for tsvwg@ietf.org; Tue, 11 Apr 2006 09:29:50 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTIwG-0001pw-Gr
	for tsvwg@ietf.org; Tue, 11 Apr 2006 09:29:50 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:cXk+UKABaY/N2jasm5SASmtH0isJPFb7@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k3BDThq11847;
	Tue, 11 Apr 2006 07:29:43 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k3BDTg023594;
	Tue, 11 Apr 2006 07:29:42 -0600
Date: Tue, 11 Apr 2006 07:29:42 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Mark Butler <butlerm@middle.net>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Message-ID: <20060411072942.A23563@openss7.org>
Mail-Followup-To: Mark Butler <butlerm@middle.net>, tsvwg@ietf.org
References: <443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
	<20060410112605.A29853@openss7.org> <443AD990.30106@middle.net>
	<20060410170359.A11262@openss7.org> <443AEEA9.9080400@middle.net>
	<20060410181639.A12819@openss7.org> <443B130D.9020005@middle.net>
	<20060410232148.A17106@openss7.org> <443B79DE.1020104@middle.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <443B79DE.1020104@middle.net>;
	from butlerm@middle.net on Tue, Apr 11, 2006 at 03:41:50AM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 21c69d3cfc2dd19218717dbe1d974352
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark,

I think you will need to scratch your own itch.  If you are not ready
for the hard work of specification writing and consensus building, you
are likely in the wrong forum.

--brian

Mark Butler wrote:                          (Tue, 11 Apr 2006 03:41:50)
> Brian F. G. Bidulock wrote:
> 
> >Mark,
> >
> >Since when does avoiding in a general purpose protocol architecture
> >specific implementation issues (for specific processors yet) make it
> >specially constrained for telecom???
> >  
> >
> It doesn't.  It is a the general practice of avoiding all protocol / API 
> generalizations, harmonizations, and normalizations - even strictly 
> optional ones - for trivial and petty reasons that I find disturbing.  
> The contrast to the environment in many working groups is striking.  
> With DCCP for example, one gets the impression that they are almost too 
> flexible.  It is more like - "that could be useful, lets add an option" 
> vs. the "over my dead body" attitude that dominates discussions here, 
> amid a background of general indifference - indifference to weaknesses 
> both large and small.  The general  impression one gets is that SCTP is 
> a victim of premature and zealously guarded fossilization.
> 
> >If it is so important to you, I respectfully suggest you write a I-D.
> >  
> >
> I am not inclined to spend time writing I-Ds that are likely to be 
> rejected for reasons having little or nothing to do with actual merit.  
> The only purpose it might serve would be to reserve a parameter slot or 
> two.
> 
>  - Mark B.
> 

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




From tsvwg-bounces@ietf.org Tue Apr 11 13:26:42 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTMdP-00057L-TV; Tue, 11 Apr 2006 13:26:35 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTMdO-00057G-G5
	for tsvwg@ietf.org; Tue, 11 Apr 2006 13:26:34 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTMdN-0002Kd-7T
	for tsvwg@ietf.org; Tue, 11 Apr 2006 13:26:34 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FTMdN-0005jP-6z; Tue, 11 Apr 2006 11:26:11 -0600
Message-ID: <443BE6CE.9040809@middle.net>
Date: Tue, 11 Apr 2006 11:26:38 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: bidulock@openss7.org
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
References: <443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
	<20060410112605.A29853@openss7.org> <443AD990.30106@middle.net>
	<20060410170359.A11262@openss7.org> <443AEEA9.9080400@middle.net>
	<20060410181639.A12819@openss7.org> <443B130D.9020005@middle.net>
	<20060410232148.A17106@openss7.org> <443B79DE.1020104@middle.net>
	<20060411072942.A23563@openss7.org>
In-Reply-To: <20060411072942.A23563@openss7.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 68c8cc8a64a9d0402e43b8eee9fc4199
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Brian F. G. Bidulock wrote:

>I think you will need to scratch your own itch.  If you are not ready
>for the hard work of specification writing and consensus building, you
>are likely in the wrong forum.
>  
>
As if there has been anything discussed here for months that you were in 
favor of... On the contrary you practice little other than arrogant, 
conceited, and insincere indirection, distraction, and disruption.

 - Mark B.




From tsvwg-bounces@ietf.org Tue Apr 11 14:46:23 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTNsY-0003LG-Ne; Tue, 11 Apr 2006 14:46:18 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTNsX-0003LB-1L
	for tsvwg@ietf.org; Tue, 11 Apr 2006 14:46:17 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTNsV-00051y-Px
	for tsvwg@ietf.org; Tue, 11 Apr 2006 14:46:17 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24) id 1FTNsW-0005ug-10
	for tsvwg@ietf.org; Tue, 11 Apr 2006 12:45:54 -0600
Message-ID: <443BF97D.3@middle.net>
Date: Tue, 11 Apr 2006 12:46:21 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: tsvwg@ietf.org
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 2870a44b67ee17965ce5ad0177e150f4
Subject: [Tsvwg] Apologies
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

My apologies for interjecting a decidedly undiplomatic statement of 
personal antagonism into the discussion here.

 - Mark B.




From tsvwg-bounces@ietf.org Tue Apr 11 16:09:33 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTPB4-0004J3-01; Tue, 11 Apr 2006 16:09:30 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTPB2-0004IW-R5
	for tsvwg@ietf.org; Tue, 11 Apr 2006 16:09:28 -0400
Received: from test-iport-1.cisco.com ([171.71.176.117])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTPB1-0000tw-HD
	for tsvwg@ietf.org; Tue, 11 Apr 2006 16:09:28 -0400
Received: from sj-core-2.cisco.com ([171.71.177.254])
	by test-iport-1.cisco.com with ESMTP; 11 Apr 2006 13:09:27 -0700
Received: from xbh-sjc-221.amer.cisco.com (xbh-sjc-221.cisco.com
	[128.107.191.63])
	by sj-core-2.cisco.com (8.12.10/8.12.6) with ESMTP id k3BK9QGv022544;
	Tue, 11 Apr 2006 13:09:26 -0700 (PDT)
Received: from xfe-sjc-212.amer.cisco.com ([171.70.151.187]) by
	xbh-sjc-221.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Tue, 11 Apr 2006 13:09:26 -0700
Received: from jmpolk-wxp.cisco.com ([10.89.16.49]) by
	xfe-sjc-212.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Tue, 11 Apr 2006 13:09:25 -0700
Message-Id: <4.3.2.7.2.20060411150329.03732d00@email.cisco.com>
X-Sender: jmpolk@email.cisco.com
X-Mailer: QUALCOMM Windows Eudora Version 4.3.2
Date: Tue, 11 Apr 2006 15:09:23 -0500
To: Mark Butler <butlerm@middle.net>, tsvwg@ietf.org
From: "James M. Polk" <jmpolk@cisco.com>
Subject: Re: [Tsvwg] Apologies
In-Reply-To: <443BF97D.3@middle.net>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
X-OriginalArrivalTime: 11 Apr 2006 20:09:25.0872 (UTC)
	FILETIME=[D4D99700:01C65DA3]
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 68c8cc8a64a9d0402e43b8eee9fc4199
Cc: 
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

At 12:46 PM 4/11/2006 -0600, Mark Butler wrote:
>My apologies for interjecting a decidedly undiplomatic statement of 
>personal antagonism into the discussion here.

Accepted - but with a comment

Non-technical list flames are not tolerated (here)

Let's all make sure this doesn't occur again, shall we?


>- Mark B.




From tsvwg-bounces@ietf.org Wed Apr 12 06:50:20 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTcvR-0008S1-6V; Wed, 12 Apr 2006 06:50:17 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTcvP-0008Rw-QL
	for tsvwg@ietf.org; Wed, 12 Apr 2006 06:50:15 -0400
Received: from sj-iport-4.cisco.com ([171.68.10.86])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTcvO-0000Li-Gz
	for tsvwg@ietf.org; Wed, 12 Apr 2006 06:50:15 -0400
Received: from sj-core-4.cisco.com ([171.68.223.138])
	by sj-iport-4.cisco.com with ESMTP; 12 Apr 2006 03:50:14 -0700
X-IronPort-AV: i="4.04,114,1144047600"; 
	d="scan'208"; a="1794181829:sNHT29963032"
Received: from [IPv6:::1] (ssh-sjc-1.cisco.com [171.68.225.134])
	by sj-core-4.cisco.com (8.12.10/8.12.6) with ESMTP id k3CAoBYg029993;
	Wed, 12 Apr 2006 03:50:15 -0700 (PDT)
Message-ID: <443CDB61.3030609@cisco.com>
Date: Wed, 12 Apr 2006 06:50:09 -0400
From: Randall Stewart <rrs@cisco.com>
User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060223
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: bidulock@openss7.org
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
	<443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
	<20060410112605.A29853@openss7.org>
In-Reply-To: <20060410112605.A29853@openss7.org>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 25620135586de10c627e3628c432b04a
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org,
	"Meyknecht, Bernward" <bernward.meyknecht@siemens.com>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Brian:

I would ask that Mark go ahead and write his own
draft... If he wants to have some cool features
to have new types of data chunks and special padding..
thats wonderful.. but lets not put it in the base spec.

He can go ahead and write a draft that has a data chunk
that can have pad and all that stuff in it..

Lets leave the PAD chunk for control and what it
was intended.. PMTU probing for Matt Mathis's stuff..

That way this whole thing can be negotiated at startup
that you "support the advanced data chunk" and then
the implementations that support this can use it instead
of the normal data chunk... If he is defining a new
data chunk he can surely add a "pad bytes" field to it
to indicate the number of bytes of trailing pad..

Thus a control chunk pad.. and this new data chunk is
all we need to handle all padding scenarios :-D

R


Brian F. G. Bidulock wrote:
> Mark,
> 
> While your defining the cool new data chunks that includes fragment
> numbers, why not have it specify the length and offset of the data
> within the chunk to allow for internal padding.
> 
> I really question the use of this ISO like feature.  Everything of
> interest (including data) is already padded to a 32-bit alignment
> boundary.  Did you have some 128-bit architecture for which that was
> insufficient?
> 
> --brian
> 
> Mark Butler wrote:                             (Mon, 10 Apr 2006 09:24:19)
> 
>>Randall Stewart wrote:
>>
>>
>>>Bernward:
>>>
>>>Yes, this is a real concern I have about putting a
>>>pad chunk between two data chunks... I am not
>>>to sure that implementations will support this... not
>>>that they should not.. but it is something that we must
>>>use care if we define this new chunk...
>>>
>>>If we make it able to go anywhere.. then maybe a restriction
>>>on not sending it unless your peer announces support
>>>for it... with the supported chunk types parameters... this
>>>would then make it so an implemenation would say.. yes I
>>>support this.. and in so doing .. say it may fall anywhere
>>>in the packet.
>>
>>I would prefer a parameter that indicates the sender allows control and 
>>data chunks to be bundled in any reasonable order, as if they were in 
>>separate packets.  Any compliant implementation implicitly supports a 
>>PAD chunk that occurs before any DATA chunks.
>>
>> - Mark B.
> 
> 


-- 
Randall Stewart
ITD - Cisco Systems Inc.
803-345-0369 <or> 815-342-5222 (cell)




From tsvwg-bounces@ietf.org Wed Apr 12 07:32:19 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTda5-0003fk-Uo; Wed, 12 Apr 2006 07:32:17 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTda4-0003ev-8E
	for tsvwg@ietf.org; Wed, 12 Apr 2006 07:32:16 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTda2-0001d5-OZ
	for tsvwg@ietf.org; Wed, 12 Apr 2006 07:32:16 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:1FFmrW+uSdqbz7wibG/x/QRRkBiGF5sR@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k3CBWDq07305;
	Wed, 12 Apr 2006 05:32:13 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k3CBWCn08475;
	Wed, 12 Apr 2006 05:32:12 -0600
Date: Wed, 12 Apr 2006 05:32:12 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <rrs@cisco.com>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Message-ID: <20060412053212.A8295@openss7.org>
Mail-Followup-To: Randall Stewart <rrs@cisco.com>,
	Mark Butler <butlerm@middle.net>,
	"Meyknecht, Bernward" <bernward.meyknecht@siemens.com>,
	Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
	<443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
	<20060410112605.A29853@openss7.org> <443CDB61.3030609@cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <443CDB61.3030609@cisco.com>;
	from rrs@cisco.com on Wed, Apr 12, 2006 at 06:50:09AM -0400
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 944ecb6e61f753561f559a497458fb4f
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org,
	"Meyknecht, Bernward" <bernward.meyknecht@siemens.com>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Randall,

Yes, I agree.  The protocol is quite extensible and it is not
all that difficult to write an extension draft; however, I also
agree with Mark that writing a draft without some interest
within the WG might be a wasted effort.

Also, I am concerned with the base protocol in that it appears
that, as Mark put it, all future chunks will be considered
control chunks by a base implementation because they do not have
the precise tag value of a DATA chunk.  This is not so good for
future extensibility.  This is one of the considerations I made
when I changed my loop to not just check for DATA chunks.  If
there is an unrecognized chunk in the sequence of DATA chunks,
it should be reported and aborted or continued in accordance
with the bits in the chunk type.  That is, unrecognized chunk
types should be considered as possibly being of a "DATA" type.

I think that the passage should be considered for revision in
the 2960bis document, if nothing more, but to pave the way for
future DATA-chunk-like extensions.

Are there comments on altering the passage in question?

--brian


Randall Stewart wrote:                         (Wed, 12 Apr 2006 06:50:09)
> Brian:
> 
> I would ask that Mark go ahead and write his own
> draft... If he wants to have some cool features
> to have new types of data chunks and special padding..
> thats wonderful.. but lets not put it in the base spec.
> 
> He can go ahead and write a draft that has a data chunk
> that can have pad and all that stuff in it..
> 
> Lets leave the PAD chunk for control and what it
> was intended.. PMTU probing for Matt Mathis's stuff..
> 
> That way this whole thing can be negotiated at startup
> that you "support the advanced data chunk" and then
> the implementations that support this can use it instead
> of the normal data chunk... If he is defining a new
> data chunk he can surely add a "pad bytes" field to it
> to indicate the number of bytes of trailing pad..
> 
> Thus a control chunk pad.. and this new data chunk is
> all we need to handle all padding scenarios :-D
> 
> R
> 
> 
> Brian F. G. Bidulock wrote:
> > Mark,
> > 
> > While your defining the cool new data chunks that includes fragment
> > numbers, why not have it specify the length and offset of the data
> > within the chunk to allow for internal padding.
> > 
> > I really question the use of this ISO like feature.  Everything of
> > interest (including data) is already padded to a 32-bit alignment
> > boundary.  Did you have some 128-bit architecture for which that was
> > insufficient?
> > 
> > --brian
> > 
> > Mark Butler wrote:                             (Mon, 10 Apr 2006 09:24:19)
> > 
> >>Randall Stewart wrote:
> >>
> >>
> >>>Bernward:
> >>>
> >>>Yes, this is a real concern I have about putting a
> >>>pad chunk between two data chunks... I am not
> >>>to sure that implementations will support this... not
> >>>that they should not.. but it is something that we must
> >>>use care if we define this new chunk...
> >>>
> >>>If we make it able to go anywhere.. then maybe a restriction
> >>>on not sending it unless your peer announces support
> >>>for it... with the supported chunk types parameters... this
> >>>would then make it so an implemenation would say.. yes I
> >>>support this.. and in so doing .. say it may fall anywhere
> >>>in the packet.
> >>
> >>I would prefer a parameter that indicates the sender allows control and 
> >>data chunks to be bundled in any reasonable order, as if they were in 
> >>separate packets.  Any compliant implementation implicitly supports a 
> >>PAD chunk that occurs before any DATA chunks.
> >>
> >> - Mark B.
> > 
> > 
> 
> 
> -- 
> Randall Stewart
> ITD - Cisco Systems Inc.
> 803-345-0369 <or> 815-342-5222 (cell)

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




From tsvwg-bounces@ietf.org Wed Apr 12 08:15:58 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTeGH-0005Io-PS; Wed, 12 Apr 2006 08:15:53 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTeG6-000520-0z
	for tsvwg@ietf.org; Wed, 12 Apr 2006 08:15:42 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTeDB-0003KI-KV
	for tsvwg@ietf.org; Wed, 12 Apr 2006 08:12:42 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:xtSEw2P60YOOOk/8oIDLu3o84m+AAL6P@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k3CCCeq08749;
	Wed, 12 Apr 2006 06:12:40 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k3CCCeu09806;
	Wed, 12 Apr 2006 06:12:40 -0600
Date: Wed, 12 Apr 2006 06:12:39 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: "Meyknecht, Bernward" <bernward.meyknecht@siemens.com>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Message-ID: <20060412061239.A9084@openss7.org>
Mail-Followup-To: "Meyknecht, Bernward" <bernward.meyknecht@siemens.com>,
	Randall Stewart <rrs@cisco.com>, Mark Butler <butlerm@middle.net>,
	Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BCE@MCHP7I6A.ww002.siemens.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BCE@MCHP7I6A.ww002.siemens.net>;
	from bernward.meyknecht@siemens.com on Wed, Apr 12, 2006 at 02:05:26PM
	+0200
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: b19722fc8d3865b147c75ae2495625f2
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>,
	Randall Stewart <rrs@cisco.com>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Meyknecht,,

Meyknecht, Bernward wrote:                                                      (Wed, 12 Apr 2006 14:05:26)
> 
> Hi all,
> 
> concerning the altered passage I can only repeat my previous remark
> on compatibility to existing and deployed implementation,
> though I like Brians idea.
> 
> Maybe it's good to follow the approach from Randy with the supported
> chunk types parameter, and we should still be free to say, that any chunk 
> type that is presented in the supported chunk types parameter,
> is allowed to appear at any reasonable position within the SCTP
> packet. Would that be something to follow?
> 
> Of course 'reasonable' may include some rules concerning association
> establishment or shutdown.

Yes, that approach would address concern about future extensions, in that
the extension itself could define further placement within the packet.

--brian

> 
> Kind regards,
> 
> Bernward

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




From tsvwg-bounces@ietf.org Wed Apr 12 08:16:12 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTeGa-0005sf-AC; Wed, 12 Apr 2006 08:16:12 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTeG9-000520-7P
	for tsvwg@ietf.org; Wed, 12 Apr 2006 08:15:45 -0400
Received: from gecko.sbs.de ([194.138.37.40])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTe75-0002mc-RI
	for tsvwg@ietf.org; Wed, 12 Apr 2006 08:06:25 -0400
Received: from mail1.sbs.de (localhost [127.0.0.1])
	by gecko.sbs.de (8.12.6/8.12.6) with ESMTP id k3CC6L1H008352;
	Wed, 12 Apr 2006 14:06:21 +0200
Received: from fthw9xoa.ww002.siemens.net (fthw9xoa.ww002.siemens.net
	[157.163.133.201])
	by mail1.sbs.de (8.12.6/8.12.6) with ESMTP id k3CC6K5W013354;
	Wed, 12 Apr 2006 14:06:20 +0200
Received: from MCHP7I6A.ww002.siemens.net ([139.25.131.137]) by
	fthw9xoa.ww002.siemens.net with Microsoft SMTPSVC(6.0.3790.1830); 
	Wed, 12 Apr 2006 14:05:26 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Subject: AW: [Tsvwg] Re: SCTP PAD chunk
Date: Wed, 12 Apr 2006 14:05:26 +0200
Message-ID: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BCE@MCHP7I6A.ww002.siemens.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [Tsvwg] Re: SCTP PAD chunk
Thread-Index: AcZeJSsER6F9vvLLTZytpMxTC0vrsQAAqZyA
From: "Meyknecht, Bernward" <bernward.meyknecht@siemens.com>
To: <bidulock@openss7.org>, "Randall Stewart" <rrs@cisco.com>
X-OriginalArrivalTime: 12 Apr 2006 12:05:26.0885 (UTC)
	FILETIME=[62AD9150:01C65E29]
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 8f374d0786b25a451ef87d82c076f593
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org


Hi all,

concerning the altered passage I can only repeat my previous remark
on compatibility to existing and deployed implementation,
though I like Brians idea.

Maybe it's good to follow the approach from Randy with the supported
chunk types parameter, and we should still be free to say, that any =
chunk=20
type that is presented in the supported chunk types parameter,
is allowed to appear at any reasonable position within the SCTP
packet. Would that be something to follow?

Of course 'reasonable' may include some rules concerning association
establishment or shutdown.

Kind regards,

Bernward

 =20

-----Urspr=FCngliche Nachricht-----
Von: Brian F. G. Bidulock [mailto:bidulock@openss7.org]=20
Gesendet: Mittwoch, 12. April 2006 13:32
An: Randall Stewart
Cc: Mark Butler; Meyknecht, Bernward; Michael Tuexen; tsvwg@ietf.org
Betreff: Re: [Tsvwg] Re: SCTP PAD chunk

Randall,

Yes, I agree.  The protocol is quite extensible and it is not
all that difficult to write an extension draft; however, I also
agree with Mark that writing a draft without some interest
within the WG might be a wasted effort.

Also, I am concerned with the base protocol in that it appears
that, as Mark put it, all future chunks will be considered
control chunks by a base implementation because they do not have
the precise tag value of a DATA chunk.  This is not so good for
future extensibility.  This is one of the considerations I made
when I changed my loop to not just check for DATA chunks.  If
there is an unrecognized chunk in the sequence of DATA chunks,
it should be reported and aborted or continued in accordance
with the bits in the chunk type.  That is, unrecognized chunk
types should be considered as possibly being of a "DATA" type.

I think that the passage should be considered for revision in
the 2960bis document, if nothing more, but to pave the way for
future DATA-chunk-like extensions.

Are there comments on altering the passage in question?

--brian


Randall Stewart wrote:                         (Wed, 12 Apr 2006 =
06:50:09)
> Brian:
>=20
> I would ask that Mark go ahead and write his own
> draft... If he wants to have some cool features
> to have new types of data chunks and special padding..
> thats wonderful.. but lets not put it in the base spec.
>=20
> He can go ahead and write a draft that has a data chunk
> that can have pad and all that stuff in it..
>=20
> Lets leave the PAD chunk for control and what it
> was intended.. PMTU probing for Matt Mathis's stuff..
>=20
> That way this whole thing can be negotiated at startup
> that you "support the advanced data chunk" and then
> the implementations that support this can use it instead
> of the normal data chunk... If he is defining a new
> data chunk he can surely add a "pad bytes" field to it
> to indicate the number of bytes of trailing pad..
>=20
> Thus a control chunk pad.. and this new data chunk is
> all we need to handle all padding scenarios :-D
>=20
> R
>=20
>=20
> Brian F. G. Bidulock wrote:
> > Mark,
> >=20
> > While your defining the cool new data chunks that includes fragment
> > numbers, why not have it specify the length and offset of the data
> > within the chunk to allow for internal padding.
> >=20
> > I really question the use of this ISO like feature.  Everything of
> > interest (including data) is already padded to a 32-bit alignment
> > boundary.  Did you have some 128-bit architecture for which that was
> > insufficient?
> >=20
> > --brian
> >=20
> > Mark Butler wrote:                             (Mon, 10 Apr 2006 =
09:24:19)
> >=20
> >>Randall Stewart wrote:
> >>
> >>
> >>>Bernward:
> >>>
> >>>Yes, this is a real concern I have about putting a
> >>>pad chunk between two data chunks... I am not
> >>>to sure that implementations will support this... not
> >>>that they should not.. but it is something that we must
> >>>use care if we define this new chunk...
> >>>
> >>>If we make it able to go anywhere.. then maybe a restriction
> >>>on not sending it unless your peer announces support
> >>>for it... with the supported chunk types parameters... this
> >>>would then make it so an implemenation would say.. yes I
> >>>support this.. and in so doing .. say it may fall anywhere
> >>>in the packet.
> >>
> >>I would prefer a parameter that indicates the sender allows control =
and=20
> >>data chunks to be bundled in any reasonable order, as if they were =
in=20
> >>separate packets.  Any compliant implementation implicitly supports =
a=20
> >>PAD chunk that occurs before any DATA chunks.
> >>
> >> - Mark B.
> >=20
> >=20
>=20
>=20
> --=20
> Randall Stewart
> ITD - Cisco Systems Inc.
> 803-345-0369 <or> 815-342-5222 (cell)

--=20
Brian F. G. Bidulock
bidulock@openss7.org
http://www.openss7.org/




From tsvwg-bounces@ietf.org Wed Apr 12 08:16:22 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTeGk-0006Qh-AN; Wed, 12 Apr 2006 08:16:22 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTeG7-000520-5b
	for tsvwg@ietf.org; Wed, 12 Apr 2006 08:15:43 -0400
Received: from stsc1260-eth-s1-s1p1-vip.va.neustar.com ([156.154.16.129]
	helo=chiedprmail1.ietf.org)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTdwp-0002Ie-2u
	for tsvwg@ietf.org; Wed, 12 Apr 2006 07:55:47 -0400
Received: from sj-iport-4.cisco.com ([171.68.10.86])
	by chiedprmail1.ietf.org with esmtp (Exim 4.43) id 1FTdmF-00021m-Ve
	for tsvwg@ietf.org; Wed, 12 Apr 2006 07:44:53 -0400
Received: from sj-core-3.cisco.com ([171.68.223.137])
	by sj-iport-4.cisco.com with ESMTP; 12 Apr 2006 04:44:51 -0700
X-IronPort-AV: i="4.04,114,1144047600"; 
	d="scan'208"; a="1794195497:sNHT30346648"
Received: from [IPv6:::1] (ssh-sjc-1.cisco.com [171.68.225.134])
	by sj-core-3.cisco.com (8.12.10/8.12.6) with ESMTP id k3CBinVI019692;
	Wed, 12 Apr 2006 04:44:50 -0700 (PDT)
Message-ID: <443CE833.9000000@cisco.com>
Date: Wed, 12 Apr 2006 07:44:51 -0400
From: Randall Stewart <rrs@cisco.com>
User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060223
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: bidulock@openss7.org
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
	<443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
	<20060410112605.A29853@openss7.org>
	<443CDB61.3030609@cisco.com> <20060412053212.A8295@openss7.org>
In-Reply-To: <20060412053212.A8295@openss7.org>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: -2.6 (--)
X-Scan-Signature: 82c9bddb247d9ba4471160a9a865a5f3
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org,
	"Meyknecht, Bernward" <bernward.meyknecht@Siemens.com>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Brian F. G. Bidulock wrote:
> Randall,
> 
> Yes, I agree.  The protocol is quite extensible and it is not
> all that difficult to write an extension draft; however, I also
> agree with Mark that writing a draft without some interest
> within the WG might be a wasted effort.
> 
> Also, I am concerned with the base protocol in that it appears
> that, as Mark put it, all future chunks will be considered
> control chunks by a base implementation because they do not have
> the precise tag value of a DATA chunk.  This is not so good for
> future extensibility.  This is one of the considerations I made
> when I changed my loop to not just check for DATA chunks.  If
> there is an unrecognized chunk in the sequence of DATA chunks,
> it should be reported and aborted or continued in accordance
> with the bits in the chunk type.  That is, unrecognized chunk
> types should be considered as possibly being of a "DATA" type.

I agree, and I think Michael and I will be adding tests to this
nature for the next inter-op..

I am not really to sure what should be done to valid CONTROL
chunks appearing inter-mixed with DATA.. but for sure the
upper bits need to be checked... One can envision at least
3 handlings of valid control in the wrong place:

a) process it anyway
b) skip it and continue
c) stop processing all together at that point.

I am thinking that BSD will change to <b> for the default
and then have <a> and <c> as sysctl options... but I have
not fully decided..

In any event BSD's processing loop will be changing soon to
at a minimum handle the upper bits :-D

> 
> I think that the passage should be considered for revision in
> the 2960bis document, if nothing more, but to pave the way for
> future DATA-chunk-like extensions.
> 
> Are there comments on altering the passage in question?
> 
I don't know if we need to change 2960 on this.. it already
says you must handle the upper bits.. I think this is more
a matter of we have just failed to check this  :-D

R


-- 
Randall Stewart
ITD - Cisco Systems Inc.
803-345-0369 <or> 815-342-5222 (cell)




From tsvwg-bounces@ietf.org Wed Apr 12 08:21:27 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTeLc-0007C8-EX; Wed, 12 Apr 2006 08:21:24 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTeLa-0007C3-P5
	for tsvwg@ietf.org; Wed, 12 Apr 2006 08:21:22 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTeLa-0003tG-AG
	for tsvwg@ietf.org; Wed, 12 Apr 2006 08:21:22 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:ZxWuEDZ0tLfeQaIrQdhBhJtWrZzkbT+j@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k3CCLLq08878;
	Wed, 12 Apr 2006 06:21:21 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k3CCLLg09888;
	Wed, 12 Apr 2006 06:21:21 -0600
Date: Wed, 12 Apr 2006 06:21:21 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <rrs@cisco.com>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Message-ID: <20060412062121.B9084@openss7.org>
Mail-Followup-To: Randall Stewart <rrs@cisco.com>,
	Mark Butler <butlerm@middle.net>,
	"Meyknecht, Bernward" <bernward.meyknecht@Siemens.com>,
	Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
	<443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
	<20060410112605.A29853@openss7.org>
	<443CDB61.3030609@cisco.com> <20060412053212.A8295@openss7.org>
	<443CE833.9000000@cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <443CE833.9000000@cisco.com>;
	from rrs@cisco.com on Wed, Apr 12, 2006 at 07:44:51AM -0400
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 00e94c813bef7832af255170dca19e36
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org,
	"Meyknecht, Bernward" <bernward.meyknecht@Siemens.com>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Randall,

Randall Stewart wrote:                                                             (Wed, 12 Apr 2006 07:44:51)
> Brian F. G. Bidulock wrote:
> > Randall,
> > 
> > Yes, I agree.  The protocol is quite extensible and it is not
> > all that difficult to write an extension draft; however, I also
> > agree with Mark that writing a draft without some interest
> > within the WG might be a wasted effort.
> > 
> > Also, I am concerned with the base protocol in that it appears
> > that, as Mark put it, all future chunks will be considered
> > control chunks by a base implementation because they do not have
> > the precise tag value of a DATA chunk.  This is not so good for
> > future extensibility.  This is one of the considerations I made
> > when I changed my loop to not just check for DATA chunks.  If
> > there is an unrecognized chunk in the sequence of DATA chunks,
> > it should be reported and aborted or continued in accordance
> > with the bits in the chunk type.  That is, unrecognized chunk
> > types should be considered as possibly being of a "DATA" type.
> 
> I agree, and I think Michael and I will be adding tests to this
> nature for the next inter-op..
> 
> I am not really to sure what should be done to valid CONTROL
> chunks appearing inter-mixed with DATA.. but for sure the
> upper bits need to be checked... One can envision at least
> 3 handlings of valid control in the wrong place:
> 
> a) process it anyway
> b) skip it and continue
> c) stop processing all together at that point.
> 
> I am thinking that BSD will change to <b> for the default
> and then have <a> and <c> as sysctl options... but I have
> not fully decided..
> 
> In any event BSD's processing loop will be changing soon to
> at a minimum handle the upper bits :-D

Should the behavior differ from unrecognized to valid?  When a chunk
changes from unrecognized to valid do you really want its handling
between DATA chunks to change?

I really lean towards a) for both.  Bundling logic can be complicated
at times, and non-SCTP savy middle boxes could concatenate or split
packets at arbitrary chunk boundaries.  Also, when assembling very
large packets, either the packet must be terminated early or a
control chunk delayed until after the packet is assembled.  I agree
with Mark in that it should not make any difference if the control
chunk starts the next packet, or continues in the current packet.

> 
> > 
> > I think that the passage should be considered for revision in
> > the 2960bis document, if nothing more, but to pave the way for
> > future DATA-chunk-like extensions.
> > 
> > Are there comments on altering the passage in question?
> > 
> I don't know if we need to change 2960 on this.. it already
> says you must handle the upper bits.. I think this is more
> a matter of we have just failed to check this  :-D

I meant reference to control chunks (being non-true-DATA chunks)
having to appear before DATA in the packet.  I think that sending
control chunks first could be a SHOULD (with consideration for
implementations that might not be able to handle them out of place).

--brian

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




From tsvwg-bounces@ietf.org Wed Apr 12 15:31:05 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTl3M-0007NM-JN; Wed, 12 Apr 2006 15:31:00 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTl3L-0007JA-Jy
	for tsvwg@ietf.org; Wed, 12 Apr 2006 15:30:59 -0400
Received: from test-iport-3.cisco.com ([171.71.176.78])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTl3K-0001l7-8c
	for tsvwg@ietf.org; Wed, 12 Apr 2006 15:30:59 -0400
Received: from sj-core-1.cisco.com ([171.71.177.237])
	by test-iport-3.cisco.com with ESMTP; 12 Apr 2006 12:30:57 -0700
Received: from xbh-sjc-221.amer.cisco.com (xbh-sjc-221.cisco.com
	[128.107.191.63])
	by sj-core-1.cisco.com (8.12.10/8.12.6) with ESMTP id k3CJUvw1015591
	for <tsvwg@ietf.org>; Wed, 12 Apr 2006 12:30:57 -0700 (PDT)
Received: from xfe-sjc-211.amer.cisco.com ([171.70.151.174]) by
	xbh-sjc-221.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Wed, 12 Apr 2006 12:30:57 -0700
Received: from jmpolk-wxp.cisco.com ([10.89.20.72]) by
	xfe-sjc-211.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Wed, 12 Apr 2006 12:30:56 -0700
Message-Id: <4.3.2.7.2.20060412142218.0193fe40@email.cisco.com>
X-Sender: jmpolk@email.cisco.com
X-Mailer: QUALCOMM Windows Eudora Version 4.3.2
Date: Wed, 12 Apr 2006 14:30:56 -0500
To: tsvwg <tsvwg@ietf.org>
From: "James M. Polk" <jmpolk@cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
X-OriginalArrivalTime: 12 Apr 2006 19:30:57.0171 (UTC)
	FILETIME=[9F2B8630:01C65E67]
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 9466e0365fc95844abaf7c3f15a05c7d
Subject: [Tsvwg] Hum confirm on draft-chan-tsvwg-diffserv-class-aggr-03
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

TSVWG

This message is to confirm a hum taken in the Dallas meeting for

	draft-chan-tsvwg-diffserv-class-aggr-03

becoming a WG item.

The chairs viewed the hum results to be:

100% for
0% against

This note to the WG is asking whether anyone on the list feels that this 
hum result misrepresents WG consensus?

Given that this Hum is recorded in the draft meeting minutes sent out last 
week, and no one has questioned the results there to date, I am going to 
allow for comments stating the hum results within the meeting regarding 
this ID were not WG consensus to end on April 19th.

In the future, list confirmations will be sent out within a week of the 
meeting.
cheers,

James
Lars
Magnus
	IETF TSVWG co-chairs




From tsvwg-bounces@ietf.org Wed Apr 12 15:31:05 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTl3R-0007Of-SL; Wed, 12 Apr 2006 15:31:05 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTl3Q-0007Ne-Mi
	for tsvwg@ietf.org; Wed, 12 Apr 2006 15:31:04 -0400
Received: from sj-iport-5.cisco.com ([171.68.10.87])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTl3N-0001lA-Cy
	for tsvwg@ietf.org; Wed, 12 Apr 2006 15:31:04 -0400
Received: from sj-core-4.cisco.com ([171.68.223.138])
	by sj-iport-5.cisco.com with ESMTP; 12 Apr 2006 12:31:00 -0700
X-IronPort-AV: i="4.04,115,1144047600"; 
	d="scan'208"; a="268330019:sNHT41087880"
Received: from xbh-sjc-211.amer.cisco.com (xbh-sjc-211.cisco.com
	[171.70.151.144])
	by sj-core-4.cisco.com (8.12.10/8.12.6) with ESMTP id k3CJUrZE025749
	for <tsvwg@ietf.org>; Wed, 12 Apr 2006 12:30:59 -0700 (PDT)
Received: from xfe-sjc-212.amer.cisco.com ([171.70.151.187]) by
	xbh-sjc-211.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Wed, 12 Apr 2006 12:30:59 -0700
Received: from jmpolk-wxp.cisco.com ([10.89.20.72]) by
	xfe-sjc-212.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Wed, 12 Apr 2006 12:30:58 -0700
Message-Id: <4.3.2.7.2.20060412135715.029d56a8@email.cisco.com>
X-Sender: jmpolk@email.cisco.com
X-Mailer: QUALCOMM Windows Eudora Version 4.3.2
Date: Wed, 12 Apr 2006 14:30:58 -0500
To: tsvwg <tsvwg@ietf.org>
From: "James M. Polk" <jmpolk@cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
X-OriginalArrivalTime: 12 Apr 2006 19:30:59.0277 (UTC)
	FILETIME=[A06CDFD0:01C65E67]
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 9466e0365fc95844abaf7c3f15a05c7d
Subject: [Tsvwg] Hum confirm on draft-tuexen-tsvwg-sctp-padding-00
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

TSVWG

This message is to confirm a hum taken in the Dallas meeting for

	draft-tuexen-tsvwg-sctp-padding-00

becoming a WG item.

The chairs viewed the hum results to be:

100% for
0% against

This note to the WG is asking whether anyone on the list feels that this 
hum result misrepresents WG consensus?

Given that this Hum is recorded in the draft meeting minutes sent out last 
week, and no one has questioned the results there to date, I am going to 
allow for comments stating the hum results within the meeting regarding 
this ID were not WG consensus to end on April 19th.

In the future, list confirmations will be sent out within a week of the 
meeting.
cheers,

James
Lars
Magnus
	IETF TSVWG co-chairs




From tsvwg-bounces@ietf.org Wed Apr 12 18:01:49 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FTnPD-00028c-Rt; Wed, 12 Apr 2006 18:01:43 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FTnPC-00028X-6q
	for tsvwg@ietf.org; Wed, 12 Apr 2006 18:01:42 -0400
Received: from cat.tcb.net ([64.78.150.134] helo=dog.tcb.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FTnP8-0002Nw-U4
	for tsvwg@ietf.org; Wed, 12 Apr 2006 18:01:42 -0400
Received: from [127.0.0.1] (dog.tcb.net [64.78.150.133])
	(using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits))
	(No client certificate requested)
	by dog.tcb.net (Postfix) with ESMTP id 61E1E644E3;
	Wed, 12 Apr 2006 16:01:31 -0600 (MDT)
Message-ID: <443D78BE.306@castlepoint.net>
Date: Wed, 12 Apr 2006 16:01:34 -0600
From: Shane Amante <shane@castlepoint.net>
User-Agent: Thunderbird 1.5 (Windows/20051201)
MIME-Version: 1.0
To: "James M. Polk" <jmpolk@cisco.com>
Subject: Re: [Tsvwg] Hum confirm on draft-chan-tsvwg-diffserv-class-aggr-03
References: <4.3.2.7.2.20060412142218.0193fe40@email.cisco.com>
In-Reply-To: <4.3.2.7.2.20060412142218.0193fe40@email.cisco.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 7baded97d9887f7a0c7e8a33c2e3ea1b
Cc: tsvwg <tsvwg@ietf.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

James, All,

I wasn't able to attend TSVWG in Dallas, therefore wasn't able to add to 
the "100% for" below, but I also support this I-D becoming a WG doc.

-shane


James M. Polk wrote:
> TSVWG
> 
> This message is to confirm a hum taken in the Dallas meeting for
> 
>     draft-chan-tsvwg-diffserv-class-aggr-03
> 
> becoming a WG item.
> 
> The chairs viewed the hum results to be:
> 
> 100% for
> 0% against
> 
> This note to the WG is asking whether anyone on the list feels that this 
> hum result misrepresents WG consensus?
> 
> Given that this Hum is recorded in the draft meeting minutes sent out 
> last week, and no one has questioned the results there to date, I am 
> going to allow for comments stating the hum results within the meeting 
> regarding this ID were not WG consensus to end on April 19th.
> 
> In the future, list confirmations will be sent out within a week of the 
> meeting.
> cheers,
> 
> James
> Lars
> Magnus
>     IETF TSVWG co-chairs
> 





From tsvwg-bounces@ietf.org Thu Apr 13 09:30:33 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FU1tr-0000zo-Hq; Thu, 13 Apr 2006 09:30:19 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FU1tq-0000zj-8i
	for tsvwg@ietf.org; Thu, 13 Apr 2006 09:30:18 -0400
Received: from ams-iport-1.cisco.com ([144.254.224.140])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FU1tp-00050N-Rp
	for tsvwg@ietf.org; Thu, 13 Apr 2006 09:30:18 -0400
Received: from ams-core-1.cisco.com ([144.254.224.150])
	by ams-iport-1.cisco.com with ESMTP; 13 Apr 2006 15:30:18 +0200
Received: from xbh-ams-331.emea.cisco.com (xbh-ams-331.cisco.com
	[144.254.231.71])
	by ams-core-1.cisco.com (8.12.10/8.12.6) with ESMTP id k3DDUGUE027011; 
	Thu, 13 Apr 2006 15:30:16 +0200 (MEST)
Received: from xmb-ams-333.cisco.com ([144.254.231.78]) by
	xbh-ams-331.emea.cisco.com with Microsoft SMTPSVC(6.0.3790.0); 
	Thu, 13 Apr 2006 15:30:16 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="US-ASCII"
Content-Transfer-Encoding: quoted-printable
Subject: RE: [Tsvwg] WGLC for draft-ietf-tsvwg-rsvp-dste-01.txt
Date: Thu, 13 Apr 2006 15:30:09 +0200
Message-ID: <A05118C6DF9320488C77F3D5459B17B7013B2674@xmb-ams-333.emea.cisco.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [Tsvwg] WGLC for draft-ietf-tsvwg-rsvp-dste-01.txt
Thread-Index: AcZCU0DibzBbgA0TSpSCsquowztrbQcqGdTA
From: "Francois Le Faucheur \(flefauch\)" <flefauch@cisco.com>
To: "tsvwg" <tsvwg@ietf.org>
X-OriginalArrivalTime: 13 Apr 2006 13:30:16.0426 (UTC)
	FILETIME=[66B1A4A0:01C65EFE]
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 3002fc2e661cd7f114cb6bae92fe88f1
Cc: magnus Westerlund <magnus.westerlund@ericsson.com>,
	"James Polk \(jmpolk\)" <jmpolk@cisco.com>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

>> TSVWG
>>=20
>> The chairs have concluded the WGLC for the document:
>>=20
>>          draft-ietf-tsvwg-rsvp-dste-01.txt
>>=20
>> which started November 8th, has closed.
>>=20
>> Allison
>> James
>> Jon
>> 	IETF TSVWG co-chairs

Hello,

I received a couple of good comments privately:
	- the draft discusses forwarding and Router Alert handling for
E2E Path messages, but it does not explicitely discusses the other E2E
RSVP messages which also use Router Alert (i.e. PathTear and ResvConf).
While the handling of those is similar, it is worth being explicit about
that.
	- When discussing the forwarding/encapsulation of E2E Path
message (in section 3.2) the current text make it sounds like the two
"methods" are mutually exclusive , while the 2nd one is more an option
on top of the 1st one.=20

I think addressing those would improve the document. So I'd like to
address those quickly in a new rev which would be used for the IESG
review. Below are the proposed edits. Please review and send
comments/concerns if any.

Thanks

Francois


Proposed Edits on draft-ietf-tsvwg-rsvp-dste-01.txt:
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D

In section 3.2,
Replace:
"The preferred method for the Aggregator to send the E2E Path message is
to address it directly to the Deaggregator by setting the destination
address in the IP Header of the E2E Path message to the Deaggregator
address. "
By:
"To send the E2E Path message, the Aggregator MUST address it directly
to the Deaggregator by setting the destination address in the IP Header
of the E2E Path message to the Deaggregator address."


In section 3.2,=20
Replace:
"An alternate method for the Aggregator to send the E2E Path is to
encapsulate the E2E Path message in an IP tunnel or in the TE tunnel
itself and unicast the E2E Path message to the Deaggregator, without the
Router Alert option. "
By:
"Optionally, the Aggregator MAY also encapsulate the E2E Path message in
an IP tunnel or in the TE tunnel itself."


In section 3.2,
Replace:
"With both methods, the Router Alert is not set."
By:
"Regardless of the encapsulation method, the Router Alert is not set."


At the very end of section 3.2,
Add:
"PathTear and ResvConf messages MUST be fowarded by the Aggregator to
the Deaggregator exactly like Path messages".


At the very end of section 3.6,
Add:
"On recipt of an E2E ResvConf from the Aggregator, the Deaggregator MUST
forward the E2E ResvConf downstream towards the receiver. In doing so,
the Deaggregator sets the destination address in the IP header of the
E2E ResvConf message to the IP address found in the RESV_CONFIRM object
of the corresponding Resv. The Deaggregator also sets the Router Alert.
"




From tsvwg-bounces@ietf.org Fri Apr 14 05:00:55 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FUKAU-0001ev-Jo; Fri, 14 Apr 2006 05:00:42 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FUKAS-0001en-Q4; Fri, 14 Apr 2006 05:00:40 -0400
Received: from mail.sbs.be ([193.109.72.11])
	by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FUKAQ-0001aS-7X; Fri, 14 Apr 2006 05:00:40 -0400
Received: from bruc001x.sbs.be (bruc001x.sbs.be [193.210.175.38])
	by mail.sbs.be (8.12.10/8.12.10/SuSE Linux 0.7) with ESMTP id
	k3E8vcbX026329; Fri, 14 Apr 2006 10:57:38 +0200
Received: from bru0007a.ww018.siemens.net (bru0007a.ww018.siemens.net
	[193.210.175.161])
	by bruc001x.sbs.be (8.12.10/8.12.10/SuSE Linux 0.7) with ESMTP id
	k3E8va4b015567; Fri, 14 Apr 2006 10:57:36 +0200
Received: from BRU0038A.ww018.siemens.net ([193.210.175.64]) by
	bru0007a.ww018.siemens.net with Microsoft SMTPSVC(6.0.3790.1830); 
	Fri, 14 Apr 2006 10:57:33 +0200
x-mimeole: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Date: Fri, 14 Apr 2006 10:57:31 +0200
Message-ID: <67043E463DDBFD4A8087ED940BFF756B9CDDCF@BRU0038A.ww018.siemens.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Sctp on Microsoft Windows machines
Thread-Index: AcZfoXaK4ZIa2uKVQ4SPApTWDYgspg==
From: "Coene, Lode" <Lode.Coene@siemens.com>
To: <tsvwg@ietf.org>, <sigtran@ietf.org>, <sctp-impl@external.cisco.com>,
	<discussion@sctp.de>
X-OriginalArrivalTime: 14 Apr 2006 08:57:33.0074 (UTC)
	FILETIME=[77CA0720:01C65FA1]
X-Spam-Score: 0.0 (/)
X-Scan-Signature: b19722fc8d3865b147c75ae2495625f2
Cc: Michael.Tuexen@micmac.franken.de
Subject: [Tsvwg] Sctp on Microsoft Windows machines
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

The SCTP library user land implementation( http://www.sctp.de )
developed by the=20
University of Essen and the University of Munster is also able to be
compiled and=20
used on a Microsoft Windows operating system.=20

I've tested it on a Windows XP machne and have made a Windows Installer
for it.
It should install on a Windows XP SP2 and also on Windows 2000.=20

You can find the Windows installer for SCTPlib at
http://www.sctp.be/sctplib
It contains the compiled sctplib library, some sample program and the
source code.=20

The source code can also be downloaded at http://www.sctp.de

Many thanks to Michael Tuxen and associates for making this possible.

If there are problems with the installer, send a mail to
email:sctp@sctp.be=20
or to the discussion forum at http://www.sctp.de
(Watch out: download at .be, discussion at .de)

Yours sincerely,
Lode Coene

Siemens COM
atealaan 34          2200 Herentals, Belgium
E-mail: lode.coene@siemens.com
Tel: +32-14-252081
Fax: +32-14-253212





From tsvwg-bounces@ietf.org Fri Apr 14 07:26:11 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FUMQs-0001sp-Sc; Fri, 14 Apr 2006 07:25:46 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FUMQr-0001oV-RL
	for tsvwg@ietf.org; Fri, 14 Apr 2006 07:25:45 -0400
Received: from sj-iport-4.cisco.com ([171.68.10.86])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FUMQr-0006Ha-Gs
	for tsvwg@ietf.org; Fri, 14 Apr 2006 07:25:45 -0400
Received: from sj-core-4.cisco.com ([171.68.223.138])
	by sj-iport-4.cisco.com with ESMTP; 14 Apr 2006 04:25:45 -0700
X-IronPort-AV: i="4.04,120,1144047600"; 
	d="scan'208"; a="1795006858:sNHT33512612"
Received: from [IPv6:::1] (ssh-sjc-1.cisco.com [171.68.225.134])
	by sj-core-4.cisco.com (8.12.10/8.12.6) with ESMTP id k3EBPgYg004236;
	Fri, 14 Apr 2006 04:25:43 -0700 (PDT)
Message-ID: <443F86B8.4020500@cisco.com>
Date: Fri, 14 Apr 2006 07:25:44 -0400
From: Randall Stewart <rrs@cisco.com>
User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060223
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: bidulock@openss7.org
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
	<443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
	<20060410112605.A29853@openss7.org>
	<443CDB61.3030609@cisco.com> <20060412053212.A8295@openss7.org>
	<443CE833.9000000@cisco.com> <20060412062121.B9084@openss7.org>
In-Reply-To: <20060412062121.B9084@openss7.org>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: a8a20a483a84f747e56475e290ee868e
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org,
	"Meyknecht, Bernward" <bernward.meyknecht@Siemens.com>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Comments in-line..

Brian F. G. Bidulock wrote:

>>
>>a) process it anyway
>>b) skip it and continue
>>c) stop processing all together at that point.
>>
>>I am thinking that BSD will change to <b> for the default
>>and then have <a> and <c> as sysctl options... but I have
>>not fully decided..
>>
>>In any event BSD's processing loop will be changing soon to
>>at a minimum handle the upper bits :-D
> 
> 
> Should the behavior differ from unrecognized to valid?  When a chunk
> changes from unrecognized to valid do you really want its handling
> between DATA chunks to change?
> 
> I really lean towards a) for both.  Bundling logic can be complicated
> at times, and non-SCTP savy middle boxes could concatenate or split
> packets at arbitrary chunk boundaries.  Also, when assembling very
> large packets, either the packet must be terminated early or a
> control chunk delayed until after the packet is assembled.  I agree
> with Mark in that it should not make any difference if the control
> chunk starts the next packet, or continues in the current packet.
> 


And you can do any of these.. and also a d) I did
not mention.. abort the association with a
"protocol violation".

This I will bet some implementations will do....

I am leaning towards making <b> the standard procedure in
BSD with sysctl options to do both <a> and <d>.

That way the default is modest.. but the sysadmin has
the ability to be liberal or very conservative... which
ever it chooses..


> 
>>>I think that the passage should be considered for revision in
>>>the 2960bis document, if nothing more, but to pave the way for
>>>future DATA-chunk-like extensions.
>>>
>>>Are there comments on altering the passage in question?
>>>
>>
>>I don't know if we need to change 2960 on this.. it already
>>says you must handle the upper bits.. I think this is more
>>a matter of we have just failed to check this  :-D
> 
> 
> I meant reference to control chunks (being non-true-DATA chunks)
> having to appear before DATA in the packet.  I think that sending
> control chunks first could be a SHOULD (with consideration for
> implementations that might not be able to handle them out of place).
> 

So let me see.. you are advocating a signifigant change to
2960 that:

a) Has no real advantage... in fact your first email argued
    as to why control should be first...

b) Will surely cause implementation incompatability... since I
    know that there are implementations out there that will probably
    do <d>. So new implementations that do this change would NOT
    inter-operate...

c) Has never come up at any inter-op in the last 6 years.


Sorry I just can't buy it...

At the last inter-op we had 14 implentations, of those 13
of them out-of-the box inter-operated.. only one, yours, had
issues and could not talk to anyone else without signifigant
code changes... you of all people must know how frustrating
this can be... I would not want any such change to the spec
that would only make things worse for developers with no
apparent advantage....

It is NOT hard to make control come first. On top of that
your previous talk of large packets is just not valid.. if
you reach the PMTU or close to it.. you send the all control
bundled packet.. and then continue bundling your next
control in the next packet... this is not hard... especially
since you must always be watching the PMTU has you bundle
and send when you have a full packet (or close to full).


R

-- 
Randall Stewart
ITD - Cisco Systems Inc.
803-345-0369 <or> 815-342-5222 (cell)




From tsvwg-bounces@ietf.org Fri Apr 14 14:27:16 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FUT0c-0001uq-Fc; Fri, 14 Apr 2006 14:27:06 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FUT0b-0001ul-Cj
	for tsvwg@ietf.org; Fri, 14 Apr 2006 14:27:05 -0400
Received: from sj-iport-4.cisco.com ([171.68.10.86])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FUT0a-0003Oa-3X
	for tsvwg@ietf.org; Fri, 14 Apr 2006 14:27:05 -0400
Received: from sj-core-1.cisco.com ([171.71.177.237])
	by sj-iport-4.cisco.com with ESMTP; 14 Apr 2006 11:27:04 -0700
X-IronPort-AV: i="4.04,121,1144047600"; 
	d="scan'208"; a="1795108471:sNHT27529942"
Received: from xbh-sjc-221.amer.cisco.com (xbh-sjc-221.cisco.com
	[128.107.191.63])
	by sj-core-1.cisco.com (8.12.10/8.12.6) with ESMTP id k3EIR3w1012846
	for <tsvwg@ietf.org>; Fri, 14 Apr 2006 11:27:03 -0700 (PDT)
Received: from xfe-sjc-211.amer.cisco.com ([171.70.151.174]) by
	xbh-sjc-221.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Fri, 14 Apr 2006 11:27:03 -0700
Received: from jmpolk-wxp.cisco.com ([10.89.20.23]) by
	xfe-sjc-211.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Fri, 14 Apr 2006 11:27:02 -0700
Message-Id: <4.3.2.7.2.20060414132041.02a01598@email.cisco.com>
X-Sender: jmpolk@email.cisco.com
X-Mailer: QUALCOMM Windows Eudora Version 4.3.2
Date: Fri, 14 Apr 2006 13:27:03 -0500
To: tsvwg <tsvwg@ietf.org>
From: "James M. Polk" <jmpolk@cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
X-OriginalArrivalTime: 14 Apr 2006 18:27:02.0948 (UTC)
	FILETIME=[069F0640:01C65FF1]
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 08170828343bcf1325e4a0fb4584481c
Subject: [Tsvwg] Final IETF Minutes uploaded
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

TSVWG

The FINAL IETF65 meeting minutes have been uploaded here:

http://www3.ietf.org/proceedings/06mar/minutes/tsvwg.txt

Thank you to those of you who provided in-meeting notes (Steve, Kwok, 
Randy), Bar BoF notes (Bob, Magnus), as well as those who provided 
additions and corrections (Ken, Francois).
cheers,

James
Lars
Magnus
	IETF TSVWG co-chairs




From tsvwg-bounces@ietf.org Fri Apr 14 16:49:40 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FUVEJ-00021N-20; Fri, 14 Apr 2006 16:49:23 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FUVEG-00021I-Vq
	for tsvwg@ietf.org; Fri, 14 Apr 2006 16:49:20 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FUVED-0008TI-MN
	for tsvwg@ietf.org; Fri, 14 Apr 2006 16:49:18 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24) id 1FUVEE-0000V7-60
	for tsvwg@ietf.org; Fri, 14 Apr 2006 14:48:56 -0600
Message-ID: <44400AD5.2060805@middle.net>
Date: Fri, 14 Apr 2006 14:49:25 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: tsvwg@ietf.org
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: b19722fc8d3865b147c75ae2495625f2
Subject: [Tsvwg] SCTP PAD chunk ordering issues
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

A couple of comments:

Aborting due to protocol violation just because it is a violation is 
enormously counterproductive, and contrary to the principle of being 
exacting in what you send and forgiving in what you accept.   Suppose 
one needs to communicate in an emergency situation, and the two 
endpoints refuse to speak to each other simply due to minor protocol 
inconsistencies - not problems of the type that would keep 
communications from occuring at all, but simple technicalities.  Dumping 
a connection under such circumstances is positively dangerous.

Valid reasons for not tolerating a protocol violation (on my view) are:

  a) Security considerations
  b) Resource constraints
  c) Implementation complexity
  d) Other undesirable side effects

On this principle, it is perfectly rational for a a specification to 
have different requirements for what a implementation sends, and what an 
implementation receives.  Short of explicit negotation, changing the 
send ordering constraints is not a good idea (even if RFC2960 
specification is silent or ambiguous about the control / data status of 
unknown chunk types).  However, recommending that implementions accept 
packets with chunks that are not in the proper order is an independent 
question. 

Sysctl variables are not necessarily for this kind of stuff, just a 
common sense implementation of the "be exact with what you send / 
forgiving in what you receive" principle.

 - Mark B.





From tsvwg-bounces@ietf.org Fri Apr 14 18:49:22 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FUX6G-0006kE-EU; Fri, 14 Apr 2006 18:49:12 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FUX6F-0006k9-Th
	for tsvwg@ietf.org; Fri, 14 Apr 2006 18:49:11 -0400
Received: from boreas.isi.edu ([128.9.160.161])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FUX6E-0003WS-Hu
	for tsvwg@ietf.org; Fri, 14 Apr 2006 18:49:11 -0400
Received: from [128.9.168.55] (upn.isi.edu [128.9.168.55])
	by boreas.isi.edu (8.11.6p2+0917/8.11.2) with ESMTP id k3EMlrM07421;
	Fri, 14 Apr 2006 15:47:53 -0700 (PDT)
Message-ID: <44402637.90809@isi.edu>
Date: Fri, 14 Apr 2006 15:46:15 -0700
From: Joe Touch <touch@ISI.EDU>
User-Agent: Thunderbird 1.5 (Windows/20051201)
MIME-Version: 1.0
To: tsvwg <tsvwg@ietf.org>
X-Enigmail-Version: 0.94.0.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
X-ISI-4-43-8-MailScanner: Found to be clean
X-MailScanner-From: touch@isi.edu
X-Spam-Score: 0.0 (/)
X-Scan-Signature: c1c65599517f9ac32519d043c37c5336
Subject: [Tsvwg] TCP port names option ID discussion on TCPM
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

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

Hi, all,

In response to some discussion on the IETF mailing list, the following
ID has been submitted. It's intended to become a TCPM work item (thus
the header) if there's sufficient interest. I'm cross-posting here
because the doc does address one possible UDP variant, as well as the
potential for use in SCTP and DCCP.

PLEASE TAKE FURTHER DISCUSSION TO THE TCPM LIST (tcpm@ietf.org)

For the original discussion, see the ietf@ietf.org archives, and look
for "Guidance needed on well known ports".

The draft available at the following URL until it's in the ID repositories:
http://www.isi.edu/touch/pubs/draft-touch-tcp-portnames-00.txt

Joe

- ----

TCPM WG                                                        J. Touch
Internet Draft                                                  USC/ISI
Expires: October 2006                                    April 14, 2006

                        A TCP Option for Port Names
                     draft-touch-tcp-portnames-00.txt
...
Abstract

   This document specifies a new TCP option that specifies services by a
   string name. This option separates the use of port numbers for
   connection demultiplexing from their use as a service identifier. The
   option allows a larger number of concurrent connections for a
   particular service, as well as potentially enabling more explicitly
   coordination of services behind NATs and firewalls. This option
   should be supported in new TCP implementations.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFEQCY3E5f5cImnZrsRApmFAKCjqnnfpSP62e3wiabHkp+vT0OoPQCgkMRU
2Ub161iKvbJirek9rtMB9Y8=
=oP9L
-----END PGP SIGNATURE-----




From tsvwg-bounces@ietf.org Sat Apr 15 08:45:36 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FUk9R-0007wr-56; Sat, 15 Apr 2006 08:45:21 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FUk9P-0007wm-NU
	for tsvwg@ietf.org; Sat, 15 Apr 2006 08:45:19 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FUk9O-0005PB-AV
	for tsvwg@ietf.org; Sat, 15 Apr 2006 08:45:19 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:UWUumwFlL5N3UG6NdeAENIzYfDfmjC59@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k3FCjHq06748;
	Sat, 15 Apr 2006 06:45:17 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k3FCjGq04025;
	Sat, 15 Apr 2006 06:45:16 -0600
Date: Sat, 15 Apr 2006 06:45:16 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Randall Stewart <rrs@cisco.com>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Message-ID: <20060415064516.B3390@openss7.org>
Mail-Followup-To: Randall Stewart <rrs@cisco.com>,
	Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>
	<443A2BA5.4000100@cisco.com> <443A78A3.5010508@middle.net>
	<20060410112605.A29853@openss7.org>
	<443CDB61.3030609@cisco.com> <20060412053212.A8295@openss7.org>
	<443CE833.9000000@cisco.com> <20060412062121.B9084@openss7.org>
	<443F86B8.4020500@cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <443F86B8.4020500@cisco.com>;
	from rrs@cisco.com on Fri, Apr 14, 2006 at 07:25:44AM -0400
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 9ed51c9d1356100bce94f1ae4ec616a9
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Randall,

Randall Stewart wrote:                  (Fri, 14 Apr 2006 07:25:44)
> 
> At the last inter-op we had 14 implentations, of those 13
> of them out-of-the box inter-operated.. only one, yours, had
> issues and could not talk to anyone else without signifigant
> code changes... you of all people must know how frustrating
> this can be... I would not want any such change to the spec
> that would only make things worse for developers with no
> apparent advantage....

As you have singled out a specific implementation, some
statements are in order:

The OpenSS7 implementation that was tested at the last interop
was experimental and was tested by a student.  I had asked
Professor Tuexen if we could sponsor one of his graduate
students to attend the event and perform interoperability
testing on an alpha implementation.

Although the alpha implementation did not show as well as our
production implementations tested at previous interops, I was
very pleased with Professor Tuexen's student's ability to patch
kernels and ferret out problems with the alpha implementation.

I was not frustrated in the least, and I believe that the
exercise was a very good learning experience for the student, and
the problems with the alpha implementation that were found more
than compensated for the small fee that was paid to sponsor the
student.  I am thankful that Professor Tuexen was able to make
the arrangement possible.

But this begs a question: do you intend to later identify
specific implementations in a negative light after the interop
in August as you have done here?

--brian

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




From tsvwg-bounces@ietf.org Sat Apr 15 18:05:52 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FUstT-0004rw-FH; Sat, 15 Apr 2006 18:05:27 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FUstR-0004rN-QV
	for tsvwg@ietf.org; Sat, 15 Apr 2006 18:05:25 -0400
Received: from adsl-070-155-160-098.sip.cae.bellsouth.net ([70.155.160.98]
	helo=lakerest.net) by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FUstN-0007sK-39
	for tsvwg@ietf.org; Sat, 15 Apr 2006 18:05:23 -0400
Received: from [IPv6:::1] (localhost [IPv6:::1])
	by lakerest.net (8.13.6/8.13.4) with ESMTP id k3FM5C7e042516;
	Sat, 15 Apr 2006 18:05:14 -0400 (EDT)
	(envelope-from randall@lakerest.net)
DKIM-Signature: a=rsa-sha1; c=simple/simple; d=lakerest.net; s=lakerest;
	t=1145138715; h=Message-ID:Date:From:User-Agent:X-Accept-Language:
	MIME-Version:To:CC:Subject:References:In-Reply-To:Content-Type:
	Content-Transfer-Encoding; b=gqE1y0gz/dVutCMTkI3JfwE5KBd+y/UJOC1Sf2
	2dcCz1gthix7tw6uJddvGUuJ/8TQBeemPQ38m2rfeokzzZ7w==
Message-ID: <44416E1D.10100@lakerest.net>
Date: Sat, 15 Apr 2006 18:05:17 -0400
From: Randall Stewart <randall@lakerest.net>
User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060223
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Mark Butler <butlerm@middle.net>
Subject: Re: [Tsvwg] SCTP PAD chunk ordering issues
References: <44400AD5.2060805@middle.net>
In-Reply-To: <44400AD5.2060805@middle.net>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.1 (/)
X-Scan-Signature: 6e922792024732fb1bb6f346e63517e4
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark:

Lixia once told me that 90% of the time whatever
the default behavior is that is what will occur, so
pick your defaults carefully..

This is exactly why I plan on allowing control
to be interspersed with data (in a future BSD
version... right now any such control would
cause a KAME-BSD implementation to just STOP processing
the packet)... but I will skip the control (since it
is in violation of the RFC).. this does not make
the two endpoints non-interoperable.. but may well
degrade performance.. in fact even current behavior
will not make the implementations inter-operable.. but
it will surely degrade performance since the data
chunks past the mixed up control would be dropped...
thus causing a "loss".

I will give the sysadmin the option of being completely
liberal.. or being very conservative... since there will
be sysadmin's that want to be liberal.. and there WILL
be those (especially in private networks with stricter
requirments) want to be more strict...aka aborting when
you do something wrong...

Using sysctl's to influence behavior in a BSD stack has
been a long accepted mechanism.. for example right now
for TCP in 7.x FreeBSD I have many controls.. here
are just a few:
-----------------------------
net.inet.tcp.syncookies: 1
net.inet.tcp.do_tcpdrain: 1
net.inet.tcp.rfc3042: 1
net.inet.tcp.rfc3390: 1
net.inet.tcp.local_slowstart_flightsize: 4
net.inet.tcp.newreno: 1
net.inet.tcp.sack.enable: 1
net.inet.tcp.icmp_may_rst: 1
net.inet.tcp.hostcache.purge: 0
net.inet.tcp.log_in_vain: 0
net.inet.tcp.blackhole: 0
net.inet.tcp.delayed_ack: 1
-------------------------------

There are actually to many to comfortably put in an email
more than I actually thought :-D

The point is these influence both a) Options and b) security
considerations of the admin.

This is exactly why I plan on putting my little sysctl
in when I get around to fixing this issue in BSD... since
right now it is broken :-<

R




Mark Butler wrote:
> A couple of comments:
> 
> Aborting due to protocol violation just because it is a violation is 
> enormously counterproductive, and contrary to the principle of being 
> exacting in what you send and forgiving in what you accept.   Suppose 
> one needs to communicate in an emergency situation, and the two 
> endpoints refuse to speak to each other simply due to minor protocol 
> inconsistencies - not problems of the type that would keep 
> communications from occuring at all, but simple technicalities.  Dumping 
> a connection under such circumstances is positively dangerous.
> 
> Valid reasons for not tolerating a protocol violation (on my view) are:
> 
>  a) Security considerations
>  b) Resource constraints
>  c) Implementation complexity
>  d) Other undesirable side effects
> 
> On this principle, it is perfectly rational for a a specification to 
> have different requirements for what a implementation sends, and what an 
> implementation receives.  Short of explicit negotation, changing the 
> send ordering constraints is not a good idea (even if RFC2960 
> specification is silent or ambiguous about the control / data status of 
> unknown chunk types).  However, recommending that implementions accept 
> packets with chunks that are not in the proper order is an independent 
> question.
> Sysctl variables are not necessarily for this kind of stuff, just a 
> common sense implementation of the "be exact with what you send / 
> forgiving in what you receive" principle.
> 
> - Mark B.
> 
> 
> 
> 


-- 
Randall Stewart
803-345-0369 <or> 815-342-5222(cell)




From tsvwg-bounces@ietf.org Sat Apr 15 19:32:49 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FUuFh-0000Yo-OO; Sat, 15 Apr 2006 19:32:29 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FUuFg-0000Yg-2T
	for tsvwg@ietf.org; Sat, 15 Apr 2006 19:32:28 -0400
Received: from adsl-070-155-160-098.sip.cae.bellsouth.net ([70.155.160.98]
	helo=lakerest.net) by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FUuFe-0002O4-Mc
	for tsvwg@ietf.org; Sat, 15 Apr 2006 19:32:28 -0400
Received: from [IPv6:::1] (localhost [IPv6:::1])
	by lakerest.net (8.13.6/8.13.4) with ESMTP id k3FNWLS0042878;
	Sat, 15 Apr 2006 19:32:24 -0400 (EDT)
	(envelope-from randall@lakerest.net)
DKIM-Signature: a=rsa-sha1; c=simple/simple; d=lakerest.net; s=lakerest;
	t=1145143944; h=Message-ID:Date:From:User-Agent:X-Accept-Language:
	MIME-Version:To:CC:Subject:References:In-Reply-To:Content-Type:
	Content-Transfer-Encoding; b=2SdWFrBTo05RgyBkQ5oeQO1xChQ/t/QYscvQyq
	/NEabke4kflFxJVslK+gSGD0GW0R2VpS8f07dryME5e47Cew==
Message-ID: <4441828A.8080702@lakerest.net>
Date: Sat, 15 Apr 2006 19:32:26 -0400
From: Randall Stewart <randall@lakerest.net>
User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060223
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: bidulock@openss7.org
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>	<443A2BA5.4000100@cisco.com>
	<443A78A3.5010508@middle.net>	<20060410112605.A29853@openss7.org>	<443CDB61.3030609@cisco.com>
	<20060412053212.A8295@openss7.org>	<443CE833.9000000@cisco.com>
	<20060412062121.B9084@openss7.org>	<443F86B8.4020500@cisco.com>
	<20060415064516.B3390@openss7.org>
In-Reply-To: <20060415064516.B3390@openss7.org>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.1 (/)
X-Scan-Signature: 69a74e02bbee44ab4f8eafdbcedd94a1
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>,
	Randall Stewart <rrs@cisco.com>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Brian F. G. Bidulock wrote:
>
> I was not frustrated in the least, and I believe that the
> exercise was a very good learning experience for the student, and
> the problems with the alpha implementation that were found more
> than compensated for the small fee that was paid to sponsor the
> student.  I am thankful that Professor Tuexen was able to make
> the arrangement possible.

My apologies..  I was NOT trying to single your implementation
out... in fact if you were not frustrated ... I was.. I dislike
seeing any new implementation's NOT have 100% success....

I want every implemenation to succeed...

This is why we MUST be very very careful in changing MUST's -> SHOULD
or SHOULD -> MUST..

If we change a MUST to a SHOULD .. we may well end up breaking
many implemenations that exist and only find it when a new
implemenation comes along.. which sets a very very bad precedent...

If we change a SHOULD to a MUST we can be in the same boat..

Changing the chunk ordering rules is a crutial thing that need
be considered with extreme care...

There may be no good reason for having Control come before Data..
I can't remember where or why that came into the protocol.. but
there is equally NO good reason not to have it be that way..
(except of course the reason you gave in your first response
  to Mark which I thought was very good :-D) especially when
many implementations are out there..

R
-- 
Randall Stewart
803-345-0369 <or> 815-342-5222(cell)




From tsvwg-bounces@ietf.org Sun Apr 16 06:21:45 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FV4Nn-0007ws-GW; Sun, 16 Apr 2006 06:21:31 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FV4Nn-0007wl-43
	for tsvwg@ietf.org; Sun, 16 Apr 2006 06:21:31 -0400
Received: from mail-n.franken.de ([193.175.24.27] helo=ilsa.franken.de)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FV4Nm-000276-Jj
	for tsvwg@ietf.org; Sun, 16 Apr 2006 06:21:31 -0400
Received: from [192.168.1.2] (p5481F03D.dip.t-dialin.net [84.129.240.61])
	by ilsa.franken.de (Postfix) with ESMTP
	id 39939245C7; Sun, 16 Apr 2006 12:21:27 +0200 (CEST)
	(KNF account authenticated via SMTP-AUTH)
In-Reply-To: <4441828A.8080702@lakerest.net>
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>	<443A2BA5.4000100@cisco.com>
	<443A78A3.5010508@middle.net>	<20060410112605.A29853@openss7.org>	<443CDB61.3030609@cisco.com>
	<20060412053212.A8295@openss7.org>	<443CE833.9000000@cisco.com>
	<20060412062121.B9084@openss7.org>	<443F86B8.4020500@cisco.com>
	<20060415064516.B3390@openss7.org> <4441828A.8080702@lakerest.net>
Mime-Version: 1.0 (Apple Message framework v749.3)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <32C89BE8-CDD0-4F5D-B8E6-0B9143F191B7@lurchi.franken.de>
Content-Transfer-Encoding: 7bit
From: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Date: Sun, 16 Apr 2006 12:21:28 +0200
To: Randall Stewart <randall@lakerest.net>
X-Mailer: Apple Mail (2.749.3)
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 3a4bc66230659131057bb68ed51598f8
Cc: bidulock@openss7.org, Randall Stewart <rrs@cisco.com>,
	TSWG <tsvwg@ietf.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Dear all,

RFC 2960 states that control chunks MUST put before DATA chunks
in SCTP packets.
Using the typographical conventions used in RFC 2960 this means only
that all DATA chunks (these are chunks with chunk type 0) MUST be
put before control chunks. Also the set of control chunks is mentioned,
but RFC 2960 does NOT define what control chunks are.
So we should fix this, by explicitly defining the notion control chunk.
 From my point of view it is clear that all chunk types defined in  
RFC 2960
except for DATA chunks are control chunks and as far as I know all  
implementations
have like this. Unfortunately, most implementations will consider all  
chunks
except DATA chunks as control chunks, but this is not true. One  
simply does not
know if an unknown chunk is a control chunks or not (it is definitely  
not
a DATA chunk). So an implementation should not consider an unknown  
chunk after
a DATA a protocol violation or so. I think this needs to be clarified.

If someone, like Mark, wants to define a new chunk which carries user  
data,
call it a N-DATA chunk, one has to define if it is a control chunk  
(probably not)
and if not, what kind of ordering restrictions exist. One could even  
state that
N-DATA can be anywhere in the packet. Using the supported chunk type  
parameter
and only sending them if the peer has listed the chunk type N-DATA in  
its
supported chunk types parameter will make sure that there are no interop
problems. Even in the case the peer does not understand it and  
consider it a control
chunk (which is wrong, but most implementation do today) this will  
not break
interoperability with current implementations.

So let us make clear in RFC 2960bis or in an IG what the notion of  
control chunk
is and what the correct behaviour on reception of unsupported chunks  
is. We will
test this at the next interop.

Now coming to the subject of the e-mail... SCTP PAD chunk. It is  
intended to
support PMTU with current implementations with as little impact as  
possible. So
I think the best is to define it as a control chunk. This makes it  
impossible
to perform Marks idea of 64/128 bit alignment.
I (personally) think that this is not critical, because I can get the  
64/128 bit
alignment by
- negotiating at the association setup what kind of alignment I want
   (the default 32bit, 64 bit, 128 bit) or whatever.
- if both sides supports this, all chunks (not only DATA) will be  
padded to
   this length. Fr 64 bit this would mean that we use 0-7 padding bytes.
- If we need to put the first chunk on some boundary, we can use the PAD
   chunk for this (except we want to AUTH it).

Mark, wouldn't give this you the alignment you want?

The only this we need to specify for this is that the PAD chunk will  
not the padded. Such
that it can be 32 bit, even if 64 bit alignment is wanted.

Best regards
Michael




On Apr 16, 2006, at 1:32 AM, Randall Stewart wrote:

> Brian F. G. Bidulock wrote:
>>
>> I was not frustrated in the least, and I believe that the
>> exercise was a very good learning experience for the student, and
>> the problems with the alpha implementation that were found more
>> than compensated for the small fee that was paid to sponsor the
>> student.  I am thankful that Professor Tuexen was able to make
>> the arrangement possible.
>
> My apologies..  I was NOT trying to single your implementation
> out... in fact if you were not frustrated ... I was.. I dislike
> seeing any new implementation's NOT have 100% success....
>
> I want every implemenation to succeed...
>
> This is why we MUST be very very careful in changing MUST's -> SHOULD
> or SHOULD -> MUST..
>
> If we change a MUST to a SHOULD .. we may well end up breaking
> many implemenations that exist and only find it when a new
> implemenation comes along.. which sets a very very bad precedent...
>
> If we change a SHOULD to a MUST we can be in the same boat..
>
> Changing the chunk ordering rules is a crutial thing that need
> be considered with extreme care...
>
> There may be no good reason for having Control come before Data..
> I can't remember where or why that came into the protocol.. but
> there is equally NO good reason not to have it be that way..
> (except of course the reason you gave in your first response
>  to Mark which I thought was very good :-D) especially when
> many implementations are out there..
>
> R
> -- 
> Randall Stewart
> 803-345-0369 <or> 815-342-5222(cell)
>





From tsvwg-bounces@ietf.org Sun Apr 16 08:23:59 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FV6I0-0005ta-HA; Sun, 16 Apr 2006 08:23:40 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FV6Hz-0005tV-FU
	for tsvwg@ietf.org; Sun, 16 Apr 2006 08:23:39 -0400
Received: from sj-iport-4.cisco.com ([171.68.10.86])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FV6Hy-0005Wf-UM
	for tsvwg@ietf.org; Sun, 16 Apr 2006 08:23:39 -0400
Received: from sj-core-3.cisco.com ([171.68.223.137])
	by sj-iport-4.cisco.com with ESMTP; 16 Apr 2006 05:23:38 -0700
X-IronPort-AV: i="4.04,123,1144047600"; 
	d="scan'208"; a="1795503973:sNHT33500792"
Received: from [IPv6:::1] (ssh-sjc-1.cisco.com [171.68.225.134])
	by sj-core-3.cisco.com (8.12.10/8.12.6) with ESMTP id k3GCNYVI021954;
	Sun, 16 Apr 2006 05:23:34 -0700 (PDT)
Message-ID: <44423749.2020301@cisco.com>
Date: Sun, 16 Apr 2006 08:23:37 -0400
From: Randall Stewart <rrs@cisco.com>
User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060223
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>	<443A2BA5.4000100@cisco.com>
	<443A78A3.5010508@middle.net>	<20060410112605.A29853@openss7.org>	<443CDB61.3030609@cisco.com>
	<20060412053212.A8295@openss7.org>	<443CE833.9000000@cisco.com>
	<20060412062121.B9084@openss7.org>	<443F86B8.4020500@cisco.com>
	<20060415064516.B3390@openss7.org> <4441828A.8080702@lakerest.net>
	<32C89BE8-CDD0-4F5D-B8E6-0B9143F191B7@lurchi.franken.de>
In-Reply-To: <32C89BE8-CDD0-4F5D-B8E6-0B9143F191B7@lurchi.franken.de>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 827a2a57ca7ab0837847220f447e8d56
Cc: bidulock@openss7.org, TSWG <tsvwg@ietf.org>,
	Randall Stewart <randall@lakerest.net>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Michael:

Though maybe a bit fuzzy.. I think your defintion is already
in 2960..

------------------------

    o  SCTP packet (or packet): The unit of data delivery across the
       interface between SCTP and the connectionless packet network
       (e.g., IP).  An SCTP packet includes the common SCTP header,
       possible SCTP control chunks, and user data encapsulated within
       SCTP DATA chunks.

----------

Notice.. it classify's all chunks as control except data.

Also there is a picture (in text) later on showing data
chunks coming in on one side and control on the other
(page 68)..

I think these defintions are good enough.... no implementor
has been confused by these...

I would agree that:
a) we need to test for this at inter-op's aka unknown chunks
    amongst the data.
b) we should probably test valid control amongst data

and

c) All future (including current drafts) extensions should declare
    that the chunk is either control or data... or as you
    any special appearance rules..


R

Michael Tuexen wrote:
> Dear all,
> 
> RFC 2960 states that control chunks MUST put before DATA chunks
> in SCTP packets.
> Using the typographical conventions used in RFC 2960 this means only
> that all DATA chunks (these are chunks with chunk type 0) MUST be
> put before control chunks. Also the set of control chunks is mentioned,
> but RFC 2960 does NOT define what control chunks are.
> So we should fix this, by explicitly defining the notion control chunk.
>  From my point of view it is clear that all chunk types defined in  RFC 
> 2960
> except for DATA chunks are control chunks and as far as I know all  
> implementations
> have like this. Unfortunately, most implementations will consider all  
> chunks
> except DATA chunks as control chunks, but this is not true. One  simply 
> does not
> know if an unknown chunk is a control chunks or not (it is definitely  not
> a DATA chunk). So an implementation should not consider an unknown  
> chunk after
> a DATA a protocol violation or so. I think this needs to be clarified.
> 
> If someone, like Mark, wants to define a new chunk which carries user  
> data,
> call it a N-DATA chunk, one has to define if it is a control chunk  
> (probably not)
> and if not, what kind of ordering restrictions exist. One could even  
> state that
> N-DATA can be anywhere in the packet. Using the supported chunk type  
> parameter
> and only sending them if the peer has listed the chunk type N-DATA in  its
> supported chunk types parameter will make sure that there are no interop
> problems. Even in the case the peer does not understand it and  consider 
> it a control
> chunk (which is wrong, but most implementation do today) this will  not 
> break
> interoperability with current implementations.
> 
> So let us make clear in RFC 2960bis or in an IG what the notion of  
> control chunk
> is and what the correct behaviour on reception of unsupported chunks  
> is. We will
> test this at the next interop.
> 
> Now coming to the subject of the e-mail... SCTP PAD chunk. It is  
> intended to
> support PMTU with current implementations with as little impact as  
> possible. So
> I think the best is to define it as a control chunk. This makes it  
> impossible
> to perform Marks idea of 64/128 bit alignment.
> I (personally) think that this is not critical, because I can get the  
> 64/128 bit
> alignment by
> - negotiating at the association setup what kind of alignment I want
>   (the default 32bit, 64 bit, 128 bit) or whatever.
> - if both sides supports this, all chunks (not only DATA) will be  
> padded to
>   this length. Fr 64 bit this would mean that we use 0-7 padding bytes.
> - If we need to put the first chunk on some boundary, we can use the PAD
>   chunk for this (except we want to AUTH it).
> 
> Mark, wouldn't give this you the alignment you want?
> 
> The only this we need to specify for this is that the PAD chunk will  
> not the padded. Such
> that it can be 32 bit, even if 64 bit alignment is wanted.
> 
> Best regards
> Michael
> 
> 
> 
> 
> On Apr 16, 2006, at 1:32 AM, Randall Stewart wrote:
> 
>> Brian F. G. Bidulock wrote:
>>
>>>
>>> I was not frustrated in the least, and I believe that the
>>> exercise was a very good learning experience for the student, and
>>> the problems with the alpha implementation that were found more
>>> than compensated for the small fee that was paid to sponsor the
>>> student.  I am thankful that Professor Tuexen was able to make
>>> the arrangement possible.
>>
>>
>> My apologies..  I was NOT trying to single your implementation
>> out... in fact if you were not frustrated ... I was.. I dislike
>> seeing any new implementation's NOT have 100% success....
>>
>> I want every implemenation to succeed...
>>
>> This is why we MUST be very very careful in changing MUST's -> SHOULD
>> or SHOULD -> MUST..
>>
>> If we change a MUST to a SHOULD .. we may well end up breaking
>> many implemenations that exist and only find it when a new
>> implemenation comes along.. which sets a very very bad precedent...
>>
>> If we change a SHOULD to a MUST we can be in the same boat..
>>
>> Changing the chunk ordering rules is a crutial thing that need
>> be considered with extreme care...
>>
>> There may be no good reason for having Control come before Data..
>> I can't remember where or why that came into the protocol.. but
>> there is equally NO good reason not to have it be that way..
>> (except of course the reason you gave in your first response
>>  to Mark which I thought was very good :-D) especially when
>> many implementations are out there..
>>
>> R
>> -- 
>> Randall Stewart
>> 803-345-0369 <or> 815-342-5222(cell)
>>
> 


-- 
Randall Stewart
ITD - Cisco Systems Inc.
803-345-0369 <or> 815-342-5222 (cell)




From tsvwg-bounces@ietf.org Sun Apr 16 12:19:38 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FV9y2-0006lD-KO; Sun, 16 Apr 2006 12:19:18 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FV9y1-0006l8-F8
	for tsvwg@ietf.org; Sun, 16 Apr 2006 12:19:17 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FV9xy-0005md-3D
	for tsvwg@ietf.org; Sun, 16 Apr 2006 12:19:15 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FV9xy-00034X-4q; Sun, 16 Apr 2006 10:18:52 -0600
Message-ID: <44426E89.9050204@middle.net>
Date: Sun, 16 Apr 2006 10:19:21 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>	<443A2BA5.4000100@cisco.com>
	<443A78A3.5010508@middle.net>	<20060410112605.A29853@openss7.org>	<443CDB61.3030609@cisco.com>
	<20060412053212.A8295@openss7.org>	<443CE833.9000000@cisco.com>
	<20060412062121.B9084@openss7.org>	<443F86B8.4020500@cisco.com>
	<20060415064516.B3390@openss7.org> <4441828A.8080702@lakerest.net>
	<32C89BE8-CDD0-4F5D-B8E6-0B9143F191B7@lurchi.franken.de>
In-Reply-To: <32C89BE8-CDD0-4F5D-B8E6-0B9143F191B7@lurchi.franken.de>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 769a46790fb42fbb0b0cc700c82f7081
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Michael Tuexen wrote:

>
> Now coming to the subject of the e-mail... SCTP PAD chunk. It is  
> intended to
> support PMTU with current implementations with as little impact as  
> possible. So
> I think the best is to define it as a control chunk. This makes it  
> impossible
> to perform Marks idea of 64/128 bit alignment.
> I (personally) think that this is not critical, because I can get the  
> 64/128 bit
> alignment by
> - negotiating at the association setup what kind of alignment I want
>   (the default 32bit, 64 bit, 128 bit) or whatever.
> - if both sides supports this, all chunks (not only DATA) will be  
> padded to
>   this length. Fr 64 bit this would mean that we use 0-7 padding bytes.
> - If we need to put the first chunk on some boundary, we can use the PAD
>   chunk for this (except we want to AUTH it).
>
> Mark, wouldn't give this you the alignment you want?

With the qualifier that I readily concede the issue with regard to 
backward compatibility, would not have brought this up in the first 
place if I had known about it, and do not want to cause any more 
distress about a relatively obscure application of arbitrary chunk 
ordering, I would add the following:

1. A negotiated alignment of that type aligns the chunk header, not the 
payload.
2. Even if (1) weren't a problem, supporting arbitrary 2^N chunk 
alignment is a rather intrusive change, where simply skipping over an 
embedded pad chunk is trivial.
3. Consequently, in the environments where support for this would be 
beneficial, an option that simply indicates that an endpoint supports 
arbitrary chunk ordering would be adequate.  No conditional code on the 
receive side processing path required.

But as far as the PAD chunk is concerned here and now, I have no 
objection to dropping the issue.

Thanks,

   Mark B.

>





From tsvwg-bounces@ietf.org Sun Apr 16 18:27:29 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FVFhx-0004mR-Pe; Sun, 16 Apr 2006 18:27:05 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FVFhw-0004mM-Nb
	for tsvwg@ietf.org; Sun, 16 Apr 2006 18:27:04 -0400
Received: from mail1.sea5.speakeasy.net ([69.17.117.3])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FVFht-0002A4-9p
	for tsvwg@ietf.org; Sun, 16 Apr 2006 18:27:02 -0400
Received: (qmail 20678 invoked from network); 16 Apr 2006 22:27:00 -0000
Received: from mm.asomi.com (HELO [66.92.45.138]) ([66.92.45.138])
	(envelope-sender <cait@asomi.com>)
	by mail1.sea5.speakeasy.net (qmail-ldap-1.03) with RC4-SHA encrypted
	SMTP for <bidulock@openss7.org>; 16 Apr 2006 22:26:59 -0000
In-Reply-To: <4441828A.8080702@lakerest.net>
References: <EB3CCF07FFD3CA418B4700C0C47EE74B8C4BC5@MCHP7I6A.ww002.siemens.net>	<443A2BA5.4000100@cisco.com>
	<443A78A3.5010508@middle.net>	<20060410112605.A29853@openss7.org>	<443CDB61.3030609@cisco.com>
	<20060412053212.A8295@openss7.org>	<443CE833.9000000@cisco.com>
	<20060412062121.B9084@openss7.org>	<443F86B8.4020500@cisco.com>
	<20060415064516.B3390@openss7.org> <4441828A.8080702@lakerest.net>
Mime-Version: 1.0 (Apple Message framework v746.3)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <149E45FC-8C5A-42DB-BF62-95D40AF394B8@asomi.com>
Content-Transfer-Encoding: 7bit
From: speakeasy <cait@asomi.com>
Subject: Re: [Tsvwg] Re: SCTP PAD chunk
Date: Sun, 16 Apr 2006 15:26:57 -0700
To: Randall Stewart <randall@lakerest.net>
X-Mailer: Apple Mail (2.746.3)
X-Spam-Score: 0.0 (/)
X-Scan-Signature: a7d6aff76b15f3f56fcb94490e1052e4
Cc: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, bidulock@openss7.org,
	Randall Stewart <rrs@cisco.com>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org


On Apr 15, 2006, at 4:32 PM, Randall Stewart wrote:

> Brian F. G. Bidulock wrote:
>>
>> I was not frustrated in the least, and I believe that the
>> exercise was a very good learning experience for the student, and
>> the problems with the alpha implementation that were found more
>> than compensated for the small fee that was paid to sponsor the
>> student.  I am thankful that Professor Tuexen was able to make
>> the arrangement possible.
>
> My apologies..  I was NOT trying to single your implementation
> out... in fact if you were not frustrated ... I was.. I dislike
> seeing any new implementation's NOT have 100% success....
>
> I want every implemenation to succeed...
>
> This is why we MUST be very very careful in changing MUST's -> SHOULD
> or SHOULD -> MUST..
>
> If we change a MUST to a SHOULD .. we may well end up breaking
> many implemenations that exist and only find it when a new
> implemenation comes along.. which sets a very very bad precedent...
>
> If we change a SHOULD to a MUST we can be in the same boat..
>
> Changing the chunk ordering rules is a crutial thing that need
> be considered with extreme care...
>
> There may be no good reason for having Control come before Data..
> I can't remember where or why that came into the protocol.. but
> there is equally NO good reason not to have it be that way..
> (except of course the reason you gave in your first response
>  to Mark which I thought was very good :-D) especially when
> many implementations are out there..
>

In that context, why couldn't a Pad chunk be defined as a data
chunk that is delivered to the bit bucket rather than as a control
chunk that had no effect?





From tsvwg-bounces@ietf.org Mon Apr 17 03:18:22 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FVNzn-0000sU-KG; Mon, 17 Apr 2006 03:18:03 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FUXUe-0003Uw-8W
	for tsvwg@ietf.org; Fri, 14 Apr 2006 19:14:24 -0400
Received: from mail.vidiator.com ([67.137.165.130]
	helo=sea1corpmail01.mediator.com)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FUXUc-0005of-W5
	for tsvwg@ietf.org; Fri, 14 Apr 2006 19:14:24 -0400
x-mimeole: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C66019.02B75364"
Date: Fri, 14 Apr 2006 16:13:16 -0700
Message-ID: <6582B762AB1433448C02EC6EA285B7E4F33293@sea1corpmail01.mediator.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: discussion about draft-watson-tsvwg-fec-sf-00.txt 
Thread-Index: AcZgGP+MOY3wOtMxT4+MzoLEbcj3ww==
From: "Glenn Hirschowitz" <glennh@vidiator.com>
To: <tsvwg@ietf.org>
X-Spam-Score: 0.1 (/)
X-Scan-Signature: 6e922792024732fb1bb6f346e63517e4
X-Mailman-Approved-At: Mon, 17 Apr 2006 03:18:01 -0400
Subject: [Tsvwg] discussion about draft-watson-tsvwg-fec-sf-00.txt 
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

This is a multi-part message in MIME format.

------_=_NextPart_001_01C66019.02B75364
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

It appeared in this draft dated July of 2005, that there was some push
to flesh out FEC oti parameter passing for streaming. However, the
momentum behind this draft seems to have waned. I was wondering if this
thread is still alive because I wish to pass FEC implementation specific
information from sender to receiver for streaming. Could you please let
me know where things stand, and if there some new documents where I
might find them?

=20

Thanks,

Glenn Hirschowitz


------_=_NextPart_001_01C66019.02B75364
Content-Type: text/html;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

<html xmlns:o=3D"urn:schemas-microsoft-com:office:office" =
xmlns:w=3D"urn:schemas-microsoft-com:office:word" =
xmlns=3D"http://www.w3.org/TR/REC-html40">

<head>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Dus-ascii">
<meta name=3DGenerator content=3D"Microsoft Word 11 (filtered medium)">
<style>
<!--
 /* Style Definitions */
 p.MsoNormal, li.MsoNormal, div.MsoNormal
	{margin:0in;
	margin-bottom:.0001pt;
	font-size:12.0pt;
	font-family:"Times New Roman";}
a:link, span.MsoHyperlink
	{color:blue;
	text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
	{color:purple;
	text-decoration:underline;}
span.EmailStyle17
	{mso-style-type:personal-compose;
	font-family:Arial;
	color:windowtext;}
@page Section1
	{size:8.5in 11.0in;
	margin:1.0in 1.25in 1.0in 1.25in;}
div.Section1
	{page:Section1;}
-->
</style>

</head>

<body lang=3DEN-US link=3Dblue vlink=3Dpurple>

<div class=3DSection1>

<p class=3DMsoNormal><font size=3D2 face=3DArial><span =
style=3D'font-size:10.0pt;
font-family:Arial'>It appeared in this draft dated July of 2005, that =
there was
some push to flesh out FEC oti parameter passing for streaming. However, =
the
momentum behind this draft seems to have waned. I was wondering if this =
thread
is still alive because I wish to pass FEC implementation specific =
information
from sender to receiver for streaming. Could you please let me know =
where
things stand, and if there some new documents where I might find =
them?<o:p></o:p></span></font></p>

<p class=3DMsoNormal><font size=3D2 face=3DArial><span =
style=3D'font-size:10.0pt;
font-family:Arial'><o:p>&nbsp;</o:p></span></font></p>

<p class=3DMsoNormal><font size=3D2 face=3DArial><span =
style=3D'font-size:10.0pt;
font-family:Arial'>Thanks,<o:p></o:p></span></font></p>

<p class=3DMsoNormal><font size=3D2 face=3DArial><span =
style=3D'font-size:10.0pt;
font-family:Arial'>Glenn Hirschowitz<o:p></o:p></span></font></p>

</div>

</body>

</html>

------_=_NextPart_001_01C66019.02B75364--




From tsvwg-bounces@ietf.org Mon Apr 17 03:23:55 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FVO5M-0002aT-Sx; Mon, 17 Apr 2006 03:23:48 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FVO5L-0002aO-4a
	for tsvwg@ietf.org; Mon, 17 Apr 2006 03:23:47 -0400
Received: from mailgw3.ericsson.se ([193.180.251.60])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FVO5I-0005nH-FD
	for tsvwg@ietf.org; Mon, 17 Apr 2006 03:23:47 -0400
Received: from esealmw129.eemea.ericsson.se (unknown [153.88.254.120])
	by mailgw3.ericsson.se (Symantec Mail Security) with ESMTP id E03961B3; 
	Mon, 17 Apr 2006 09:23:42 +0200 (CEST)
Received: from esealmw127.eemea.ericsson.se ([153.88.254.171]) by
	esealmw129.eemea.ericsson.se with Microsoft SMTPSVC(6.0.3790.1830); 
	Mon, 17 Apr 2006 09:23:42 +0200
Received: from [153.88.16.50] ([153.88.16.50]) by esealmw127.eemea.ericsson.se
	with Microsoft SMTPSVC(6.0.3790.1830); 
	Mon, 17 Apr 2006 09:23:42 +0200
Message-ID: <4443427E.5010107@ericsson.com>
Date: Mon, 17 Apr 2006 09:23:42 +0200
From: Magnus Westerlund <magnus.westerlund@ericsson.com>
User-Agent: Thunderbird 1.5 (Windows/20051201)
MIME-Version: 1.0
To: Glenn Hirschowitz <glennh@vidiator.com>
Subject: Re: [Tsvwg] discussion about draft-watson-tsvwg-fec-sf-00.txt
References: <6582B762AB1433448C02EC6EA285B7E4F33293@sea1corpmail01.mediator.com>
In-Reply-To: <6582B762AB1433448C02EC6EA285B7E4F33293@sea1corpmail01.mediator.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-OriginalArrivalTime: 17 Apr 2006 07:23:42.0054 (UTC)
	FILETIME=[DAADD060:01C661EF]
X-Brightmail-Tracker: AAAAAA==
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 21c69d3cfc2dd19218717dbe1d974352
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Hi Glenn,

This has indeed gone forward. There is this week an ongoing IETF review 
of the proposal to form a new WG (FECFrame).
http://www1.ietf.org/mail-archive/web/ietf-announce/current/msg02415.html

There is also a mailing list:
https://www1.ietf.org/mailman/listinfo/fecframe

I would also like to comment that this work takes a wider scope than 
just the OTI information.

Cheers

Magnus


Glenn Hirschowitz wrote:
> It appeared in this draft dated July of 2005, that there was some push
> to flesh out FEC oti parameter passing for streaming. However, the
> momentum behind this draft seems to have waned. I was wondering if this
> thread is still alive because I wish to pass FEC implementation specific
> information from sender to receiver for streaming. Could you please let
> me know where things stand, and if there some new documents where I
> might find them?
> 
>  
> 
> Thanks,
> 
> Glenn Hirschowitz
> 
> 


-- 

Magnus Westerlund

Multimedia Technologies, Ericsson Research EAB/TVA/A
----------------------------------------------------------------------
Ericsson AB                | Phone +46 8 4048287
Torshamsgatan 23           | Fax   +46 8 7575550
S-164 80 Stockholm, Sweden | mailto: magnus.westerlund@ericsson.com




From tsvwg-bounces@ietf.org Mon Apr 17 11:09:12 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FVVLS-0005GH-3E; Mon, 17 Apr 2006 11:08:54 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FVVLQ-0005GC-7A
	for tsvwg@ietf.org; Mon, 17 Apr 2006 11:08:52 -0400
Received: from lennon.multicasttech.com ([63.105.122.7] helo=multicasttech.com)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FVVLP-0002c7-0V
	for tsvwg@ietf.org; Mon, 17 Apr 2006 11:08:52 -0400
Received: from [63.105.122.7] (account marshall_eubanks HELO [IPv6:::1])
	by multicasttech.com (CommuniGate Pro SMTP 3.4.8)
	with ESMTP id 4446502; Mon, 17 Apr 2006 11:07:39 -0400
In-Reply-To: <6582B762AB1433448C02EC6EA285B7E4F33293@sea1corpmail01.mediator.com>
References: <6582B762AB1433448C02EC6EA285B7E4F33293@sea1corpmail01.mediator.com>
Mime-Version: 1.0 (Apple Message framework v749.3)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <B9AAF9F0-7896-4EAD-87BE-D47EE29930D9@multicasttech.com>
Content-Transfer-Encoding: 7bit
From: Marshall Eubanks <tme@multicasttech.com>
Subject: Re: [Tsvwg] discussion about draft-watson-tsvwg-fec-sf-00.txt 
Date: Mon, 17 Apr 2006 11:09:08 -0400
To: Glenn Hirschowitz <glennh@vidiator.com>
X-Mailer: Apple Mail (2.749.3)
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 97adf591118a232206bdb5a27b217034
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Hello;

This draft is now under consideration by the FecFrame BOF, which  
hopes to become a WG in short order.
You might want to join the FecFrame mailing list

https://www1.ietf.org/mailman/listinfo/fecframe

and you might want to comment on the new charter

http://www1.ietf.org/mail-archive/web/ietf-announce/current/ 
msg02415.html

Regards
Marshall Eubanks

On Apr 14, 2006, at 7:13 PM, Glenn Hirschowitz wrote:

> It appeared in this draft dated July of 2005, that there was some  
> push to flesh out FEC oti parameter passing for streaming. However,  
> the momentum behind this draft seems to have waned. I was wondering  
> if this thread is still alive because I wish to pass FEC  
> implementation specific information from sender to receiver for  
> streaming. Could you please let me know where things stand, and if  
> there some new documents where I might find them?
>
>
>
> Thanks,
>
> Glenn Hirschowitz
>
>





From tsvwg-bounces@ietf.org Mon Apr 17 19:57:14 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FVdaY-0000CO-RY; Mon, 17 Apr 2006 19:57:02 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FVdaX-0000CJ-2C
	for tsvwg@ietf.org; Mon, 17 Apr 2006 19:57:01 -0400
Received: from louie.udel.edu ([128.4.40.12] helo=mail.eecis.udel.edu)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FVdaV-0005rv-SC
	for tsvwg@ietf.org; Mon, 17 Apr 2006 19:57:01 -0400
Received: by mail.eecis.udel.edu (Postfix, from userid 62)
	id C777D102; Mon, 17 Apr 2006 19:56:59 -0400 (EDT)
X-Spam-Checker-Version: SpamAssassin 3.1.1 (2006-03-10) on louie.udel.edu
X-Spam-Level: 
X-Spam-Status: No, score=-2.3 required=4.1 tests=ALL_TRUSTED,BAYES_00 
	autolearn=ham version=3.1.1
Received: from stimpy.eecis.udel.edu (stimpy.eecis.udel.edu [128.4.40.17])
	by mail.eecis.udel.edu (Postfix) with ESMTP id 5CC3B39
	for <tsvwg@ietf.org>; Mon, 17 Apr 2006 19:56:59 -0400 (EDT)
Date: Mon, 17 Apr 2006 19:56:59 -0400 (EDT)
From: Ilknur Aydin <aydin@mail.eecis.udel.edu>
X-X-Sender: aydin@stimpy.eecis.udel.edu
To: tsvwg@ietf.org
Message-ID: <Pine.SOC.4.64.0604171956070.22053@stimpy.eecis.udel.edu>
X-Sanitizer: This message has been sanitized!
X-Sanitizer-URL: http://mailtools.anomy.net/
X-Sanitizer-Rev: UDEL-ECECIS: Sanitizer.pm, v 1.64 2002/10/22 MIME-Version: 1.0
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset="US-ASCII"; format=flowed
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 798b2e660f1819ae38035ac1d8d5e3ab
Subject: [Tsvwg] Release 2.0 of SCTP QualNet module
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Hello All,

Release 2.0 of our SCTP module for QualNet(3.9) network simulator is available 
for download.

The module supports most of the features in RFC 2960 in addition to Dynamic
Address Reconfiguration (addip09.txt) and Partial Reliability (RFC 3758).


For more details and to download the module, visit:

http://degas.cis.udel.edu/SCTP/

Thanks.

Best Regards,

Ilknur Aydin
DEGAS Networking Group
Dept. of Computer and Information Sciences
University of Delaware
http://www.cis.udel.edu/~aydin




From tsvwg-bounces@ietf.org Tue Apr 18 05:38:31 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FVmf2-0000dL-FJ; Tue, 18 Apr 2006 05:38:16 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FVdPC-00047c-4r; Mon, 17 Apr 2006 19:45:18 -0400
Received: from louie.udel.edu ([128.4.40.12] helo=mail.eecis.udel.edu)
	by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FVdPA-0005ZJ-Ue; Mon, 17 Apr 2006 19:45:18 -0400
Received: by mail.eecis.udel.edu (Postfix, from userid 62)
	id B16D3FE; Mon, 17 Apr 2006 19:45:16 -0400 (EDT)
X-Spam-Checker-Version: SpamAssassin 3.1.1 (2006-03-10) on louie.udel.edu
X-Spam-Level: 
X-Spam-Status: No, score=-2.2 required=4.1 tests=ALL_TRUSTED,AWL,BAYES_00 
	autolearn=ham version=3.1.1
Received: from stimpy.eecis.udel.edu (stimpy.eecis.udel.edu [128.4.40.17])
	by mail.eecis.udel.edu (Postfix) with ESMTP id 4D08039;
	Mon, 17 Apr 2006 19:45:16 -0400 (EDT)
Date: Mon, 17 Apr 2006 19:45:16 -0400 (EDT)
From: Ilknur Aydin <aydin@mail.eecis.udel.edu>
X-X-Sender: aydin@stimpy.eecis.udel.edu
To: manet emailing list <manet@ietf.org>
In-Reply-To: <67043E463DDBFD4A8087ED940BFF756B9CDDCF@BRU0038A.ww018.siemens.net>
Message-ID: <Pine.SOC.4.64.0604171937060.21625@stimpy.eecis.udel.edu>
References: <67043E463DDBFD4A8087ED940BFF756B9CDDCF@BRU0038A.ww018.siemens.net>
X-Sanitizer: This message has been sanitized!
X-Sanitizer-URL: http://mailtools.anomy.net/
X-Sanitizer-Rev: UDEL-ECECIS: Sanitizer.pm, v 1.64 2002/10/22 MIME-Version: 1.0
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset="US-ASCII"; format=flowed
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 798b2e660f1819ae38035ac1d8d5e3ab
X-Mailman-Approved-At: Tue, 18 Apr 2006 05:38:14 -0400
Cc: 
Subject: [Tsvwg] Release 2.0 of SCTP QualNet module
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Hello All,

Release 2.0 of our SCTP module for QualNet(3.9) network simulator is available 
for download.

The module supports most of the features in RFC 2960 in addition to Dynamic
Address Reconfiguration (addip09.txt) and Partial Reliability (RFC 3758).


For more details and to download the module, visit:

http://degas.cis.udel.edu/SCTP/

Thanks.

Best Regards,

Ilknur Aydin
DEGAS Networking Group
Dept. of Computer and Information Sciences
University of Delaware
http://www.cis.udel.edu/~aydin




From tsvwg-bounces@ietf.org Tue Apr 18 10:38:36 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FVrLe-0002mG-Rx; Tue, 18 Apr 2006 10:38:34 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FVrLd-0002m6-Oe
	for tsvwg@ietf.org; Tue, 18 Apr 2006 10:38:33 -0400
Received: from hermes.hosts.co.uk ([212.84.175.24] helo=minerva.hosts.co.uk)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FVrLd-0002pp-A9
	for tsvwg@ietf.org; Tue, 18 Apr 2006 10:38:33 -0400
Received: from [69.138.71.89] (helo=[192.168.1.3])
	by minerva.hosts.co.uk with esmtpsa (TLSv1:RC4-SHA:128) (Exim 4.60)
	(envelope-from <carlberg@g11.org.uk>)
	id 1FVrLQ-0005Eu-ST; Tue, 18 Apr 2006 15:38:21 +0100
Mime-Version: 1.0 (Apple Message framework v749.3)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <C454E278-F94B-410C-92E0-D0C24692134B@g11.org.uk>
Content-Transfer-Encoding: 7bit
From: ken carlberg <carlberg@g11.org.uk>
Date: Tue, 18 Apr 2006 10:38:28 -0400
To: bob.briscoe@bt.com
X-Mailer: Apple Mail (2.749.3)
X-Spam-Score: -1.4 (-)
X-Spam-Score: 0.0 (/)
X-Scan-Signature: a87a9cdae4ac5d3fbeee75cd0026d632
Cc: tsvwg <tsvwg@ietf.org>
Subject: [Tsvwg] draft-briscoe-tsvwg-cl-architecture-02.txt
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Bob (and gang :-),

here are some comments/questions about your draft.

-ken
===================================

Page 4, Figure 1: Overall QoS Architecture.  Would there be a problem  
in extending the target architecture from that shown in Figure 1 to  
one that potentially could include two or more "islands" of CL- 
regions?  The reason I ask is that the CL-region over the internet is  
only as large as one can extend diff-serv to multiple autonomous  
domains, which is difficult at best.  But if one could accomplish  
what you set out to do between two or more leaf ISPs with non-diff- 
serv transit backbones, then it would seem that your design is more  
flexible, and possibly more tractable.

page 5, 3'rd paragraph.  This is definitely just a nit, but could you  
add some text to be more precise when you state that PCN "can  
supplement" any PHB?  The word "can" leaves the reader with a measure  
of doubt, and the possibility that the word "supplement" may involve  
more than just a superset of functionality.

page 6, 4'th paragraph.  When you mention the dependency on  
measurement, I was wondering if relatively short transaction flows  
becomes an issue with your design.  Phone calls today still last just  
~4.5 minutes in duration, while on the other side of the spectrum,  
SMS over telco's obviously are just brief blips.  If we extend these  
to the thousands and on, will a need arise for more than one  
measurement algorithm?  Note, this is just food for thought and  
doesn't really require an answer in your draft.  But this does lead  
to the next point....

page 6, 5'th paragraph.  Here, and in other parts of your draft, you  
state the use of EWMA to dampen and smoothen the measurements of CL  
traffic.  In reality, what you want is a moving average of some type  
-- it could be simple, weighted, non-weighted, etc.  But given that  
this is an architecture document, is it necessary to pin it down to  
one specific type of moving average?  perhaps EWMA could be the  
default, but even then one would eventually need to state the best  
common practice regarding its parameters.

page 7, 1'st paragraph.  In this segment, and throughout the draft,  
you state that interior nodes within the CL-region do not hold  
"state".  Isn't it more accurate to say that they do not hold "per- 
flow state"?  The diff-serv capable interior nodes still hold  
aggregate state, correct?

Small nit.  You use the term "nodes" throughout the draft.  When I  
hear the term node, I tend to view it as an IP capable object that  
could be either a router or a host/server.  And yet in that first  
paragraph of page 7, the "node" you speak of is just a router.  I  
think it would be better if you could use the term routers and hosts  
when it is clear that is what you are speaking about, instead of  
adding ambiguity when using the term "node".

Section 1.1.2 Flow Pre-emption.  When I first came across this  
section, the questions that came to mind were: where does this occur  
-- at the edge, or interior to the CL-region?...and how does it occur  
-- through the use of signaling, or in-band state via DCSP?  These  
questions didn't get (somewhat) answered until a hint is given on the  
last paragraph of Page 24 and Section 7 on Security.  I think you  
need to address these questions in Section 1.1.2

Also, when capacity constraints occur, would it be possible to  
introduce another model besides pre-emption that focuses on  
probability using engineered spare capacity?  Instead of a lengthy  
explanation written here, let me point you to
the following draft:

ftp://ftp.rfc-editor.org/in-notes/internet-drafts/draft-lefaucheur- 
emergency-rsvp-01.txt

and specifically, Section 2.1.3.  If you have further questions on  
the draft or my request, we can discuss that off-line.


Page 8, 1'st paragraph.  Small nit, but the first sentence is unclear.

Page 12, 4'th Item.  The timeline of 1-2 seconds really applies to  
voice.  User expectations for video would not seem to be under the  
same constraints.

Page 13, 1'st Item.  When you speak of "session priorities", are you  
referring to the priorities one may see from signaling protocols like  
SIP and the Resource Priority header?  Or are you referring to the  
session as a flow identified by RSVP?

Page 15, 2'nd Item.  You should either substitute "NSIS" with "GIST",  
or expand the current text to say "signaling protocol produced by the  
NSIS working group".

Page 21, 1'st Item & 1'st para.  You state "packets are pre-emption  
marked".  My question is, how and where does this occur?  Is this a  
specific (potentially new?) DSCP marking?  If so, it would be clearer  
to state this explicitly as opposed to referring to a new behavior.

Section 4.6 Tunneling.  Do you foresee marking the DSCP field in the  
outer packet header?

Section 6.1.  I ended up getting confused by what you are trying to  
say.  It may be helpful to break up the first sentence into 2 or more  
smaller ones.





From tsvwg-bounces@ietf.org Wed Apr 19 12:58:03 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FWG08-0000SN-Vx; Wed, 19 Apr 2006 12:58:00 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FWG07-0000SI-Tt
	for tsvwg@ietf.org; Wed, 19 Apr 2006 12:57:59 -0400
Received: from test-iport-1.cisco.com ([171.71.176.117])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FWG07-0004Nq-Io
	for tsvwg@ietf.org; Wed, 19 Apr 2006 12:57:59 -0400
Received: from sj-core-1.cisco.com ([171.71.177.237])
	by test-iport-1.cisco.com with ESMTP; 19 Apr 2006 09:57:59 -0700
Received: from xbh-sjc-231.amer.cisco.com (xbh-sjc-231.cisco.com
	[128.107.191.100])
	by sj-core-1.cisco.com (8.12.10/8.12.6) with ESMTP id k3JGvxw1028647
	for <tsvwg@ietf.org>; Wed, 19 Apr 2006 09:57:59 -0700 (PDT)
Received: from xfe-sjc-212.amer.cisco.com ([171.70.151.187]) by
	xbh-sjc-231.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Wed, 19 Apr 2006 09:57:58 -0700
Received: from jmpolk-wxp.cisco.com ([10.89.20.143]) by
	xfe-sjc-212.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Wed, 19 Apr 2006 09:57:58 -0700
Message-Id: <4.3.2.7.2.20060419111019.02937990@email.cisco.com>
X-Sender: jmpolk@email.cisco.com
X-Mailer: QUALCOMM Windows Eudora Version 4.3.2
Date: Wed, 19 Apr 2006 11:57:57 -0500
To: tsvwg <tsvwg@ietf.org>
From: "James M. Polk" <jmpolk@cisco.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
X-OriginalArrivalTime: 19 Apr 2006 16:57:58.0750 (UTC)
	FILETIME=[694BDFE0:01C663D2]
X-Spam-Score: 0.6 (/)
X-Scan-Signature: cd26b070c2577ac175cd3a6d878c6248
Subject: [Tsvwg] Fwd: IESG Statement: Normative and Informative References 
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

TSVWG Authors/Editors

A few words of guidance from the IESG Secretary about references in all IDs 
wanting to become RFCs (in case you missed this on the IETF announce list).

>From: IESG Secretary <iesg-secretary@ietf.org>
>Date: Wed, 19 Apr 2006 09:50:00 -0400
>Subject: IESG Statement: Normative and Informative References
>
>Normative and Informative References
>
>Nearly all RFCs contain citations to other documents, and these are
>listed in a References section near the end of the RFC. There are many
>styles for references, and the RFCs have one of their own. Please
>follow the reference style used in recent RFCs. Please note that for
>documents that have been assigned an STD or BCP number, the number must
>be included in the reference.
>
>Within an RFC, references to other documents fall into two general
>categories: "normative" and "informative". Normative references specify
>documents that must be read to understand or implement the technology
>in the new RFC, or whose technology must be present for the technology
>in the new RFC to work. An informative reference is not normative;
>rather, it only provides additional information. For example, an
>informative reference might provide background or historical
>information. Informative references are not required to implement the
>technology in the RFC.
>
>Note 1: Even references that are relevant only for optional features
>must be classified as normative if they meet the above conditions for
>normative references.
>
>Note 2: It is not considered necessary to cite basic specifications
>that may be safely assumed to be known to practitioners (for example,
>RFC 791 need not be cited in every specification that mentions IPv4).
>
>Note 3: The normative/informative distinction is relevant in
>any document that amounts to a technical specification, even
>if its intended status is Experimental or Informational.
>
>Note 4: Normative references in RFCs cannot be to "work in progress"
>documents such as Internet Drafts. Drafts with such references will
>not be published as RFCs until the references are also published.
>
>The distinction between normative and informative references is often
>important. The IETF standards process according to RFC 2026 and RFC 3967,
>and the RFC Editor publication process, both need to know whether a
>reference to a work in progress is normative. An RFC cannot be published
>until all of the documents that it lists as normative references have been
>published. In practice, this often results in the simultaneous publication
>of a group of interrelated RFCs.
>
>For these reasons, the IESG and the RFC Editor have established
>guidelines that will request separate reference lists for normative
>and informative references in Internet Drafts and RFCs. For example,
>if both types are present, there would be two reference subsections,
>numbered s.1 and s.2 for example:
>
>s.1. Normative References
>
>xxx
>...
>xxx
>
>s.2. Informative References
>
>xxx
>...
>xxx
>
>Of course, if there is only one type of reference, only one
>section is needed.
>
>The IESG
>
>_______________________________________________
>IETF-Announce mailing list
>IETF-Announce@ietf.org
>https://www1.ietf.org/mailman/listinfo/ietf-announce




From tsvwg-bounces@ietf.org Wed Apr 19 15:56:59 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FWImx-0008ON-9A; Wed, 19 Apr 2006 15:56:35 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FWImv-0008OC-Cm
	for tsvwg@ietf.org; Wed, 19 Apr 2006 15:56:33 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FWImt-0003vQ-CW
	for tsvwg@ietf.org; Wed, 19 Apr 2006 15:56:31 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FWImr-0003kt-2V; Wed, 19 Apr 2006 13:56:07 -0600
Message-ID: <444695F8.5020000@middle.net>
Date: Wed, 19 Apr 2006 13:56:40 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: "RS, Moorthy (STSD)" <rsm@hp.com>
References: <44465220.3070705@hp.com>
In-Reply-To: <44465220.3070705@hp.com>
Content-Type: multipart/alternative;
	boundary="------------030101010305030209080703"
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.6 (/)
X-Scan-Signature: 5011df3e2a27abcc044eaa15befcaa87
Cc: tsvwg@ietf.org
Subject: [Tsvwg] Re: Should SCTP send ABORT on its Listen Queue Full?
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

This is a multi-part message in MIME format.
--------------030101010305030209080703
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

RS, Moorthy (STSD) wrote:

> All,
>
> Currently, when SCTP INIT packet is received by server and when the 
> server's listen queue is full - ( i.e sk->sk_ack_backlog exceeds the 
> listen() value specified), SCTP server sends an ABORT packet to the 
> client. This means that the client gets a "connection refused" error - 
> when there is load on the server, and is not able to process its 
> incoming connections faster.
>
> While TCP behaves differently - it just drops the SYN packet, when its 
> listen queue is full. However, in this case, applications will retry 
> and connect eventually. Given that, most TCP clients do not retry on 
> "connection refused" errors, they would have problems while using SCTP 
> (TCP-style one-to-one sockets) and would need to change their code. (I 
> have test scenarios, where even after increasing listen() queue value 
> - 10 times of 10000 connection requests are ABORTed).
>
> I would like to see, SCTP too dropping the INIT packet and not sending 
> an ABORT, when its listen queue is full.
>
> Any thoughts of why that should not be the case?

I agree completely - SCTP should behave like TCP in this matter.  It do 
not believe RFC2960 has anything to say either way.  It might be nice 
someday to even have a control chunk indicating the expected accept 
queue wait time, such that an application could switch to a different 
mirror when a primary was overloaded.

[Cross posting to TSVWG, as this is really an SCTP socket issue, not an 
LKSCTP issue].

>
> Thanks
> Moorthy
>


--------------030101010305030209080703
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
  <title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
RS, Moorthy (STSD) wrote:<br>
<blockquote cite="mid44465220.3070705@hp.com" type="cite">
  <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
  <font face="Arial">All,<br>
  <br>
Currently, when SCTP INIT packet is received by server and when the
server's listen queue is full - ( i.e sk-&gt;sk_ack_backlog exceeds the
listen() value specified), SCTP server sends an ABORT packet to the
client. This means that the client gets a "connection refused" error -
when there is load on the server, and is not able to process its
incoming connections faster.<br>
  <br>
While TCP behaves differently - it just drops the SYN packet, when its
listen queue is full. However, in this case, applications will retry
and connect eventually. Given that, most TCP clients do not retry on
"connection refused" errors, they would have problems while using SCTP
(TCP-style one-to-one sockets) and would need to change their code. (I
have test scenarios, where even after increasing listen() queue value -
10 times of 10000 connection requests are ABORTed).<br>
  <br>
I would like to see, SCTP too dropping the INIT packet and not sending
an ABORT, when its listen queue is full.<br>
  <br>
Any thoughts of why that should not be the case?<br>
  </font></blockquote>
I agree completely - SCTP should behave like TCP in this matter.&nbsp; It do
not believe RFC2960 has anything to say either way.&nbsp; It might be nice
someday to even have a control chunk indicating the expected accept
queue wait time, such that an application could switch to a different
mirror when a primary was overloaded.<br>
<br>
[Cross posting to TSVWG, as this is really an SCTP socket issue, not an
LKSCTP issue].<br>
<blockquote cite="mid44465220.3070705@hp.com" type="cite"><font
 face="Arial"><br>
Thanks<br>
Moorthy<br>
  <br>
  </font></blockquote>
<br>
</body>
</html>

--------------030101010305030209080703--




From tsvwg-bounces@ietf.org Wed Apr 19 19:45:37 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FWMMY-00010y-NB; Wed, 19 Apr 2006 19:45:34 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FWMMX-00010d-6W
	for tsvwg@ietf.org; Wed, 19 Apr 2006 19:45:33 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FWMMV-00087B-PU
	for tsvwg@ietf.org; Wed, 19 Apr 2006 19:45:33 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:rz+7nBCIIFuPCPfpCmLpIToIrJxHIs1u@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k3JNjUq18891;
	Wed, 19 Apr 2006 17:45:30 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k3JNjTr12900;
	Wed, 19 Apr 2006 17:45:29 -0600
Date: Wed, 19 Apr 2006 17:45:29 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Mark Butler <butlerm@middle.net>
Subject: Re: [Tsvwg] Re: Should SCTP send ABORT on its Listen Queue Full?
Message-ID: <20060419174529.A12701@openss7.org>
Mail-Followup-To: Mark Butler <butlerm@middle.net>,
	"RS, Moorthy (STSD)" <rsm@hp.com>, tsvwg@ietf.org
References: <44465220.3070705@hp.com> <444695F8.5020000@middle.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <444695F8.5020000@middle.net>;
	from butlerm@middle.net on Wed, Apr 19, 2006 at 01:56:40PM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 538aad3a3c4f01d8b6a6477ca4248793
Cc: "RS, Moorthy \(STSD\)" <rsm@hp.com>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark,

The ABORT should include cause "out of resource."  If the sender translates
that to "connection refused" that is its error.  I return ECONNABORTED
instead of ECONNREFUSED.

IMO, in fitting with the purposes of SCTP, it is better to provide an
indication back (ECONNABORTED) rather than waiting some number of minutes
for the connection establishment process to time out (ETIMEDOUT) as in the
case of TCP.

These are implementation details.  The protocol already provides the
necessary ABORT cause codepoints.

--brian


Mark Butler wrote:                              (Wed, 19 Apr 2006 13:56:40)
> 
>    RS, Moorthy (STSD) wrote:
> 
>    All,
>    Currently,  when  SCTP  INIT packet is received by server and when the
>    server's  listen  queue is full - ( i.e sk->sk_ack_backlog exceeds the
>    listen()  value  specified),  SCTP server sends an ABORT packet to the
>    client. This means that the client gets a "connection refused" error -
>    when  there  is  load  on  the  server, and is not able to process its
>    incoming connections faster.
>    While TCP behaves differently - it just drops the SYN packet, when its
>    listen  queue  is full. However, in this case, applications will retry
>    and  connect  eventually. Given that, most TCP clients do not retry on
>    "connection refused" errors, they would have problems while using SCTP
>    (TCP-style one-to-one sockets) and would need to change their code. (I
>    have  test scenarios, where even after increasing listen() queue value
>    - 10 times of 10000 connection requests are ABORTed).
>    I would like to see, SCTP too dropping the INIT packet and not sending
>    an ABORT, when its listen queue is full.
>    Any thoughts of why that should not be the case?
>    I  agree  completely - SCTP should behave like TCP in this matter.  It
>    do  not  believe  RFC2960 has anything to say either way.  It might be
>    nice  someday  to  even  have  a control chunk indicating the expected
>    accept  queue  wait  time,  such that an application could switch to a
>    different mirror when a primary was overloaded.
>    [Cross  posting  to TSVWG, as this is really an SCTP socket issue, not
>    an LKSCTP issue].
> 
>      Thanks
>      Moorthy

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




From tsvwg-bounces@ietf.org Wed Apr 19 19:57:15 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FWMXh-0004gl-59; Wed, 19 Apr 2006 19:57:05 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FWMXf-0004gg-U8
	for tsvwg@ietf.org; Wed, 19 Apr 2006 19:57:03 -0400
Received: from palrel12.hp.com ([156.153.255.237])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FWMXe-0008Vv-LN
	for tsvwg@ietf.org; Wed, 19 Apr 2006 19:57:03 -0400
Received: from tardy.cup.hp.com (tardy.cup.hp.com [15.244.56.217])
	by palrel12.hp.com (Postfix) with ESMTP id 1059634399;
	Wed, 19 Apr 2006 16:57:00 -0700 (PDT)
Received: from hp.com (localhost [127.0.0.1])
	by tardy.cup.hp.com (8.9.3 (PHNE_28810)/8.9.3 SMKit7.02) with ESMTP id
	QAA13804; Wed, 19 Apr 2006 16:56:59 -0700 (PDT)
Message-ID: <4446CE4A.6050301@hp.com>
Date: Wed, 19 Apr 2006 16:56:58 -0700
From: Rick Jones <rick.jones2@hp.com>
User-Agent: Mozilla/5.0 (X11; U; HP-UX 9000/785; en-US; rv:1.6) Gecko/20040304
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: bidulock@openss7.org
Subject: Re: [Tsvwg] Re: Should SCTP send ABORT on its Listen Queue Full?
References: <44465220.3070705@hp.com> <444695F8.5020000@middle.net>
	<20060419174529.A12701@openss7.org>
In-Reply-To: <20060419174529.A12701@openss7.org>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 8b30eb7682a596edff707698f4a80f7d
Cc: "RS, Moorthy \(STSD\)" <rsm@hp.com>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Brian F. G. Bidulock wrote:
> Mark,
> 
> The ABORT should include cause "out of resource."  If the sender translates
> that to "connection refused" that is its error.  I return ECONNABORTED
> instead of ECONNREFUSED.

Is there (will there be) a writeup somewhere that says what an 
application should do with an ECONNABORTED (or ECONNREFUSED) on a 
connect() on an SCTP end-point?

> 
> IMO, in fitting with the purposes of SCTP, it is better to provide an
> indication back (ECONNABORTED) rather than waiting some number of minutes
> for the connection establishment process to time out (ETIMEDOUT) as in the
> case of TCP.

OK, I'll step-down from the il-informed peanut gallery and bite - 
perhaps on my own toes :) - why?

A TCP application that concerned about timely connection completion will 
do a non-blocking connect() and give-up if the socket doesn't become 
writable in the length of time it wants anyway.  Which it still must do 
even if connect() returns errors on listen queues being full because 
there is no guarantee that the connection request actually makes it to 
the remote to return the "queue full" indication in the first place. 
So, connect() returning an error on remote queue full doesn't really 
save the application programmer anything.

If the application does not especially care about how quickly the 
connect() completes, should it really be required to make N connect() 
calls to establish its connection when the remote queue is full? (Asked 
he who was ass-u-me-ing that is what would happen with those error 
returns...)  And just what value of N should it use?

sincerely,

rick jones





From tsvwg-bounces@ietf.org Wed Apr 19 20:33:41 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FWN71-0004V8-Kr; Wed, 19 Apr 2006 20:33:35 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FWN70-0004V3-JM
	for tsvwg@ietf.org; Wed, 19 Apr 2006 20:33:34 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FWN6z-0001gn-4b
	for tsvwg@ietf.org; Wed, 19 Apr 2006 20:33:34 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:QwJXS76PzhaCaZtxvfgsiGbMa4g3TSGG@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k3K0XWq19964;
	Wed, 19 Apr 2006 18:33:32 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k3K0XVL13575;
	Wed, 19 Apr 2006 18:33:31 -0600
Date: Wed, 19 Apr 2006 18:33:31 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Rick Jones <rick.jones2@hp.com>
Subject: Re: [Tsvwg] Re: Should SCTP send ABORT on its Listen Queue Full?
Message-ID: <20060419183331.A13026@openss7.org>
Mail-Followup-To: Rick Jones <rick.jones2@hp.com>,
	Mark Butler <butlerm@middle.net>, "RS, Moorthy (STSD)" <rsm@hp.com>,
	tsvwg@ietf.org
References: <44465220.3070705@hp.com> <444695F8.5020000@middle.net>
	<20060419174529.A12701@openss7.org> <4446CE4A.6050301@hp.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <4446CE4A.6050301@hp.com>;
	from rick.jones2@hp.com on Wed, Apr 19, 2006 at 04:56:58PM -0700
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 0fa76816851382eb71b0a882ccdc29ac
Cc: "RS, Moorthy \(STSD\)" <rsm@hp.com>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Rick,

Rick Jones wrote:                                                    (Wed, 19 Apr 2006 16:56:58)
> Brian F. G. Bidulock wrote:
> > Mark,
> > 
> > The ABORT should include cause "out of resource."  If the sender translates
> > that to "connection refused" that is its error.  I return ECONNABORTED
> > instead of ECONNREFUSED.
> 
> Is there (will there be) a writeup somewhere that says what an 
> application should do with an ECONNABORTED (or ECONNREFUSED) on a 
> connect() on an SCTP end-point?

Sorry, different part of the code (ERROR chunks in response to INIT-ACK).
If I receive an ABORT in COOKIE-WAIT or COOKIE-ECHOED state I return
ECONNREFUSED too.

But as the receiver I do not check max_ack_backlog on receipt of an INIT.
That strikes me as against the concept of not maintaining any state associated
with INITs and invites a denial of service attack.

When I receive a COOKIE-ECHO I check max_ack_backlog before queuing the
connection indication.  If it is exceeded I send ABORT "out of resource".
rather than silently discarding the COOKIE-ECHO (which would be against
RFC 2960).

I believe that ABORT(out of resource) in the COOKIE-WAIT state should be
because there is no listening socket or because the receiver simply cannot
process INITs at this time.  ABORT(out of resource) in the COOKIE-ECHOED
state should be because of ack_backlog issues.

So, rathert than dropping the COOKIE-ECHO chunk, as was suggested, if this
behaviour is important to you, the receiver should reattempt after receiving
ABORT() in the COOKIE-ECHOED state.  Note that it can even use the same
cookie until it expires.

> 
> > 
> > IMO, in fitting with the purposes of SCTP, it is better to provide an
> > indication back (ECONNABORTED) rather than waiting some number of minutes
> > for the connection establishment process to time out (ETIMEDOUT) as in the
> > case of TCP.
> 
> OK, I'll step-down from the il-informed peanut gallery and bite - 
> perhaps on my own toes :) - why?
> 
> A TCP application that concerned about timely connection completion will 
> do a non-blocking connect() and give-up if the socket doesn't become 
> writable in the length of time it wants anyway.  Which it still must do 
> even if connect() returns errors on listen queues being full because 
> there is no guarantee that the connection request actually makes it to 
> the remote to return the "queue full" indication in the first place. 
> So, connect() returning an error on remote queue full doesn't really 
> save the application programmer anything.
> 
> If the application does not especially care about how quickly the 
> connect() completes, should it really be required to make N connect() 
> calls to establish its connection when the remote queue is full? (Asked 
> he who was ass-u-me-ing that is what would happen with those error 
> returns...)  And just what value of N should it use?

I suppose that makes sense.  I'll change the openss7 sockets implementation
to retransmit the COOKIE-ECHO and give it a whirl at the interop.

--brian


> 
> sincerely,
> 
> rick jones

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




From tsvwg-bounces@ietf.org Wed Apr 19 21:47:36 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FWOGS-00040v-NC; Wed, 19 Apr 2006 21:47:24 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FWOGR-0003zx-E3
	for tsvwg@ietf.org; Wed, 19 Apr 2006 21:47:23 -0400
Received: from stsc1260-eth-s1-s1p1-vip.va.neustar.com ([156.154.16.129]
	helo=chiedprmail1.ietf.org)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FWMhN-0000J5-B8
	for tsvwg@ietf.org; Wed, 19 Apr 2006 20:07:05 -0400
Received: from gw.openss7.com ([142.179.199.224])
	by chiedprmail1.ietf.org with esmtp (Exim 4.43) id 1FWMTp-0002Ky-5X
	for tsvwg@ietf.org; Wed, 19 Apr 2006 19:53:06 -0400
Received: from ns.pigworks.openss7.net
	(IDENT:qaaaClZoGV6YoPXbKDCI1r8v/p8wEvBy@ns1.evil.openss7.net
	[192.168.9.1])
	by gw.openss7.com (8.11.6/8.11.6) with ESMTP id k3JNr2q19134;
	Wed, 19 Apr 2006 17:53:02 -0600
Received: (from brian@localhost)
	by ns.pigworks.openss7.net (8.11.6/8.11.6) id k3JNr1k12997;
	Wed, 19 Apr 2006 17:53:01 -0600
Date: Wed, 19 Apr 2006 17:53:01 -0600
From: "Brian F. G. Bidulock" <bidulock@openss7.org>
To: Mark Butler <butlerm@middle.net>
Subject: Re: [Tsvwg] Re: Should SCTP send ABORT on its Listen Queue Full?
Message-ID: <20060419175301.B12701@openss7.org>
Mail-Followup-To: Mark Butler <butlerm@middle.net>,
	"RS, Moorthy (STSD)" <rsm@hp.com>, tsvwg@ietf.org
References: <44465220.3070705@hp.com> <444695F8.5020000@middle.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <444695F8.5020000@middle.net>;
	from butlerm@middle.net on Wed, Apr 19, 2006 at 01:56:40PM -0600
Organization: http://www.openss7.org/
Dsn-Notification-To: <bidulock@openss7.org>
X-Spam-Score: -1.5 (-)
X-Scan-Signature: c1c65599517f9ac32519d043c37c5336
Cc: "RS, Moorthy \(STSD\)" <rsm@hp.com>, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: bidulock@openss7.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark,

Just one other thing.  Are you sure that max_ack_backlog is being checked
when an INIT is received?  I always respond to INIT with INIT-ACK if there
is a listening socket (as is spelled out in RFC2960).  I don't check
max_ack_backlog until a COOKIE-ECHO is received...

Anyway, the sctp-impl list is probably a better place for these discussions.

--brian

Mark Butler wrote:                             (Wed, 19 Apr 2006 13:56:40)
> 
>    RS, Moorthy (STSD) wrote:
> 
>    All,
>    Currently,  when  SCTP  INIT packet is received by server and when the
>    server's  listen  queue is full - ( i.e sk->sk_ack_backlog exceeds the
>    listen()  value  specified),  SCTP server sends an ABORT packet to the
>    client. This means that the client gets a "connection refused" error -
>    when  there  is  load  on  the  server, and is not able to process its
>    incoming connections faster.
>    While TCP behaves differently - it just drops the SYN packet, when its
>    listen  queue  is full. However, in this case, applications will retry
>    and  connect  eventually. Given that, most TCP clients do not retry on
>    "connection refused" errors, they would have problems while using SCTP
>    (TCP-style one-to-one sockets) and would need to change their code. (I
>    have  test scenarios, where even after increasing listen() queue value
>    - 10 times of 10000 connection requests are ABORTed).
>    I would like to see, SCTP too dropping the INIT packet and not sending
>    an ABORT, when its listen queue is full.
>    Any thoughts of why that should not be the case?
>    I  agree  completely - SCTP should behave like TCP in this matter.  It
>    do  not  believe  RFC2960 has anything to say either way.  It might be
>    nice  someday  to  even  have  a control chunk indicating the expected
>    accept  queue  wait  time,  such that an application could switch to a
>    different mirror when a primary was overloaded.
>    [Cross  posting  to TSVWG, as this is really an SCTP socket issue, not
>    an LKSCTP issue].
> 
>      Thanks
>      Moorthy

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




From tsvwg-bounces@ietf.org Thu Apr 20 00:30:13 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FWQnv-0007Xh-Q1; Thu, 20 Apr 2006 00:30:07 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FWQnt-0007Xc-MM
	for tsvwg@ietf.org; Thu, 20 Apr 2006 00:30:05 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FWQnp-0005p9-8e
	for tsvwg@ietf.org; Thu, 20 Apr 2006 00:30:05 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FWQnp-0004tW-1m; Wed, 19 Apr 2006 22:29:39 -0600
Message-ID: <44470E55.3060903@middle.net>
Date: Wed, 19 Apr 2006 22:30:13 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Rick Jones <rick.jones2@hp.com>
Subject: Re: [Tsvwg] Re: Should SCTP send ABORT on its Listen Queue Full?
References: <44465220.3070705@hp.com> <444695F8.5020000@middle.net>
	<20060419174529.A12701@openss7.org> <4446CE4A.6050301@hp.com>
In-Reply-To: <4446CE4A.6050301@hp.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: f607d15ccc2bc4eaf3ade8ffa8af02a0
Cc: "RS, Moorthy \(STSD\)" <rsm@hp.com>, bidulock@openss7.org, tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Rick Jones wrote:

> Brian F. G. Bidulock wrote:
>
>> Mark,
>>
>> The ABORT should include cause "out of resource."  If the sender 
>> translates
>> that to "connection refused" that is its error.  I return ECONNABORTED
>> instead of ECONNREFUSED.
>
>
> Is there (will there be) a writeup somewhere that says what an 
> application should do with an ECONNABORTED (or ECONNREFUSED) on a 
> connect() on an SCTP end-point?

That is up to the application.  What is important to specify is under 
what conditions those errors will be generated.

>
>>
>> IMO, in fitting with the purposes of SCTP, it is better to provide an
>> indication back (ECONNABORTED) rather than waiting some number of 
>> minutes
>> for the connection establishment process to time out (ETIMEDOUT) as 
>> in the
>> case of TCP.
>
>
> OK, I'll step-down from the il-informed peanut gallery and bite - 
> perhaps on my own toes :) - why?
>
> A TCP application that concerned about timely connection completion 
> will do a non-blocking connect() and give-up if the socket doesn't 
> become writable in the length of time it wants anyway.  Which it still 
> must do even if connect() returns errors on listen queues being full 
> because there is no guarantee that the connection request actually 
> makes it to the remote to return the "queue full" indication in the 
> first place. So, connect() returning an error on remote queue full 
> doesn't really save the application programmer anything.
>
> If the application does not especially care about how quickly the 
> connect() completes, should it really be required to make N connect() 
> calls to establish its connection when the remote queue is full? 
> (Asked he who was ass-u-me-ing that is what would happen with those 
> error returns...)  And just what value of N should it use?

The theory behind dropping an INIT or COOKIE ECHO instead of returning 
an ABORT when the accept queue is full is that way the initiator will 
timeout and retry in the SCTP layer, instead of requiring another level 
of connect / retry code in the application layer.

I agree that this is not an SCTP protocol issue per se, but it is 
definitely an SCTP socket issue.  The SCTP socket draft ought to 
describe what errors are returned under what conditions, whenever the 
errors or conditions diverge from POSIX specified behavior.

 - Mark






From tsvwg-bounces@ietf.org Fri Apr 21 01:28:25 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FWoBm-00042N-W8; Fri, 21 Apr 2006 01:28:18 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FWoBl-00042I-A2
	for tsvwg@ietf.org; Fri, 21 Apr 2006 01:28:17 -0400
Received: from web37407.mail.mud.yahoo.com ([209.191.87.60])
	by ietf-mx.ietf.org with smtp (Exim 4.43) id 1FWoBk-00007B-Dl
	for tsvwg@ietf.org; Fri, 21 Apr 2006 01:28:17 -0400
Received: (qmail 1730 invoked by uid 60001); 21 Apr 2006 05:28:15 -0000
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com;
	h=Message-ID:Received:Date:From:Subject:To:MIME-Version:Content-Type:Content-Transfer-Encoding;
	b=x8PrgGEH0WOJWK05UwNuXTiErIuwBa4zQ9CB/tSZsOJJgj/S+KZNFqBOVTSxzeka0nzbw4W1+kawDB1b45NCaP5Zb76hRTVtygY8YO1GsmyQgHXhQHZYBEY7sQPH5d7cPVYs3l+C+X/oseurR5ywbfKHC3y3APdya9FCmGwco7k=
	; 
Message-ID: <20060421052815.1728.qmail@web37407.mail.mud.yahoo.com>
Received: from [203.145.132.100] by web37407.mail.mud.yahoo.com via HTTP;
	Thu, 20 Apr 2006 22:28:15 PDT
Date: Thu, 20 Apr 2006 22:28:15 -0700 (PDT)
From: ash kat <ashwani_groups@yahoo.com>
To: tsvwg@ietf.org
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="0-1310778812-1145597295=:98111"
Content-Transfer-Encoding: 8bit
X-Spam-Score: 0.1 (/)
X-Scan-Signature: 5011df3e2a27abcc044eaa15befcaa87
Subject: [Tsvwg] ambiguity in handling invalid cookie in COOKIE ECHO chunks
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

--0-1310778812-1145597295=:98111
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit

Hi All
   
  I was looking at the behavior of SCTP stack when the COOKIE ECHO chunk contains the invalid cookie.
   
  The following section of RFC2960 are conflicting each other:
   
  Section 4 (Page# 50)
   Notes:
     1) If the State Cookie in the received COOKIE ECHO is invalid (i.e.,
      failed to pass the integrity check), the receiver MUST silently
      discard the packet.  Or, if the received State Cookie is expired
      (see Section 5.1.5), the receiver MUST send back an ERROR chunk.
      In either case, the receiver stays in the CLOSED state.
   
  Section 5.1 (Page# 52)
     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.
   
  Section 5.1.5 (Page# 56)
   
     2) Authenticate the State Cookie as one that it previously generated
      by comparing the computed MAC against the one carried in the State
      Cookie.  If this comparison fails, the SCTP packet, including the
      COOKIE ECHO and any DATA chunks, should be silently discarded.
   
   
  While section 2.47 of draft-ietf-tsvwg-sctpimpguide-16 has some modifications on this:
   
  2.47.  Sending of ABORT chunks
  2.47.1.  Description of the problem
     Section 5.1 of RFC2960 [7] requires that an ABORT chunk is sent in
   response to an INIT chunk when there is no listening end point.  To
   make port scanning harder someone might not want these ABORTs to be
   received by the sender of the INIT chunks.  Currently the only way to
   enforce this is by using a firewall discarding the packets containing
   the INIT chunks or the packets containing the ABORT chunks.  It is
   desirable that the same can be done without a middle box.
  2.47.2.  Text changes to the document
     ---------
   Old text: (Section 5.1)
   ---------
      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.
     ---------
   New text: (Section 5.1)
   ---------
        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 SHOULD respond
      with an ABORT chunk.
   
   
  But still the conflict is there.
   
  In my point of view there should be only one of these three actions, which should be specified when the SCTP receive the COOKIE ECHO chunk with an invalid cookie.
  i) MUST silently discard (section 4 of RFC2960)
  ii) SHOULD respond with an ABORT (section 2.47 of draft-ietf-tsvwg-sctpimpguide-16)
  iii) should be silently discarded. (section 5.1.5 of RFC2960)
   
    Please help me resolving this conflict.
   

  -K. Äsh
   
   
  ---------------------------------------------------------------------------------------
  Sooner or later, those who win are those who think they can.
  ---------------------------------------------------------------------------------------

		
---------------------------------
How low will we go? Check out Yahoo! Messenger’s low  PC-to-Phone call rates.
--0-1310778812-1145597295=:98111
Content-Type: text/html; charset=iso-8859-1
Content-Transfer-Encoding: 8bit

<DIV>Hi All</DIV>  <DIV>&nbsp;</DIV>  <DIV>I was looking at the behavior of SCTP stack when the COOKIE ECHO chunk contains the invalid cookie.</DIV>  <DIV>&nbsp;</DIV>  <DIV>The following section of RFC2960 are conflicting each other:</DIV>  <DIV>&nbsp;</DIV>  <DIV>Section 4 (Page# 50)<BR>&nbsp;&nbsp; Notes:</DIV>  <DIV>&nbsp;&nbsp; 1) If the State Cookie in the received COOKIE ECHO is invalid (i.e.,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; failed to pass the integrity check), <FONT color=#ff0000><STRONG>the receiver MUST silently<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; discard the packet</STRONG>.</FONT>&nbsp; Or, if the received State Cookie is expired<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; (see Section 5.1.5), the receiver MUST send back an ERROR chunk.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; In either case, the receiver stays in the CLOSED state.</DIV>  <DIV>&nbsp;</DIV>  <DIV>Section 5.1 (Page# 52)</DIV>  <DIV>&nbsp;&nbsp; If an endpoint receives an INIT, INIT ACK, or COOKIE ECHO chunk
 but<BR>&nbsp;&nbsp; decides not to establish the new association due to missing mandatory<BR>&nbsp;&nbsp; parameters in the received INIT or INIT ACK, invalid parameter<BR>&nbsp;&nbsp; values, or lack of local resources, <FONT color=#ff0000><STRONG>it MUST respond with an ABORT<BR>&nbsp;&nbsp; chunk.</STRONG></FONT>&nbsp; It SHOULD also specify the cause of abort, such as the type<BR>&nbsp;&nbsp; of the missing mandatory parameters, etc., by including the error<BR>&nbsp;&nbsp; cause parameters with the ABORT chunk.&nbsp; The Verification Tag field in<BR>&nbsp;&nbsp; the common header of the outbound SCTP packet containing the ABORT<BR>&nbsp;&nbsp; chunk MUST be set to the Initiate Tag value of the peer.</DIV>  <DIV>&nbsp;</DIV>  <DIV>Section 5.1.5 (Page# 56)</DIV>  <DIV>&nbsp;</DIV>  <DIV>&nbsp;&nbsp; 2) Authenticate the State Cookie as one that it previously generated<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; by comparing the computed MAC against the one carried in the
 State<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Cookie.&nbsp; If this comparison fails, the SCTP packet, including the<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; COOKIE ECHO and any DATA chunks, <STRONG><FONT color=#ff0000>should be silently discarded.</FONT></STRONG></DIV>  <DIV><FONT color=#000000></FONT>&nbsp;</DIV>  <DIV>&nbsp;</DIV>  <DIV>While section 2.47 of draft-ietf-tsvwg-sctpimpguide-16&nbsp;has some modifications on this:</DIV>  <DIV>&nbsp;</DIV>  <DIV>2.47.&nbsp; Sending of ABORT chunks</DIV>  <DIV>2.47.1.&nbsp; Description of the problem</DIV>  <DIV>&nbsp;&nbsp; Section 5.1 of RFC2960 [7] requires that an ABORT chunk is sent in<BR>&nbsp;&nbsp; response to an INIT chunk when there is no listening end point.&nbsp; To<BR>&nbsp;&nbsp; make port scanning harder someone might not want these ABORTs to be<BR>&nbsp;&nbsp; received by the sender of the INIT chunks.&nbsp; Currently the only way to<BR>&nbsp;&nbsp; enforce this is by using a firewall discarding the packets
 containing<BR>&nbsp;&nbsp; the INIT chunks or the packets containing the ABORT chunks.&nbsp; It is<BR>&nbsp;&nbsp; desirable that the same can be done without a middle box.</DIV>  <DIV>2.47.2.&nbsp; Text changes to the document</DIV>  <DIV>&nbsp;&nbsp; ---------<BR>&nbsp;&nbsp; Old text: (Section 5.1)<BR>&nbsp;&nbsp; ---------<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; If an endpoint receives an INIT, INIT ACK, or COOKIE ECHO chunk<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; but decides not to establish the new association due to missing<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; mandatory parameters in the received INIT or INIT ACK, invalid<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; parameter values, or lack of local resources, <STRONG><FONT color=#ff0000>it MUST respond with<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; an ABORT chunk.</FONT></STRONG></DIV>  <DIV>&nbsp;&nbsp; ---------<BR>&nbsp;&nbsp; New text: (Section 5.1)<BR>&nbsp;&nbsp; ---------</DIV>  <DIV>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; If an endpoint receives an
 INIT, INIT ACK, or COOKIE ECHO chunk<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; but decides not to establish the new association due to missing<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; mandatory parameters in the received INIT or INIT ACK, invalid<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; parameter values, or lack of local resources, <FONT color=#ff0000><STRONG>it SHOULD respond<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; with an ABORT chunk.</STRONG></FONT></DIV>  <DIV>&nbsp;</DIV>  <DIV>&nbsp;</DIV>  <DIV>But still the conflict is there.</DIV>  <DIV>&nbsp;</DIV>  <DIV>In my point of view&nbsp;there should be only&nbsp;one of these three actions, which&nbsp;should be specified when&nbsp;the SCTP&nbsp;receive the COOKIE ECHO chunk with an&nbsp;invalid cookie.</DIV>  <DIV>i) MUST silently discard (section 4 of RFC2960)</DIV>  <DIV>ii)&nbsp;SHOULD respond with an ABORT (section&nbsp;2.47 of draft-ietf-tsvwg-sctpimpguide-16)</DIV>  <DIV>iii) should be silently discarded. (section&nbsp;5.1.5 of RFC2960)</DIV> 
 <DIV>&nbsp;</DIV>  <DIV>  <DIV>Please help me resolving this conflict.</DIV>  <DIV>&nbsp;</DIV></DIV>  <DIV>-K. Äsh</DIV>  <DIV>&nbsp;</DIV>  <DIV>&nbsp;</DIV>  <DIV>---------------------------------------------------------------------------------------</DIV>  <DIV>Sooner or later, those who win are those who think they can.</DIV>  <DIV>---------------------------------------------------------------------------------------</DIV><p>
		<hr size=1>How low will we go? Check out Yahoo! Messenger’s low <a href="http://us.rd.yahoo.com/mail_us/taglines/postman8/*http://us.rd.yahoo.com/evt=39663/*http://voice.yahoo.com"> PC-to-Phone call rates.
--0-1310778812-1145597295=:98111--




From tsvwg-bounces@ietf.org Fri Apr 21 08:52:34 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FWv7Z-0000od-SP; Fri, 21 Apr 2006 08:52:25 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FWv7X-0000oY-WA
	for tsvwg@ietf.org; Fri, 21 Apr 2006 08:52:24 -0400
Received: from adsl-070-155-160-098.sip.cae.bellsouth.net ([70.155.160.98]
	helo=lakerest.net) by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FWv7X-00051h-GY
	for tsvwg@ietf.org; Fri, 21 Apr 2006 08:52:23 -0400
Received: from [IPv6:::1] (localhost [IPv6:::1])
	by lakerest.net (8.13.6/8.13.4) with ESMTP id k3LCqMUt020256;
	Fri, 21 Apr 2006 08:52:22 -0400 (EDT)
	(envelope-from randall@lakerest.net)
Message-ID: <4448D58D.3040002@lakerest.net>
Date: Fri, 21 Apr 2006 08:52:29 -0400
From: Randall Stewart <randall@lakerest.net>
User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060223
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: ash kat <ashwani_groups@yahoo.com>
Subject: Re: [Tsvwg] ambiguity in handling invalid cookie in COOKIE ECHO chunks
References: <20060421052815.1728.qmail@web37407.mail.mud.yahoo.com>
In-Reply-To: <20060421052815.1728.qmail@web37407.mail.mud.yahoo.com>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: quoted-printable
X-MIME-Autoconverted: from 8bit to quoted-printable by lakerest.net id
	k3LCqMUt020256
X-Spam-Score: 0.1 (/)
X-Scan-Signature: 8fbbaa16f9fd29df280814cb95ae2290
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

ash kat wrote:
> Hi All
>   =20
>   I was looking at the behavior of SCTP stack when the COOKIE ECHO chun=
k contains the invalid cookie.
>   =20
>   The following section of RFC2960 are conflicting each other:
>   =20
>   Section 4 (Page# 50)
>    Notes:
>      1) If the State Cookie in the received COOKIE ECHO is invalid (i.e=
.,
>       failed to pass the integrity check), the receiver MUST silently
>       discard the packet.  Or, if the received State Cookie is expired
>       (see Section 5.1.5), the receiver MUST send back an ERROR chunk.
>       In either case, the receiver stays in the CLOSED state.
>   =20
>   Section 5.1 (Page# 52)
>      If an endpoint receives an INIT, INIT ACK, or COOKIE ECHO chunk bu=
t
>    decides not to establish the new association due to missing mandator=
y
>    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 i=
n
>    the common header of the outbound SCTP packet containing the ABORT
>    chunk MUST be set to the Initiate Tag value of the peer.
>   =20
>   Section 5.1.5 (Page# 56)
>   =20
>      2) Authenticate the State Cookie as one that it previously generat=
ed
>       by comparing the computed MAC against the one carried in the Stat=
e
>       Cookie.  If this comparison fails, the SCTP packet, including the
>       COOKIE ECHO and any DATA chunks, should be silently discarded.
>   =20
>   =20
>   While section 2.47 of draft-ietf-tsvwg-sctpimpguide-16 has some modif=
ications on this:
>   =20
>   2.47.  Sending of ABORT chunks
>   2.47.1.  Description of the problem
>      Section 5.1 of RFC2960 [7] requires that an ABORT chunk is sent in
>    response to an INIT chunk when there is no listening end point.  To
>    make port scanning harder someone might not want these ABORTs to be
>    received by the sender of the INIT chunks.  Currently the only way t=
o
>    enforce this is by using a firewall discarding the packets containin=
g
>    the INIT chunks or the packets containing the ABORT chunks.  It is
>    desirable that the same can be done without a middle box.
>   2.47.2.  Text changes to the document
>      ---------
>    Old text: (Section 5.1)
>    ---------
>       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 wit=
h
>       an ABORT chunk.
>      ---------
>    New text: (Section 5.1)
>    ---------
>         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 SHOULD respond
>       with an ABORT chunk.
>   =20
>   =20
>   But still the conflict is there.
>   =20
>   In my point of view there should be only one of these three actions, =
which should be specified when the SCTP receive the COOKIE ECHO chunk wit=
h an invalid cookie.
>   i) MUST silently discard (section 4 of RFC2960)
>   ii) SHOULD respond with an ABORT (section 2.47 of draft-ietf-tsvwg-sc=
tpimpguide-16)
>   iii) should be silently discarded. (section 5.1.5 of RFC2960)
>   =20
>     Please help me resolving this conflict.
>   =20
>=20
>   -K. =C4sh
>   =20
>   =20
>   ---------------------------------------------------------------------=
------------------
>   Sooner or later, those who win are those who think they can.
>   ---------------------------------------------------------------------=
------------------
>=20
> 	=09
> ---------------------------------
> How low will we go? Check out Yahoo! Messenger=92s low  PC-to-Phone cal=
l rates.
K. Ash:

Here I disagree with you..

In the case of a COOKIE that has been modified, the integrity
check will fail... i.e. most likely you are being attacked... In this
case, the only logical thing to do is silently discard the packet.

In the case of being short on resources... or NOT having a listener then
it really helps if an ABORT is sent back.... i.e. if an innocent
guy trys to do 'sftp xxx.com' (where sftp is an SCTP based ftp)... it is
nice if an ABORT comes back instead of him hanging around waiting for
INIT time-outs... but some, for security reasons, don't like the idea
of giving clues or hints...


These are two different classes of problems... IMO.. and thus they
are not in conflict.. they would be in conflict if you can site
two places that have the SAME thing that tell you to do different
things... i.e. if the two places that mention a corrupted cookie, not
passing the security check, told you to in one place abort and in
the other silently discard....

The document does not do that...

R

--
Randall Stewart
803-345-0369 <or> 815-342-5222(cell)




From tsvwg-bounces@ietf.org Fri Apr 21 10:48:46 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FWwvj-0002rd-U9; Fri, 21 Apr 2006 10:48:19 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FWwvi-0002rY-S4
	for tsvwg@ietf.org; Fri, 21 Apr 2006 10:48:18 -0400
Received: from stsc1260-eth-s1-s1p1-vip.va.neustar.com ([156.154.16.129]
	helo=chiedprmail1.ietf.org)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FWvQQ-0006I3-5M
	for tsvwg@ietf.org; Fri, 21 Apr 2006 09:11:54 -0400
Received: from dee.erg.abdn.ac.uk ([139.133.204.82] helo=erg.abdn.ac.uk)
	by chiedprmail1.ietf.org with esmtp (Exim 4.43) id 1FWvGv-0007q7-O6
	for tsvwg@ietf.org; Fri, 21 Apr 2006 09:02:06 -0400
Received: from [139.133.207.151] (dhcp-207-151.erg.abdn.ac.uk
	[139.133.207.151])
	by erg.abdn.ac.uk (8.13.4/8.13.4) with ESMTP id k3LD1dtC006658;
	Fri, 21 Apr 2006 14:01:39 +0100 (BST)
Message-ID: <4448D7B3.4020805@erg.abdn.ac.uk>
Date: Fri, 21 Apr 2006 14:01:39 +0100
From: Gorry Fairhurst <gorry@erg.abdn.ac.uk>
Organization: University of Aberdeen, UK
User-Agent: Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: tsvwg@ietf.org
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-ERG-MailScanner: Found to be clean
X-ERG-MailScanner-From: gorry@erg.abdn.ac.uk
X-Spam-Status: No
X-Spam-Score: -2.6 (--)
X-Scan-Signature: f49c97ce49302a02285a2d36a99eef8c
Cc: Gorry Fairhurst <gorry@erg.abdn.ac.uk>, pasi.sarolahti@iki.fi,
	floyd@icir.org, mallman@icir.org
Subject: [Tsvwg] Comments on draft-ietf-tsvwg-quickstart-02.txt
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: gorry@erg.abdn.ac.uk
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org


Sally, authors, list. I started a detailed read of the latest rev of the 
QS draft this week, and have a few comments. Alas I ran short of time 
and shall be away on another job next week - so rather than postpone 
this, I am sending now, with the intention of the sending the final 
comments after next week.

Best wishes,

Gorry

1) Quick Start Interval
- This isn't a new point, but I think there should be a maximum time 
limit for a QS approval. I have the notion that a QS rate should be 
soft-state (or no-state) in the router. If it is soft-state (implies 
some accounting of requests for a period of time), I'd much rather see 
the time bounded.

This has some implications as well on the relationship betwen RTT and 
sending rate: I can't see a problem with saying the state MUST expire 
after (say) 3 seconds. This is a LONG time compared to most RTT's and 
RTTs that are actually greater than this can perhaps rate-pace their use 
of the window over the 3 second interval.

Fixing an upper limit could also help to get "predictable" performance 
in challenging envirobments - e.g. when BoD (e.g. an ISDN 
circuit/Wireless infrastructure) requires a time to setup a flow - but 
that additional setup time does not appear as a increase in subsequent 
RTT measurements.

2) Page 11: On relatively stable paths
"shows the assumption should generally be valid, although [RFC3819] 
discusses a range of Bandwidth on Demand subnets."
- This is not clear, I don't think the issue is necessarily the 
stability of the PATH (in terms of the set of routers involved).
- To me, the point concerning BoD subnets is that the capacity they 
offer (and RTT they exhibit) can be difficult to characterise at the 
transport layer, and may be a function of the overall traffic carried by 
the subnet.

3) Section 3.1
- The document seems not to mention Router-Alert. This is probably an 
oversight, in that at least this should be mentioned - if QS does not 
require Router-Alert, then we should say so.

4) Page 19 Section 3.3 para 2 - Non-approval.
"A router that wishes to terminate a Quick-Start Request SHOULD delete 
the Quick-Start Request from the IP Header... but.. then the router 
SHOULD zero the QS TTL, QS Nonce, and Rate Request Fields."
- I don't like this. The first SHOULD is ok. The second one leaves too 
much undefined. Can we consider mandating at least one of these actions 
if the request is not approved and the option has not been removed. I 
suggest:
"then the router MUST zero the Rate Request Fields and SHOULD zero the 
QS TTL, and QS Nonce."

5) Page 21, section 3.4 final para.
"then the router may decide to deny future Quick-Start Requests"
                  ^^^
- may or MAY?

6) Page 27, section 4.2, final para.
"Packet loss is an indication of congestion"
              ^^^^^^^^^^^^^^^^
- I'd argue that it is an indication of potential congestion. TCP needs 
to be conservative, it's not stupid, it makes this decision because 
there may be a problem and it can not know otherwise.

7) Section 4.6.2 I don't necessarily think of a simple ACL rule being a 
midbox function.
- Clearly the issues with more sophisticated midboxes are explained. 
However, this speaks of midboxes in a generalised way, and that's fine - 
but I would not generally have relate this discussion with router ACLS - 
which in my net is actually the place were decisions on options are 
made. I wonder if the wording could be clearer here.

8) Section 6 talks abourt IP-Tunnels.

- The opening text however only mentions IPsec and IP-in-IP. As far as I 
can see this text also relates to GRE and other tunnel encapsulations. 
Since we are frequenctly requested to consider GRE by the IESG, can we 
have more  comprehensive first sentence please?


Issues for my next mail:

Appendices (still to be read)
Approval - various issues in the motivation, security and coding, 
although I think I like the principle.

--------------------------------------------------------------

Editorial NiTs:

---
At various places single and double quotes do not use the RFC-Ed characters.

---
Page 2:
"is designed to allow connections to use higher sending rates"
                       ^                  ^                  ^
I'd have prefered:
"is designed to allow a connection to use a higher sending rate"
---
Page 13: Section 2.1
"The transport receiver returns the rate and information about the TTL"
- And the NONCE also :-)
---
Page 18, section 3.2:
"Hop Limit field in order to calculate"
                  ^^^^^^^^
Say simply
"Hop Limit field to calculate"
- The use of "in order" confused me here, since there is no implied 
ordering. I suggest you avoid the use of this phrase except when it 
implies that something else follows?

---
Page 19, Section 3.3:
"This saves resources as downstream routers"
                       ^^
Better to say:
"This saves resources because downstream routers"
---
Page 24, Section 4:
"The sender sends a Report of Approved Rate."
                              ^
- Insert "the" after "of"
---
Pge 29, Section 4.5
Quotes in first para seem odd (and may be unecessary?)
---
Page 32 top para:
"Is more likely to be blocked due to IP Options than due.."
- Poor wording, it's more likely to be dropped because of filtering 
policies that discard packets that carry IP Options (or packets with 
unrecognised IP Options).
---






From tsvwg-bounces@ietf.org Fri Apr 21 11:11:13 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FWxHk-0000Cx-Tp; Fri, 21 Apr 2006 11:11:04 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FWxHk-0000Cs-En
	for tsvwg@ietf.org; Fri, 21 Apr 2006 11:11:04 -0400
Received: from wyvern.icir.org ([192.150.187.14])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FWxHj-0005U0-4B
	for tsvwg@ietf.org; Fri, 21 Apr 2006 11:11:04 -0400
Received: from guns.icir.org (adsl-69-222-35-58.dsl.bcvloh.ameritech.net
	[69.222.35.58])
	by wyvern.icir.org (8.12.11/8.12.11) with ESMTP id k3LFAw8p099446;
	Fri, 21 Apr 2006 08:10:58 -0700 (PDT)
	(envelope-from mallman@guns.icir.org)
Received: from guns.icir.org (localhost [127.0.0.1])
	by guns.icir.org (Postfix) with ESMTP
	id 5F3A577AB77; Fri, 21 Apr 2006 11:10:57 -0400 (EDT)
To: gorry@erg.abdn.ac.uk
From: Mark Allman <mallman@icir.org>
In-Reply-To: <4448D7B3.4020805@erg.abdn.ac.uk> 
Organization: ICSI Center for Internet Research (ICIR)
Song-of-the-Day: The Rising
MIME-Version: 1.0
Content-Type: multipart/signed; boundary="=_bOundary";
	micalg=pgp-sha1; protocol="application/pgp-signature"
Date: Fri, 21 Apr 2006 11:10:57 -0400
Message-Id: <20060421151057.5F3A577AB77@guns.icir.org>
X-Spam-Score: 0.1 (/)
X-Scan-Signature: ea4ac80f790299f943f0a53be7e1a21a
Cc: pasi.sarolahti@iki.fi, tsvwg@ietf.org, floyd@icir.org
Subject: [Tsvwg] Re: Comments on draft-ietf-tsvwg-quickstart-02.txt 
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: mallman@icir.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

--=_bOundary
Content-Type: text/plain
Content-Disposition: inline


Gorry-

Thanks for the note.  I happened to be hacking the draft when your note
arrived.  I fixed the nits you outlined, but do not have time to process
your larger comments right now.  However, they all look to be useful and
we will get them addressed.

Thanks!

allman




--=_bOundary
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (FreeBSD)

iD8DBQFESPYBWyrrWs4yIs4RAinJAJ9uYGP11GXhJggmxX3QNgT5HeW2lQCglnn/
JL7QxIkaQ466xA0HYlHM0f8=
=X51a
-----END PGP SIGNATURE-----
--=_bOundary--




From tsvwg-bounces@ietf.org Fri Apr 21 14:20:57 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FX0F4-0001Zp-Qs; Fri, 21 Apr 2006 14:20:30 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FX0F3-0001Zk-HY
	for tsvwg@ietf.org; Fri, 21 Apr 2006 14:20:29 -0400
Received: from cougar.icir.org ([192.150.187.76])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FX0F1-0007w6-0X
	for tsvwg@ietf.org; Fri, 21 Apr 2006 14:20:29 -0400
Received: from cougar.icir.org (localhost [127.0.0.1])
	by cougar.icir.org (8.12.11/8.12.11) with ESMTP id k3LIKIFB060065;
	Fri, 21 Apr 2006 11:20:18 -0700 (PDT)
	(envelope-from floyd@cougar.icir.org)
Message-Id: <200604211820.k3LIKIFB060065@cougar.icir.org>
To: gorry@erg.abdn.ac.uk
From: Sally Floyd <floyd@icir.org>
Date: Fri, 21 Apr 2006 11:20:18 -0700
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 200d029292fbb60d25b263122ced50fc
Cc: pasi.sarolahti@iki.fi, tsvwg@ietf.org, mallman@icir.org
Subject: [Tsvwg] Re: Comments on draft-ietf-tsvwg-quickstart-02.txt 
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Gorry -

Many thanks for the feedback.  We have been editing the draft,
and hope to have a version ready for Working Group Last Call
next week.

>1) Quick Start Interval
>- This isn't a new point, but I think there should be a maximum time 
>limit for a QS approval. I have the notion that a QS rate should be 
>soft-state (or no-state) in the router. If it is soft-state (implies 
>some accounting of requests for a period of time), I'd much rather see 
>the time bounded.
>
>This has some implications as well on the relationship betwen RTT and 
>sending rate: I can't see a problem with saying the state MUST expire 
>after (say) 3 seconds. This is a LONG time compared to most RTT's and 
>RTTs that are actually greater than this can perhaps rate-pace their use 
>of the window over the 3 second interval.
>
>Fixing an upper limit could also help to get "predictable" performance 
>in challenging envirobments - e.g. when BoD (e.g. an ISDN 
>circuit/Wireless infrastructure) requires a time to setup a flow - but 
>that additional setup time does not appear as a increase in subsequent 
>RTT measurements.

The revised version says this:

    If the checks of the TTL Diff and the Rate Request are successful,
    and the TCP host is going to use the Quick-Start Request, it MUST
    start using it within one round-trip time of receiving the Quick-
    Start Response. 

We could revise this to say within one RTT or within three seconds,
whichever is smaller.

>2) Page 11: On relatively stable paths
>"shows the assumption should generally be valid, although [RFC3819] 
>discusses a range of Bandwidth on Demand subnets."
>- This is not clear, I don't think the issue is necessarily the 
>stability of the PATH (in terms of the set of routers involved).
>- To me, the point concerning BoD subnets is that the capacity they 
>offer (and RTT they exhibit) can be difficult to characterise at the 
>transport layer, and may be a function of the overall traffic carried by 
>the subnet.

Yep.  Though the assumption about relatively stable paths matters because
we want the Quick-Start data packets to follow the same path as
the Quick-Start request the previous RTT.  The revised version
of the draft says this, and hopefully is more clear:

    Assumptions:
...
    * The path between the two endpoints is relatively stable, such that
    the path used by the Quick-Start request is generally the same path
    used by the Quick-Start packets one round-trip time later.  [ZDPS01]
    shows this assumption should be generally valid.  However, [RFC3819]
    discusses a range of Bandwidth on Demand subnets that could cause
    the characteristics of the path to change over time.

>3) Section 3.1
>- The document seems not to mention Router-Alert. This is probably an 
>oversight, in that at least this should be mentioned - if QS does not 
>require Router-Alert, then we should say so.

Section A.1.1 on ICMP says the following:

    The ICMP Quick-Start Request would  
    have to pass by the routers on the path to the receiver, possibly   
    using the IP Router Alert option [RFC2113].

My assumption is that QS with the IP opion does not require
Router Alert.  I will add that.

>4) Page 19 Section 3.3 para 2 - Non-approval.
>"A router that wishes to terminate a Quick-Start Request SHOULD delete 
>the Quick-Start Request from the IP Header... but.. then the router 
>SHOULD zero the QS TTL, QS Nonce, and Rate Request Fields."
>- I don't like this. The first SHOULD is ok. The second one leaves too 
>much undefined. Can we consider mandating at least one of these actions 
>if the request is not approved and the option has not been removed. I 
>suggest:
>"then the router MUST zero the Rate Request Fields and SHOULD zero the 
>QS TTL, and QS Nonce."

Well, there already exist non-Quick-Start-capable routers that
won't be able to zero anything in the QS request, but will
implicitly deny the request by not decrementing the QS TTL counter.

And we already say that the router should protect itself from
QS-request-floods by rate-limiting the number of QS requests that
it considers (and implicitly denying the others by not decrementing
the QS TTL counter.  I think it is necessary to leave that in.
And that requires allowing Quick-Start-capable routers to deny
requests just by ignoring them.

Yes?

>5) Page 21, section 3.4 final para.
>"then the router may decide to deny future Quick-Start Requests"
>                  ^^^
>- may or MAY?

I will change it to MAY.

>6) Page 27, section 4.2, final para.
>"Packet loss is an indication of congestion"
>              ^^^^^^^^^^^^^^^^
>- I'd argue that it is an indication of potential congestion. TCP needs 
>to be conservative, it's not stupid, it makes this decision because 
>there may be a problem and it can not know otherwise.

Will do.  (Packet loss could also be an indication of corruption on
the return path.)

>7) Section 4.6.2 I don't necessarily think of a simple ACL rule being a 
>midbox function.
>- Clearly the issues with more sophisticated midboxes are explained. 
>However, this speaks of midboxes in a generalised way, and that's fine - 
>but I would not generally have relate this discussion with router ACLS - 
>which in my net is actually the place were decisions on options are 
>made. I wonder if the wording could be clearer here.

I am not sure what you want.  You want the text to make it more
explicit that the entity dropping the packet with the QS request could
be a router?  That would be fine by me.

>8) Section 6 talks abourt IP-Tunnels.
>
>- The opening text however only mentions IPsec and IP-in-IP. As far as I 
>can see this text also relates to GRE and other tunnel encapsulations. 
>Since we are frequenctly requested to consider GRE by the IESG, can we 
>have more  comprehensive first sentence please?

Will do.  We also have added a section about MPLS...  

>Issues for my next mail:
>
>Appendices (still to be read)
>Approval - various issues in the motivation, security and coding, 
>although I think I like the principle.

I will send you the revised version, even though it is not
yet submitted, so that you can read the most recent version.

Many thanks.
- Sally
http://www.icir.org/floyd/




From tsvwg-bounces@ietf.org Fri Apr 21 14:58:21 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FX0pT-0003hi-L4; Fri, 21 Apr 2006 14:58:07 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FX0pR-0003gb-Vq
	for tsvwg@ietf.org; Fri, 21 Apr 2006 14:58:05 -0400
Received: from cougar.icir.org ([192.150.187.76])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FX0pQ-0000bo-K1
	for tsvwg@ietf.org; Fri, 21 Apr 2006 14:58:05 -0400
Received: from cougar.icir.org (localhost [127.0.0.1])
	by cougar.icir.org (8.12.11/8.12.11) with ESMTP id k3LIw10g060809;
	Fri, 21 Apr 2006 11:58:01 -0700 (PDT)
	(envelope-from floyd@cougar.icir.org)
Message-Id: <200604211858.k3LIw10g060809@cougar.icir.org>
To: gorry@erg.abdn.ac.uk
From: Sally Floyd <floyd@icir.org>
Date: Fri, 21 Apr 2006 11:58:01 -0700
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 68c8cc8a64a9d0402e43b8eee9fc4199
Cc: pasi.sarolahti@iki.fi, tsvwg@ietf.org, mallman@icir.org
Subject: [Tsvwg] Re: Comments on draft-ietf-tsvwg-quickstart-02.txt 
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Gorry -

I have put the revised version of the QS draft at

  http://www.icir.org/floyd/papers/draft-ietf-tsvwg-quickstart-03a.txt

if you want to read from that one.
We should be submitting it in a few days, ready for 
Working Group Last Call.

Many thanks.
- Sally




From tsvwg-bounces@ietf.org Fri Apr 21 18:17:59 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FX3wf-0001e1-Bz; Fri, 21 Apr 2006 18:17:45 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FX3we-0001dw-IN
	for tsvwg@ietf.org; Fri, 21 Apr 2006 18:17:44 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FX3wd-0003SV-9X
	for tsvwg@ietf.org; Fri, 21 Apr 2006 18:17:44 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FX3wY-0002yP-Uj; Fri, 21 Apr 2006 16:17:17 -0600
Message-ID: <44495A0C.50707@middle.net>
Date: Fri, 21 Apr 2006 16:17:48 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: tsvwg@ietf.org
Subject: Re: [Tsvwg] ambiguity in handling invalid cookie in COOKIE ECHO chunks
References: <20060421052815.1728.qmail@web37407.mail.mud.yahoo.com>
	<4448D58D.3040002@lakerest.net>
In-Reply-To: <4448D58D.3040002@lakerest.net>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 08170828343bcf1325e4a0fb4584481c
Cc: Randall Stewart <randall@lakerest.net>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

I agree that if the COOKIE ECHO is corrupt (as opposed to stale) it 
should be silently discarded.

The other conditions where one might ABORT could use some clarification. 
Notably in the case of lack of resources whether you should ABORT or 
discard depends on how long you expect the problem to last.  If 
indefinitely, then ABORT is the correct response.  However, if the 
problem is temporary, silently discarding the packet is much to be 
preferred. 

In the case of a full accept queue, the latter behavior should be 
standardized in the SCTP socket draft, so that connect(2) calls do not 
error out pre-maturely.

 - Mark B.




From tsvwg-bounces@ietf.org Fri Apr 21 20:36:07 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FX66I-0002ln-4U; Fri, 21 Apr 2006 20:35:50 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FX66H-0002li-JD
	for tsvwg@ietf.org; Fri, 21 Apr 2006 20:35:49 -0400
Received: from cougar.icir.org ([192.150.187.76])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FX66F-00030d-Jw
	for tsvwg@ietf.org; Fri, 21 Apr 2006 20:35:49 -0400
Received: from cougar.icir.org (localhost [127.0.0.1])
	by cougar.icir.org (8.12.11/8.12.11) with ESMTP id k3M0ZjLa065560;
	Fri, 21 Apr 2006 17:35:45 -0700 (PDT)
	(envelope-from floyd@cougar.icir.org)
Message-Id: <200604220035.k3M0ZjLa065560@cougar.icir.org>
To: tsvwg@ietf.org
From: Sally Floyd <floyd@icir.org>
Date: Fri, 21 Apr 2006 17:35:45 -0700
X-Spam-Score: 0.0 (/)
X-Scan-Signature: cd26b070c2577ac175cd3a6d878c6248
Cc: Pasi Sarolahti <pasi.sarolahti@iki.fi>, Mark Allman <mallman@icir.org>
Subject: [Tsvwg] Working Group Last Call for
	draft-ietf-tsvwg-quickstart-03.txt?
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

We have just submitted a revised version of
draft-ietf-tsvwg-quickstart-03.txt.  We believe that this is now
ready for Working Group Last Call (finally!).

http://www.icir.org/floyd/papers/draft-ietf-tsvwg-quickstart-03.txt
http://www.icir.org/floyd/papers/draft-ietf-tsvwg-quickstart-03.ps

- Sally

The changes, listed in the draft, are as following:

     Changes from draft-ietf-tsvwg-quickstart-02:

     * Some general editing.

     * Said that if the receiver receives a Quick-Start Request
       with a rate of zero, then the receiver SHOULD NOT send
       a Quick-Start response.  And that if the sender
       receives an acknowledgement of its packet with no
       Quick-Start response, then the sender MUST assume that the
       request was denied, and send a Report of Approved Rate
       with a rate of zero.

     * Said that if a Quick-Start packet is dropped or marked,
       the sender should not make more Quick-Start requests in this
       connection.

     * Said that the Quick-Start Request SHOULD be sent on a packet
       that requires an acknowledgement, e.g., a SYN, SYN/ACK, or data
       packet.

     * Made changes to the section on "TCP: A Quick-Start Request in the
       Middle of a Connection".

     * Added that if the TCP host is going to use the successful
       Quick-Start Request, it MUST start using it within one
       round-trip time of receiving the Quick-Start Response,
       or within three seconds, whichever is smaller.

     * Added a stronger applicability statement, in the abstract
       and in Section 10 on "Implementation and Deployment Issues".
       From feedback from the working group.

     * Added a section about MPLS.  From feedback from Mitchell
       Erblichs.

     * Strengthened the language of the difficulties presented by
       multi-access links.

     * Added a discussion in Section 10.3 about the deployment of
       Quick-Start on single-hop paths.  From feedback from
       Mitchell Erblichs.

     * Clarified that the "router" function of approving
       Quick-Start requests includes the IP-layer processing
       at the sender.

     * Clarified in Section 3.3 on "Processing the Quick-Start
       Request at Routers" that this document standardizes only
       the semantics of Quick-Start, and not the specific
       algorithms for processing Quick-Start requests at routers.

     * Clarified in Section 3.3 on "Processing the Quick-Start
       Request at Routers" that a router will have to consider
       the previous Quick-Start requests in approving a new one.

     * In Section 9.3 on "Quick-Start with QoS-enabled Traffic",
       which says that routers are free to take into account
       the diff-serv codepoint in considering QS requests, clarified
       that routers are also free to take into account their own
       understanding of priorities.

     * Added the Temporary bit to Appendix D on "Possible Additional
       Uses for the Quick-Start Option".  Clarified that the Quick-Start
       mechanism is not designed to help routers achieve full link
       utilization.

     * Editing from feedback from Gorry Fairhurst.





From tsvwg-bounces@ietf.org Sat Apr 22 04:29:11 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FXDTT-0003Ez-9l; Sat, 22 Apr 2006 04:28:15 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FXDTR-0003Et-JS
	for tsvwg@ietf.org; Sat, 22 Apr 2006 04:28:13 -0400
Received: from dee.erg.abdn.ac.uk ([139.133.204.82] helo=erg.abdn.ac.uk)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FXDTQ-0001N7-1q
	for tsvwg@ietf.org; Sat, 22 Apr 2006 04:28:13 -0400
Received: from [10.0.1.13] (maxp7.dialup.abdn.ac.uk [139.133.201.166])
	(authenticated bits=0)
	by erg.abdn.ac.uk (8.13.4/8.13.4) with ESMTP id k3M8Rq5M000782
	(version=TLSv1/SSLv3 cipher=DES-CBC3-SHA bits=168 verify=NOT);
	Sat, 22 Apr 2006 09:27:55 +0100 (BST)
User-Agent: Microsoft-Entourage/11.2.1.051004
Date: Sat, 22 Apr 2006 09:29:42 +0100
Subject: Re: [Tsvwg] Re: Comments on draft-ietf-tsvwg-quickstart-02.txt
From: Gorry Fairhurst <gorry@erg.abdn.ac.uk>
To: Sally Floyd <floyd@icir.org>
Message-ID: <C06FA806.4D00%gorry@erg.abdn.ac.uk>
Thread-Topic: [Tsvwg] Re: Comments on draft-ietf-tsvwg-quickstart-02.txt
Thread-Index: AcZl5ucOJVplQtHaEdqI3QAKlc/qXg==
In-Reply-To: <200604211820.k3LIKIFB060065@cougar.icir.org>
Mime-version: 1.0
Content-type: text/plain;
	charset="US-ASCII"
Content-transfer-encoding: 7bit
X-ERG-MailScanner: Found to be clean
X-ERG-MailScanner-From: gorry@erg.abdn.ac.uk
X-Spam-Status: No
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 48472a944c87678fcfe8db15ffecdfff
Cc: pasi.sarolahti@iki.fi, tsvwg@ietf.org, mallman@icir.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

On 21/4/06 19:20, "Sally Floyd" <floyd@icir.org> wrote:

> Gorry -
> 
> Many thanks for the feedback.  We have been editing the draft,
> and hope to have a version ready for Working Group Last Call
> next week.
> 
Look forward to this - and will postpone my final set of comments until this
rev is issued...

>> 1) Quick Start Interval
>> - This isn't a new point, but I think there should be a maximum time
>> limit for a QS approval. I have the notion that a QS rate should be
>> soft-state (or no-state) in the router. If it is soft-state (implies
>> some accounting of requests for a period of time), I'd much rather see
>> the time bounded.
>> 
>> This has some implications as well on the relationship betwen RTT and
>> sending rate: I can't see a problem with saying the state MUST expire
>> after (say) 3 seconds. This is a LONG time compared to most RTT's and
>> RTTs that are actually greater than this can perhaps rate-pace their use
>> of the window over the 3 second interval.
>> 
>> Fixing an upper limit could also help to get "predictable" performance
>> in challenging envirobments - e.g. when BoD (e.g. an ISDN
>> circuit/Wireless infrastructure) requires a time to setup a flow - but
>> that additional setup time does not appear as a increase in subsequent
>> RTT measurements.
> 
> The revised version says this:
> 
>     If the checks of the TTL Diff and the Rate Request are successful,
>     and the TCP host is going to use the Quick-Start Request, it MUST
>     start using it within one round-trip time of receiving the Quick-
>     Start Response.
> 
> We could revise this to say within one RTT or within three seconds,
> whichever is smaller.
> 
The rule is fine from the perspective of a host that knows the PRTT (e.g.
from TCP). Routers on the other hand know nothing of the PRTT of individual
flows, and therefore do not know how to behave. If they do take different
stances, then the hosts are less able to know what will happen.

What do others think about a recommended upper limit to the time (3 secs)? -
Or put another way: Can we agree a reasonable limit?

>> 2) Page 11: On relatively stable paths
>> "shows the assumption should generally be valid, although [RFC3819]
>> discusses a range of Bandwidth on Demand subnets."
>> - This is not clear, I don't think the issue is necessarily the
>> stability of the PATH (in terms of the set of routers involved).
>> - To me, the point concerning BoD subnets is that the capacity they
>> offer (and RTT they exhibit) can be difficult to characterise at the
>> transport layer, and may be a function of the overall traffic carried by
>> the subnet.
> 
> Yep.  Though the assumption about relatively stable paths matters because
> we want the Quick-Start data packets to follow the same path as
> the Quick-Start request the previous RTT.  The revised version
> of the draft says this, and hopefully is more clear:
> 
>     Assumptions:
> ...
>     * The path between the two endpoints is relatively stable, such that
>     the path used by the Quick-Start request is generally the same path
>     used by the Quick-Start packets one round-trip time later.  [ZDPS01]
>     shows this assumption should be generally valid.  However, [RFC3819]
>     discusses a range of Bandwidth on Demand subnets that could cause
>     the characteristics of the path to change over time.
> 
Fine with me.

>> 3) Section 3.1
>> - The document seems not to mention Router-Alert. This is probably an
>> oversight, in that at least this should be mentioned - if QS does not
>> require Router-Alert, then we should say so.
> 
> Section A.1.1 on ICMP says the following:
> 
>     The ICMP Quick-Start Request would
>     have to pass by the routers on the path to the receiver, possibly
>     using the IP Router Alert option [RFC2113].
> 
> My assumption is that QS with the IP opion does not require
> Router Alert.  I will add that.
OK - I thought that was meant.
> 
>> 4) Page 19 Section 3.3 para 2 - Non-approval.
>> "A router that wishes to terminate a Quick-Start Request SHOULD delete
>> the Quick-Start Request from the IP Header... but.. then the router
>> SHOULD zero the QS TTL, QS Nonce, and Rate Request Fields."
>> - I don't like this. The first SHOULD is ok. The second one leaves too
>> much undefined. Can we consider mandating at least one of these actions
>> if the request is not approved and the option has not been removed. I
>> suggest:
>> "then the router MUST zero the Rate Request Fields and SHOULD zero the
>> QS TTL, and QS Nonce."
> 
> Well, there already exist non-Quick-Start-capable routers that
> won't be able to zero anything in the QS request, but will
> implicitly deny the request by not decrementing the QS TTL counter.
True.

> And we already say that the router should protect itself from
> QS-request-floods by rate-limiting the number of QS requests that
> it considers (and implicitly denying the others by not decrementing
> the QS TTL counter.  I think it is necessary to leave that in.
> And that requires allowing Quick-Start-capable routers to deny
> requests just by ignoring them.
> 
> Yes?
OK - I see that logic.
> 
>> 5) Page 21, section 3.4 final para.
>> "then the router may decide to deny future Quick-Start Requests"
>>                  ^^^
>> - may or MAY?
> 
> I will change it to MAY.
OK.
> 
>> 6) Page 27, section 4.2, final para.
>> "Packet loss is an indication of congestion"
>>              ^^^^^^^^^^^^^^^^
>> - I'd argue that it is an indication of potential congestion. TCP needs
>> to be conservative, it's not stupid, it makes this decision because
>> there may be a problem and it can not know otherwise.
> 
> Will do.  (Packet loss could also be an indication of corruption on
> the return path.)
Indeed!
> 
>> 7) Section 4.6.2 I don't necessarily think of a simple ACL rule being a
>> midbox function.
>> - Clearly the issues with more sophisticated midboxes are explained.
>> However, this speaks of midboxes in a generalised way, and that's fine -
>> but I would not generally have relate this discussion with router ACLS -
>> which in my net is actually the place were decisions on options are
>> made. I wonder if the wording could be clearer here.
> 
> I am not sure what you want.  You want the text to make it more
> explicit that the entity dropping the packet with the QS request could
> be a router?  That would be fine by me.
> 
One sentence that indicated that the filtering could be a simply policy
decision in a router would be nice at the beginning. In other documents, I
sometimes skip "modbox" considerations if I'm just thinking of routers and
end-hosts, in this case EVERYONE needs to read this section.

>> 8) Section 6 talks abourt IP-Tunnels.
>> 
>> - The opening text however only mentions IPsec and IP-in-IP. As far as I
>> can see this text also relates to GRE and other tunnel encapsulations.
>> Since we are frequenctly requested to consider GRE by the IESG, can we
>> have more  comprehensive first sentence please?
> 
> Will do.  We also have added a section about MPLS...
Good.
>
>> Issues for my next mail:
>> 
>> Appendices (still to be read)
>> Approval - various issues in the motivation, security and coding,
>> although I think I like the principle.
> 
> I will send you the revised version, even though it is not
> yet submitted, so that you can read the most recent version.
> 
> Many thanks.
> - Sally
> http://www.icir.org/floyd/
> 

Best wishes,

Gorry







From tsvwg-bounces@ietf.org Sat Apr 22 10:07:18 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FXIlW-0001pa-QR; Sat, 22 Apr 2006 10:07:14 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FXIlV-0001pV-FJ
	for tsvwg@ietf.org; Sat, 22 Apr 2006 10:07:13 -0400
Received: from adsl-070-155-160-098.sip.cae.bellsouth.net ([70.155.160.98]
	helo=lakerest.net) by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FXIlU-0002Zx-5Z
	for tsvwg@ietf.org; Sat, 22 Apr 2006 10:07:13 -0400
Received: from [IPv6:::1] (localhost [IPv6:::1])
	by lakerest.net (8.13.6/8.13.4) with ESMTP id k3ME7BhF027588;
	Sat, 22 Apr 2006 10:07:11 -0400 (EDT)
	(envelope-from randall@lakerest.net)
Message-ID: <444A388E.6060704@lakerest.net>
Date: Sat, 22 Apr 2006 10:07:10 -0400
From: Randall Stewart <randall@lakerest.net>
User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060223
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Mark Butler <butlerm@middle.net>
Subject: Re: [Tsvwg] ambiguity in handling invalid cookie in COOKIE ECHO chunks
References: <20060421052815.1728.qmail@web37407.mail.mud.yahoo.com>	<4448D58D.3040002@lakerest.net>
	<44495A0C.50707@middle.net>
In-Reply-To: <44495A0C.50707@middle.net>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.1 (/)
X-Scan-Signature: 69a74e02bbee44ab4f8eafdbcedd94a1
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark Butler wrote:
> I agree that if the COOKIE ECHO is corrupt (as opposed to stale) it 
> should be silently discarded.
> 
> The other conditions where one might ABORT could use some clarification. 
> Notably in the case of lack of resources whether you should ABORT or 
> discard depends on how long you expect the problem to last.  If 
> indefinitely, then ABORT is the correct response.  However, if the 
> problem is temporary, silently discarding the packet is much to be 
> preferred.
> In the case of a full accept queue, the latter behavior should be 
> standardized in the SCTP socket draft, so that connect(2) calls do not 
> error out pre-maturely.

I agree that in a full-q one-2-one model case you probably want
to silently discard the COOKIE and let the retransmission of the
cookie hopefully help you get the association established... for
BSD this is the default behavior.. but there IS a sysctl to turn
this into a abort response too... since there will be some
that want this..

And as to if you hit the system limit of associations (another
type of resource constraint) then you will get an abort ... since
that is unlikely to clear before your retransmission timer goes
off..

R

> 
> - Mark B.
> 
> 
> 


-- 
Randall Stewart
803-345-0369 <or> 815-342-5222(cell)




From tsvwg-bounces@ietf.org Sat Apr 22 12:19:26 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FXKp9-0006gy-Dp; Sat, 22 Apr 2006 12:19:07 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FXKp8-0006gt-Dp
	for tsvwg@ietf.org; Sat, 22 Apr 2006 12:19:06 -0400
Received: from ams-iport-1.cisco.com ([144.254.224.140])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FXKp8-0000Qr-2M
	for tsvwg@ietf.org; Sat, 22 Apr 2006 12:19:06 -0400
Received: from ams-core-1.cisco.com ([144.254.224.150])
	by ams-iport-1.cisco.com with ESMTP; 22 Apr 2006 18:19:07 +0200
Received: from xbh-ams-331.emea.cisco.com (xbh-ams-331.cisco.com
	[144.254.231.71])
	by ams-core-1.cisco.com (8.12.10/8.12.6) with ESMTP id k3MGJ5UE022265; 
	Sat, 22 Apr 2006 18:19:05 +0200 (MEST)
Received: from xmb-ams-333.cisco.com ([144.254.231.78]) by
	xbh-ams-331.emea.cisco.com with Microsoft SMTPSVC(6.0.3790.0); 
	Sat, 22 Apr 2006 18:19:05 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="US-ASCII"
Content-Transfer-Encoding: quoted-printable
Subject: RE: [Tsvwg] WGLC for draft-ietf-tsvwg-rsvp-dste-01.txt
Date: Sat, 22 Apr 2006 18:19:00 +0200
Message-ID: <A05118C6DF9320488C77F3D5459B17B7013B26A2@xmb-ams-333.emea.cisco.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: [Tsvwg] WGLC for draft-ietf-tsvwg-rsvp-dste-01.txt
Thread-Index: AcZCU0DibzBbgA0TSpSCsquowztrbQcqGdTAAcsdyyA=
From: "Francois Le Faucheur \(flefauch\)" <flefauch@cisco.com>
To: "Francois Le Faucheur \(flefauch\)" <flefauch@cisco.com>,
	"tsvwg" <tsvwg@ietf.org>,
	"magnus Westerlund" <magnus.westerlund@ericsson.com>,
	"James Polk \(jmpolk\)" <jmpolk@cisco.com>
X-OriginalArrivalTime: 22 Apr 2006 16:19:05.0228 (UTC)
	FILETIME=[79A5ECC0:01C66628]
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 2086112c730e13d5955355df27e3074b
Cc: 
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Hello again,=20

>> -----Original Message-----
>> From: Francois Le Faucheur (flefauch)=20
>> Sent: jeudi 13 avril 2006 15:30
>> To: tsvwg
>> Cc: magnus Westerlund; James Polk (jmpolk)
>> Subject: RE: [Tsvwg] WGLC for draft-ietf-tsvwg-rsvp-dste-01.txt
>>=20
>> >> TSVWG
>> >>=20
>> >> The chairs have concluded the WGLC for the document:
>> >>=20
>> >>          draft-ietf-tsvwg-rsvp-dste-01.txt
>> >>=20
>> >> which started November 8th, has closed.
>> >>=20
>> >> Allison
>> >> James
>> >> Jon
>> >> 	IETF TSVWG co-chairs
>>=20
>> Hello,
>>=20
>> I received a couple of good comments privately:
>> 	- the draft discusses forwarding and Router Alert handling for
>> E2E Path messages, but it does not explicitely discusses the=20
>> other E2E
>> RSVP messages which also use Router Alert (i.e. PathTear and=20
>> ResvConf).
>> While the handling of those is similar, it is worth being=20
>> explicit about
>> that.
>> 	- When discussing the forwarding/encapsulation of E2E Path
>> message (in section 3.2) the current text make it sounds like the two
>> "methods" are mutually exclusive , while the 2nd one is more=20
>> an option
>> on top of the 1st one.=20
>>=20
>> I think addressing those would improve the document. So I'd like to
>> address those quickly in a new rev which would be used for the IESG
>> review. Below are the proposed edits. Please review and send
>> comments/concerns if any.
=20
As noone raised any comments/concerns, I just posted an updated version
incorporating the proposed edits below.
This is draft-ietf-tsvwg-rsvp-dste-02.txt and should be available on
IETF server shortly.

This version is ready for further IESG review.

Thanks

Francois




>> Thanks
>>=20
>> Francois
>>=20
>>=20
>> Proposed Edits on draft-ietf-tsvwg-rsvp-dste-01.txt:
>> =
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D
>>=20
>> In section 3.2,
>> Replace:
>> "The preferred method for the Aggregator to send the E2E=20
>> Path message is
>> to address it directly to the Deaggregator by setting the destination
>> address in the IP Header of the E2E Path message to the Deaggregator
>> address. "
>> By:
>> "To send the E2E Path message, the Aggregator MUST address=20
>> it directly
>> to the Deaggregator by setting the destination address in=20
>> the IP Header
>> of the E2E Path message to the Deaggregator address."
>>=20
>>=20
>> In section 3.2,=20
>> Replace:
>> "An alternate method for the Aggregator to send the E2E Path is to
>> encapsulate the E2E Path message in an IP tunnel or in the TE tunnel
>> itself and unicast the E2E Path message to the Deaggregator,=20
>> without the
>> Router Alert option. "
>> By:
>> "Optionally, the Aggregator MAY also encapsulate the E2E=20
>> Path message in
>> an IP tunnel or in the TE tunnel itself."
>>=20
>>=20
>> In section 3.2,
>> Replace:
>> "With both methods, the Router Alert is not set."
>> By:
>> "Regardless of the encapsulation method, the Router Alert is=20
>> not set."
>>=20
>>=20
>> At the very end of section 3.2,
>> Add:
>> "PathTear and ResvConf messages MUST be fowarded by the Aggregator to
>> the Deaggregator exactly like Path messages".
>>=20
>>=20
>> At the very end of section 3.6,
>> Add:
>> "On recipt of an E2E ResvConf from the Aggregator, the=20
>> Deaggregator MUST
>> forward the E2E ResvConf downstream towards the receiver. In=20
>> doing so,
>> the Deaggregator sets the destination address in the IP header of the
>> E2E ResvConf message to the IP address found in the=20
>> RESV_CONFIRM object
>> of the corresponding Resv. The Deaggregator also sets the=20
>> Router Alert.
>> "
>>=20




From tsvwg-bounces@ietf.org Sat Apr 22 19:39:39 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FXRhD-0005SP-RA; Sat, 22 Apr 2006 19:39:23 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FXRhC-0005SF-Hi
	for tsvwg@ietf.org; Sat, 22 Apr 2006 19:39:22 -0400
Received: from [166.70.100.114] (helo=mail.middle.net)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FXRhB-0003RL-7z
	for tsvwg@ietf.org; Sat, 22 Apr 2006 19:39:22 -0400
Received: from citius ([67.137.150.193] helo=[192.168.0.61])
	by mail.middle.net with esmtp (Exim 4.24)
	id 1FXRhC-0005ck-7h; Sat, 22 Apr 2006 17:39:00 -0600
Message-ID: <444ABEB4.5080800@middle.net>
Date: Sat, 22 Apr 2006 17:39:32 -0600
From: Mark Butler <butlerm@middle.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
	rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Randall Stewart <randall@lakerest.net>
Subject: Re: [Tsvwg] ambiguity in handling invalid cookie in COOKIE ECHO chunks
References: <20060421052815.1728.qmail@web37407.mail.mud.yahoo.com>	<4448D58D.3040002@lakerest.net>
	<44495A0C.50707@middle.net> <444A388E.6060704@lakerest.net>
In-Reply-To: <444A388E.6060704@lakerest.net>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-SA-Exim-Connect-IP: 67.137.150.193
X-SA-Exim-Mail-From: butlerm@middle.net
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 21c69d3cfc2dd19218717dbe1d974352
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Randall Stewart wrote:

> Mark Butler wrote:
>
>> I agree that if the COOKIE ECHO is corrupt (as opposed to stale) it 
>> should be silently discarded.
>>
>> The other conditions where one might ABORT could use some 
>> clarification. Notably in the case of lack of resources whether you 
>> should ABORT or discard depends on how long you expect the problem to 
>> last.  If indefinitely, then ABORT is the correct response.  However, 
>> if the problem is temporary, silently discarding the packet is much 
>> to be preferred.
>> In the case of a full accept queue, the latter behavior should be 
>> standardized in the SCTP socket draft, so that connect(2) calls do 
>> not error out pre-maturely.
>
>
> I agree that in a full-q one-2-one model case you probably want
> to silently discard the COOKIE and let the retransmission of the
> cookie hopefully help you get the association established... for
> BSD this is the default behavior.. but there IS a sysctl to turn
> this into a abort response too... since there will be some
> that want this..

The problem for using sysctls for stuff like this is the way 
applications are written is dependent on what behavior is assumed, such 
that you end up applications that cannot be installed on the same system 
as other applications, because they require incompatible behavior on the 
part of the SCTP stack, no matter what the sysctl option is set to.  
Consequently, whereever incompatible modes need to be supported, socket 
options should be used instead of system level options.

In the long run, I think it would be great to have control chunks that 
carried the following information:

1. Connect request has been queued.  Expected wait time (to COOKIE ACK) 
is x milliseconds.
2. Unable to queue request due to lack of resources. Retry again in x 
milliseconds.

 - Mark B.






From tsvwg-bounces@ietf.org Mon Apr 24 07:33:01 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FXzIp-0000z4-Su; Mon, 24 Apr 2006 07:32:27 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FXzIo-0000yz-1w
	for tsvwg@ietf.org; Mon, 24 Apr 2006 07:32:26 -0400
Received: from web37402.mail.mud.yahoo.com ([209.191.87.55])
	by ietf-mx.ietf.org with smtp (Exim 4.43) id 1FXzIn-0006Fp-Di
	for tsvwg@ietf.org; Mon, 24 Apr 2006 07:32:26 -0400
Received: (qmail 75595 invoked by uid 60001); 24 Apr 2006 11:32:24 -0000
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com;
	h=Message-ID:Received:Date:From:Subject:To:Cc:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding;
	b=iuqw4IZCO0SmjPxypoQC8XdnzL0a0tEpBP0JpUcIIvo5QtAln4H55DZsnnaRLOakW244+7LI3kEybIeXXM+V6DzecTZLK1kUR9LsNrGua8A8+2WgJLDOvjEZZxAEuEDop+vAPPTBfySe8mLeSPAjqvBt1hpoStUn6rraJe8Mye8=
	; 
Message-ID: <20060424113224.75593.qmail@web37402.mail.mud.yahoo.com>
Received: from [203.145.132.100] by web37402.mail.mud.yahoo.com via HTTP;
	Mon, 24 Apr 2006 04:32:24 PDT
Date: Mon, 24 Apr 2006 04:32:24 -0700 (PDT)
From: ash kat <ashwani_groups@yahoo.com>
Subject: Re: [Tsvwg] ambiguity in handling invalid cookie in COOKIE ECHO chunks
To: Randall Stewart <randall@lakerest.net>
In-Reply-To: <4448D58D.3040002@lakerest.net>
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="0-1734498613-1145878344=:68792"
Content-Transfer-Encoding: 8bit
X-Spam-Score: 1.9 (+)
X-Scan-Signature: 0cff8c3ec906d056784362c06f5f88c1
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

--0-1734498613-1145878344=:68792
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit

Hi
   
  Thanks for your response.
   
  I agree with you that the above section 5.1 deals with the lack of resources but it also says about the invalid parameter value. 
   
  However, there is still an ambiguity in section 4 and modified section 5.1 (section 2.47 of draft-ietf-tsvwg-sctpimpguide-16). Both deal with the handling of “invalid cookie (parameter) value in the COOKIE ECHO chunk”, but suggest different actions for the same.
   
  Section 4:
  =======
  1) If the State Cookie in the received COOKIE ECHO is invalid (i.e.,
      failed to pass the integrity check), the receiver MUST silently
      discard the packet.  
   
   
   
  Modified section 5.1 (section 2.47 of draft-ietf-tsvwg-sctpimpguide-16)
  ==================================================
        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 SHOULD respond with an ABORT chunk.
   
  Please note that, in case the corrupted cookie in COOKIE ECHO chunk it will be treated as the invalid parameter value. Hence the line 
  ****invalid parameter values,**** of the above section 5.1 creates ambiguity.
   
  --K.  Äsh
  

Randall Stewart <randall@lakerest.net> wrote:
  ash kat wrote:
> Hi All
> 
> I was looking at the behavior of SCTP stack when the COOKIE ECHO chunk contains the invalid cookie.
> 
> The following section of RFC2960 are conflicting each other:
> 
> Section 4 (Page# 50)
> Notes:
> 1) If the State Cookie in the received COOKIE ECHO is invalid (i.e.,
> failed to pass the integrity check), the receiver MUST silently
> discard the packet. Or, if the received State Cookie is expired
> (see Section 5.1.5), the receiver MUST send back an ERROR chunk.
> In either case, the receiver stays in the CLOSED state.
> 
> Section 5.1 (Page# 52)
> 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.
> 
> Section 5.1.5 (Page# 56)
> 
> 2) Authenticate the State Cookie as one that it previously generated
> by comparing the computed MAC against the one carried in the State
> Cookie. If this comparison fails, the SCTP packet, including the
> COOKIE ECHO and any DATA chunks, should be silently discarded.
> 
> 
> While section 2.47 of draft-ietf-tsvwg-sctpimpguide-16 has some modifications on this:
> 
> 2.47. Sending of ABORT chunks
> 2.47.1. Description of the problem
> Section 5.1 of RFC2960 [7] requires that an ABORT chunk is sent in
> response to an INIT chunk when there is no listening end point. To
> make port scanning harder someone might not want these ABORTs to be
> received by the sender of the INIT chunks. Currently the only way to
> enforce this is by using a firewall discarding the packets containing
> the INIT chunks or the packets containing the ABORT chunks. It is
> desirable that the same can be done without a middle box.
> 2.47.2. Text changes to the document
> ---------
> Old text: (Section 5.1)
> ---------
> 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.
> ---------
> New text: (Section 5.1)
> ---------
> 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 SHOULD respond
> with an ABORT chunk.
> 
> 
> But still the conflict is there.
> 
> In my point of view there should be only one of these three actions, which should be specified when the SCTP receive the COOKIE ECHO chunk with an invalid cookie.
> i) MUST silently discard (section 4 of RFC2960)
> ii) SHOULD respond with an ABORT (section 2.47 of draft-ietf-tsvwg-sctpimpguide-16)
> iii) should be silently discarded. (section 5.1.5 of RFC2960)
> 
> Please help me resolving this conflict.
> 
> 
> -K. Äsh
> 
> 
> ---------------------------------------------------------------------------------------
> Sooner or later, those who win are those who think they can.
> ---------------------------------------------------------------------------------------
> 
> 
> ---------------------------------
> How low will we go? Check out Yahoo! Messenger’s low PC-to-Phone call rates.
K. Ash:

Here I disagree with you..

In the case of a COOKIE that has been modified, the integrity
check will fail... i.e. most likely you are being attacked... In this
case, the only logical thing to do is silently discard the packet.

In the case of being short on resources... or NOT having a listener then
it really helps if an ABORT is sent back.... i.e. if an innocent
guy trys to do 'sftp xxx.com' (where sftp is an SCTP based ftp)... it is
nice if an ABORT comes back instead of him hanging around waiting for
INIT time-outs... but some, for security reasons, don't like the idea
of giving clues or hints...


These are two different classes of problems... IMO.. and thus they
are not in conflict.. they would be in conflict if you can site
two places that have the SAME thing that tell you to do different
things... i.e. if the two places that mention a corrupted cookie, not
passing the security check, told you to in one place abort and in
the other silently discard....

The document does not do that...

R

--
Randall Stewart
803-345-0369 815-342-5222(cell)


			
---------------------------------
Celebrate Earth Day everyday!  Discover 10 things you can do to help slow climate change. Yahoo! Earth Day
--0-1734498613-1145878344=:68792
Content-Type: text/html; charset=iso-8859-1
Content-Transfer-Encoding: 8bit

<DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">Hi<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" /><o:p></o:p></SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial"></SPAN>&nbsp;</DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">Thanks for your response.</SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">&nbsp;<o:p></o:p></SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">I agree with you that the above section 5.1 deals with the lack of resources but it also says about the invalid parameter value. <o:p></o:p></SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial"><o:p>&nbsp;</o:p></SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">However, there is still an
 ambiguity in section 4 and modified section 5.1 (section 2.47 of draft-ietf-tsvwg-sctpimpguide-16). Both deal with the handling of “invalid cookie (parameter) value in the COOKIE ECHO chunk”, but suggest different actions for the same.<o:p></o:p></SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial"><o:p>&nbsp;</o:p></SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">Section 4:<o:p></o:p></SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">=======<o:p></o:p></SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">1) If the State Cookie in the received COOKIE ECHO is invalid (i.e.,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; failed to pass the integrity check), <STRONG><SPAN style="COLOR: red; FONT-FAMILY: Arial">the receiver MUST silently</SPAN></STRONG><B><SPAN style="COLOR: red"><BR><STRONG><SPAN
 style="FONT-FAMILY: Arial">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; discard the packet</SPAN></STRONG></SPAN></B><SPAN style="COLOR: red">.</SPAN>&nbsp; <o:p></o:p></SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial"><o:p>&nbsp;</o:p></SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial"><o:p>&nbsp;</o:p></SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">&nbsp;<o:p></o:p></SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">Modified section 5.1 (section 2.47 of draft-ietf-tsvwg-sctpimpguide-16)<o:p></o:p></SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">==================================================<o:p></o:p></SPAN></DIV>  <DIV class=MsoNormal style="MARGIN: 0in 0in 0pt 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; If
 an endpoint receives an INIT, INIT ACK, or COOKIE ECHO chunk<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; but decides not to establish the new association due to missing<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; mandatory parameters in the received INIT or INIT ACK, <o:p></o:p></SPAN></DIV>  <DIV class=MsoNormal style="MARGIN: 0in 0in 0pt 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">****invalid parameter values,****<o:p></o:p></SPAN></DIV>  <DIV class=MsoNormal style="MARGIN: 0in 0in 0pt 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">****&nbsp;or lack of local resources,****<o:p></o:p></SPAN></DIV>  <DIV class=MsoNormal style="MARGIN: 0in 0in 0pt 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">&nbsp;<STRONG><SPAN style="COLOR: red; FONT-FAMILY: Arial">it SHOULD respond with an ABORT chunk.</SPAN></STRONG><o:p></o:p></SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">&nbsp;<o:p></o:p></SPAN></DIV>  <DIV
 style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">Please note that, in case the corrupted cookie in COOKIE ECHO chunk it will be treated as the invalid parameter value. Hence the line </SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">****invalid parameter values,**** of </SPAN>the above section 5.1 creates ambiguity.</SPAN></DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial"></SPAN>&nbsp;</DIV>  <DIV style="MARGIN-LEFT: 0.5in"><SPAN style="FONT-SIZE: 10pt; FONT-FAMILY: Arial">--K.&nbsp; Äsh</SPAN></DIV>  <DIV><BR><BR><B><I>Randall Stewart &lt;randall@lakerest.net&gt;</I></B> wrote:</DIV>  <BLOCKQUOTE class=replbq style="PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #1010ff 2px solid">ash kat wrote:<BR>&gt; Hi All<BR>&gt; <BR>&gt; I was looking at the behavior of SCTP stack when the COOKIE ECHO chunk contains
 the invalid cookie.<BR>&gt; <BR>&gt; The following section of RFC2960 are conflicting each other:<BR>&gt; <BR>&gt; Section 4 (Page# 50)<BR>&gt; Notes:<BR>&gt; 1) If the State Cookie in the received COOKIE ECHO is invalid (i.e.,<BR>&gt; failed to pass the integrity check), the receiver MUST silently<BR>&gt; discard the packet. Or, if the received State Cookie is expired<BR>&gt; (see Section 5.1.5), the receiver MUST send back an ERROR chunk.<BR>&gt; In either case, the receiver stays in the CLOSED state.<BR>&gt; <BR>&gt; Section 5.1 (Page# 52)<BR>&gt; If an endpoint receives an INIT, INIT ACK, or COOKIE ECHO chunk but<BR>&gt; decides not to establish the new association due to missing mandatory<BR>&gt; parameters in the received INIT or INIT ACK, invalid parameter<BR>&gt; values, or lack of local resources, it MUST respond with an ABORT<BR>&gt; chunk. It SHOULD also specify the cause of abort, such as the type<BR>&gt; of the missing mandatory parameters, etc., by including
 the error<BR>&gt; cause parameters with the ABORT chunk. The Verification Tag field in<BR>&gt; the common header of the outbound SCTP packet containing the ABORT<BR>&gt; chunk MUST be set to the Initiate Tag value of the peer.<BR>&gt; <BR>&gt; Section 5.1.5 (Page# 56)<BR>&gt; <BR>&gt; 2) Authenticate the State Cookie as one that it previously generated<BR>&gt; by comparing the computed MAC against the one carried in the State<BR>&gt; Cookie. If this comparison fails, the SCTP packet, including the<BR>&gt; COOKIE ECHO and any DATA chunks, should be silently discarded.<BR>&gt; <BR>&gt; <BR>&gt; While section 2.47 of draft-ietf-tsvwg-sctpimpguide-16 has some modifications on this:<BR>&gt; <BR>&gt; 2.47. Sending of ABORT chunks<BR>&gt; 2.47.1. Description of the problem<BR>&gt; Section 5.1 of RFC2960 [7] requires that an ABORT chunk is sent in<BR>&gt; response to an INIT chunk when there is no listening end point. To<BR>&gt; make port scanning harder someone might not want
 these ABORTs to be<BR>&gt; received by the sender of the INIT chunks. Currently the only way to<BR>&gt; enforce this is by using a firewall discarding the packets containing<BR>&gt; the INIT chunks or the packets containing the ABORT chunks. It is<BR>&gt; desirable that the same can be done without a middle box.<BR>&gt; 2.47.2. Text changes to the document<BR>&gt; ---------<BR>&gt; Old text: (Section 5.1)<BR>&gt; ---------<BR>&gt; If an endpoint receives an INIT, INIT ACK, or COOKIE ECHO chunk<BR>&gt; but decides not to establish the new association due to missing<BR>&gt; mandatory parameters in the received INIT or INIT ACK, invalid<BR>&gt; parameter values, or lack of local resources, it MUST respond with<BR>&gt; an ABORT chunk.<BR>&gt; ---------<BR>&gt; New text: (Section 5.1)<BR>&gt; ---------<BR>&gt; If an endpoint receives an INIT, INIT ACK, or COOKIE ECHO chunk<BR>&gt; but decides not to establish the new association due to missing<BR>&gt; mandatory parameters in
 the received INIT or INIT ACK, invalid<BR>&gt; parameter values, or lack of local resources, it SHOULD respond<BR>&gt; with an ABORT chunk.<BR>&gt; <BR>&gt; <BR>&gt; But still the conflict is there.<BR>&gt; <BR>&gt; In my point of view there should be only one of these three actions, which should be specified when the SCTP receive the COOKIE ECHO chunk with an invalid cookie.<BR>&gt; i) MUST silently discard (section 4 of RFC2960)<BR>&gt; ii) SHOULD respond with an ABORT (section 2.47 of draft-ietf-tsvwg-sctpimpguide-16)<BR>&gt; iii) should be silently discarded. (section 5.1.5 of RFC2960)<BR>&gt; <BR>&gt; Please help me resolving this conflict.<BR>&gt; <BR>&gt; <BR>&gt; -K. Äsh<BR>&gt; <BR>&gt; <BR>&gt; ---------------------------------------------------------------------------------------<BR>&gt; Sooner or later, those who win are those who think they can.<BR>&gt; ---------------------------------------------------------------------------------------<BR>&gt; <BR>&gt;
 <BR>&gt; ---------------------------------<BR>&gt; How low will we go? Check out Yahoo! Messenger’s low PC-to-Phone call rates.<BR>K. Ash:<BR><BR>Here I disagree with you..<BR><BR>In the case of a COOKIE that has been modified, the integrity<BR>check will fail... i.e. most likely you are being attacked... In this<BR>case, the only logical thing to do is silently discard the packet.<BR><BR>In the case of being short on resources... or NOT having a listener then<BR>it really helps if an ABORT is sent back.... i.e. if an innocent<BR>guy trys to do 'sftp xxx.com' (where sftp is an SCTP based ftp)... it is<BR>nice if an ABORT comes back instead of him hanging around waiting for<BR>INIT time-outs... but some, for security reasons, don't like the idea<BR>of giving clues or hints...<BR><BR><BR>These are two different classes of problems... IMO.. and thus they<BR>are not in conflict.. they would be in conflict if you can site<BR>two places that have the SAME thing that tell you to
 do different<BR>things... i.e. if the two places that mention a corrupted cookie, not<BR>passing the security check, told you to in one place abort and in<BR>the other silently discard....<BR><BR>The document does not do that...<BR><BR>R<BR><BR>--<BR>Randall Stewart<BR>803-345-0369 <OR>815-342-5222(cell)<BR></BLOCKQUOTE><BR><p>
	
		<hr size=1>Celebrate Earth Day everyday!  Discover 10 things you can do to help slow climate change. <a href="http://us.rd.yahoo.com/mail_us/taglines/earthday/*http://earth.yahoo.com">Yahoo! Earth Day</a>
--0-1734498613-1145878344=:68792--




From tsvwg-bounces@ietf.org Mon Apr 24 14:35:35 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FY5uH-0000ta-66; Mon, 24 Apr 2006 14:35:33 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FY5uF-0000tV-Tr
	for tsvwg@ietf.org; Mon, 24 Apr 2006 14:35:31 -0400
Received: from stsc1260-eth-s1-s1p1-vip.va.neustar.com ([156.154.16.129]
	helo=chiedprmail1.ietf.org)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FY5uF-00026l-RR
	for tsvwg@ietf.org; Mon, 24 Apr 2006 14:35:31 -0400
Received: from pop-altamira.atl.sa.earthlink.net ([207.69.195.62])
	by chiedprmail1.ietf.org with esmtp (Exim 4.43) id 1FY5gp-00078i-QH
	for tsvwg@ietf.org; Mon, 24 Apr 2006 14:21:41 -0400
Received: from h-68-164-86-13.snvacaid.dynamic.covad.net ([68.164.86.13]
	helo=earthlink.net)
	by pop-altamira.atl.sa.earthlink.net with esmtp (Exim 3.36 #10)
	id 1FY5gl-0007FO-00; Mon, 24 Apr 2006 14:21:35 -0400
Message-ID: <444D03F8.2A9CB6D0@earthlink.net>
Date: Mon, 24 Apr 2006 09:59:36 -0700
From: Erblichs <erblichs@earthlink.net>
X-Sender: "Erblichs" <@smtp.earthlink.net> (Unverified)
X-Mailer: Mozilla 4.72 [en]C-gatewaynet  (Win98; I)
X-Accept-Language: en
MIME-Version: 1.0
To: Sally Floyd <floyd@icir.org>
Subject: Re: [Tsvwg] WG LC fordraft-ietf-tsvwg-quickstart-03.txt? DS bit  
	Suggestion
References: <200604220035.k3M0ZjLa065560@cougar.icir.org>
Content-Type: text/plain; charset=iso-8859-2
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 22bbb45ef41b733eb2d03ee71ece8243
Cc: Pasi Sarolahti <pasi.sarolahti@iki.fi>, tsvwg@ietf.org,
	Mark Allman <mallman@icir.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Sally Floyd, Mark Allman, et al,

	If clarification is requested, I will offer a
	followup email.

	Given our past discussions and some additional
	thinking about your concern for 
	slow speed (wireless/ GPRS (Generalized Packet Radio
	Service) / etc, support the following suggestion
	is offered.

	Reference RFC 3150 : E2E Perf Implication of Slow Links

	The suggestion is a DS (Downward Scale bit) against the
	specified bps scale. This should allow a finer spacing
	of bps bits for the higher speed links and still allow
	the support for slower speed links. The scale should
	use shift operation to minimze the overhead.

	The bit is overloaded in that it SHOULD also indicate
	that:

	-- The timeout for the initial request to establish a TCP
           connection should be upwardly scaled.
           Reason : Ludwig, A multi-layer Tracing of TCP over a Reliable
	   Wireless Link, ACM Sigmetrics 1999 -- publicly indicates that 
           this is a issue to establish a number of low speed
connections.

	-- 3150: Due to the low bps scaled rate, the number of MTU sized
	   pkts is minimal. Section 2.3 Choosing MTU sizes of 3150 should
           be consulted. This would also allow a larger number of
	   pkts to be initially placed on the wire, decrease queueing
	   delays, etc..

	-- Due to the low speed environment, the coarse timeout should
	   be upwardly scaled to 30 secs.

        -- Lastly, which is probably the most controverisal, upon
percieved
	   dropouts the standard TCP MD should be scaled back (if implimented
	   at all) because LOSS in wireless / slow links are NOT an indication
	   of congestion.


	Mitchell Erblich
	------------------
	

Sally Floyd wrote:
> 
> We have just submitted a revised version of
> draft-ietf-tsvwg-quickstart-03.txt.  We believe that this is now
> ready for Working Group Last Call (finally!).
> 
> http://www.icir.org/floyd/papers/draft-ietf-tsvwg-quickstart-03.txt
> http://www.icir.org/floyd/papers/draft-ietf-tsvwg-quickstart-03.ps
> 
> - Sally
> 
> The changes, listed in the draft, are as following:
> 
>      Changes from draft-ietf-tsvwg-quickstart-02:
> 
>      * Some general editing.
> 
>      * Said that if the receiver receives a Quick-Start Request
>        with a rate of zero, then the receiver SHOULD NOT send
>        a Quick-Start response.  And that if the sender
>        receives an acknowledgement of its packet with no
>        Quick-Start response, then the sender MUST assume that the
>        request was denied, and send a Report of Approved Rate
>        with a rate of zero.
> 
>      * Said that if a Quick-Start packet is dropped or marked,
>        the sender should not make more Quick-Start requests in this
>        connection.
> 
>      * Said that the Quick-Start Request SHOULD be sent on a packet
>        that requires an acknowledgement, e.g., a SYN, SYN/ACK, or data
>        packet.
> 
>      * Made changes to the section on "TCP: A Quick-Start Request in the
>        Middle of a Connection".
> 
>      * Added that if the TCP host is going to use the successful
>        Quick-Start Request, it MUST start using it within one
>        round-trip time of receiving the Quick-Start Response,
>        or within three seconds, whichever is smaller.
> 
>      * Added a stronger applicability statement, in the abstract
>        and in Section 10 on "Implementation and Deployment Issues".
>        From feedback from the working group.
> 
>      * Added a section about MPLS.  From feedback from Mitchell
>        Erblichs.
> 
>      * Strengthened the language of the difficulties presented by
>        multi-access links.
> 
>      * Added a discussion in Section 10.3 about the deployment of
>        Quick-Start on single-hop paths.  From feedback from
>        Mitchell Erblichs.
> 
>      * Clarified that the "router" function of approving
>        Quick-Start requests includes the IP-layer processing
>        at the sender.
> 
>      * Clarified in Section 3.3 on "Processing the Quick-Start
>        Request at Routers" that this document standardizes only
>        the semantics of Quick-Start, and not the specific
>        algorithms for processing Quick-Start requests at routers.
> 
>      * Clarified in Section 3.3 on "Processing the Quick-Start
>        Request at Routers" that a router will have to consider
>        the previous Quick-Start requests in approving a new one.
> 
>      * In Section 9.3 on "Quick-Start with QoS-enabled Traffic",
>        which says that routers are free to take into account
>        the diff-serv codepoint in considering QS requests, clarified
>        that routers are also free to take into account their own
>        understanding of priorities.
> 
>      * Added the Temporary bit to Appendix D on "Possible Additional
>        Uses for the Quick-Start Option".  Clarified that the Quick-Start
>        mechanism is not designed to help routers achieve full link
>        utilization.
> 
>      * Editing from feedback from Gorry Fairhurst.




From tsvwg-bounces@ietf.org Mon Apr 24 15:50:19 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FY74U-0006BP-1Y; Mon, 24 Apr 2006 15:50:10 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FY74M-0006AH-QA; Mon, 24 Apr 2006 15:50:02 -0400
Received: from stsc1260-eth-s1-s1p1-vip.va.neustar.com ([156.154.16.129]
	helo=pine.neustar.com) by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FY74M-0006Nx-9u; Mon, 24 Apr 2006 15:50:02 -0400
Received: from stiedprstage1.ietf.org (stiedprstage1.va.neustar.com
	[10.31.47.10])
	by pine.neustar.com (8.12.8/8.12.8) with ESMTP id k3OJo1vP007909
	(version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT);
	Mon, 24 Apr 2006 19:50:02 GMT
Received: from ietf by stiedprstage1.ietf.org with local (Exim 4.43)
	id 1FY74L-0002a4-Qd; Mon, 24 Apr 2006 15:50:01 -0400
Content-Type: Multipart/Mixed; Boundary="NextPart"
Mime-Version: 1.0
To: i-d-announce@ietf.org
From: Internet-Drafts@ietf.org
Message-Id: <E1FY74L-0002a4-Qd@stiedprstage1.ietf.org>
Date: Mon, 24 Apr 2006 15:50:01 -0400
X-Spam-Score: -2.5 (--)
X-Scan-Signature: 32b73d73e8047ed17386f9799119ce43
Cc: tsvwg@ietf.org
Subject: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-rsvp-dste-02.txt 
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@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		: Aggregation of RSVP Reservations over MPLS TE/DS-TE Tunnels
	Author(s)	: F. Le Faucheur, et al.
	Filename	: draft-ietf-tsvwg-rsvp-dste-02.txt
	Pages		: 29
	Date		: 2006-4-24
	
RFC 3175 specifies aggregation of RSVP end-to-end reservations over 
aggregate RSVP reservations. This document specifies aggregation of 
RSVP end-to-end reservations over MPLS Traffic Engineering (TE) 
tunnels or MPLS Diffserv-aware MPLS Traffic Engineering (DS-TE) 
Tunnels. This approach is based on RFC 3175 and simply modifies the 
corresponding procedures for operations over MPLS TE tunnels instead 
of aggregate RSVP reservations. This approach can be used to achieve 
admission control of a very large number of flows in a scalable 
manner since the devices in the core of the network are unaware of 
the end-to-end RSVP reservations and are only aware of the MPLS TE 
tunnels.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-rsvp-dste-02.txt

To remove yourself from the I-D Announcement list, send a message to 
i-d-announce-request@ietf.org with the word unsubscribe in the body of the message.  
You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce 
to change your subscription settings.


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-rsvp-dste-02.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-rsvp-dste-02.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: <2006-4-24141658.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-tsvwg-rsvp-dste-02.txt

--OtherAccess
Content-Type: Message/External-body; name="draft-ietf-tsvwg-rsvp-dste-02.txt";
	site="ftp.ietf.org"; access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID: <2006-4-24141658.I-D@ietf.org>


--OtherAccess--

--NextPart--




From tsvwg-bounces@ietf.org Mon Apr 24 18:50:15 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FY9se-0006bA-CJ; Mon, 24 Apr 2006 18:50:08 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FY9sY-00063J-A9; Mon, 24 Apr 2006 18:50:02 -0400
Received: from stsc1260-eth-s1-s1p1-vip.va.neustar.com ([156.154.16.129]
	helo=cypress.neustar.com) by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FY9sY-0008ST-11; Mon, 24 Apr 2006 18:50:02 -0400
Received: from stiedprstage1.ietf.org (stiedprstage1.va.neustar.com
	[10.31.47.10])
	by cypress.neustar.com (8.12.8/8.12.8) with ESMTP id k3OMo10e011231
	(version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO);
	Mon, 24 Apr 2006 22:50:01 GMT
Received: from ietf by stiedprstage1.ietf.org with local (Exim 4.43)
	id 1FY9sX-0007OK-Jw; Mon, 24 Apr 2006 18:50:01 -0400
Content-Type: Multipart/Mixed; Boundary="NextPart"
Mime-Version: 1.0
To: i-d-announce@ietf.org
From: Internet-Drafts@ietf.org
Message-Id: <E1FY9sX-0007OK-Jw@stiedprstage1.ietf.org>
Date: Mon, 24 Apr 2006 18:50:01 -0400
X-Spam-Score: -2.5 (--)
X-Scan-Signature: 057ebe9b96adec30a7efb2aeda4c26a4
Cc: tsvwg@ietf.org
Subject: [Tsvwg] I-D ACTION:draft-ietf-tsvwg-quickstart-03.txt 
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@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		: Quick-Start for TCP and IP
	Author(s)	: S. Floyd, et al.
	Filename	: draft-ietf-tsvwg-quickstart-03.txt,.ps
	Pages		: 90
	Date		: 2006-4-24
	
This document specifies an optional Quick-Start mechanism for
transport protocols, in cooperation with routers, to determine an
allowed sending rate at the start and at times in the middle of a
data transfer (e.g., after an idle period).  While Quick-Start is
designed to be used by a range of transport protocols, in this
document we only specify its use with TCP.  Quick-Start is designed
to allow connections to use higher sending rates when there is
significant unused bandwidth along the path and the sender and all
of the routers along the path approve the Quick-Start Request.

As this document describes, environments where Quick-Start requests
would not be approved include paths with routers, IP tunnels, MPLS
paths, and the like that do not support Quick-Start, or paths with
routers or middleboxes that drop packets containing IP options.
Quick-Start requests could be difficult to approve over paths that
include multi-access layer-two networks.  This document also
describes environments where the Quick-Start process could fail with
false positives, with the sender incorrectly assuming that the
Quick-Start request had been approved by all of the routers along
the path.  As a result of these concerns, and as a result of the
difficulties and seeming absence of motivation for routers such as
core routers to deploy Quick-Start, Quick-Start is being proposed as
a mechanism that could be of use in controlled environments, and not
as a mechanism that would be intended or appropriate for ubiquitous
deployment in the global Internet.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-quickstart-03.txt

To remove yourself from the I-D Announcement list, send a message to 
i-d-announce-request@ietf.org with the word unsubscribe in the body of the message.  
You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce 
to change your subscription settings.


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-quickstart-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-quickstart-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: <2006-4-24154404.I-D@ietf.org>

ENCODING mime
FILE /internet-drafts/draft-ietf-tsvwg-quickstart-03.txt

--OtherAccess
Content-Type: Message/External-body; name="draft-ietf-tsvwg-quickstart-03.txt";
	site="ftp.ietf.org"; access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID: <2006-4-24154404.I-D@ietf.org>


--OtherAccess--

--NextPart--




From tsvwg-bounces@ietf.org Mon Apr 24 22:47:35 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FYDaD-0006Ic-Hl; Mon, 24 Apr 2006 22:47:21 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FYDaC-0006IX-7L
	for tsvwg@ietf.org; Mon, 24 Apr 2006 22:47:20 -0400
Received: from wyvern.icir.org ([192.150.187.14])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FYDaB-0002oj-Rc
	for tsvwg@ietf.org; Mon, 24 Apr 2006 22:47:20 -0400
Received: from guns.icir.org (adsl-69-222-35-58.dsl.bcvloh.ameritech.net
	[69.222.35.58])
	by wyvern.icir.org (8.12.11/8.12.11) with ESMTP id k3P2lGoE074501;
	Mon, 24 Apr 2006 19:47:16 -0700 (PDT)
	(envelope-from mallman@icir.org)
Received: from lawyers.icir.org (guns.icir.org [69.222.35.58])
	by guns.icir.org (Postfix) with ESMTP
	id 452BB77B164; Mon, 24 Apr 2006 22:47:14 -0400 (EDT)
Received: from lawyers.icir.org (localhost [127.0.0.1])
	by lawyers.icir.org (Postfix) with ESMTP id 1FD3F3FA23B;
	Mon, 24 Apr 2006 22:46:23 -0400 (EDT)
To: Erblichs <erblichs@earthlink.net>
From: Mark Allman <mallman@icir.org>
Subject: Re: [Tsvwg] WG LC fordraft-ietf-tsvwg-quickstart-03.txt? DS bit
	Suggestion 
In-Reply-To: <444D03F8.2A9CB6D0@earthlink.net> 
Organization: ICSI Center for Internet Research (ICIR)
Song-of-the-Day: Good Ol' Boys
MIME-Version: 1.0
Content-Type: multipart/signed; boundary="=_bOundary";
	micalg=pgp-sha1; protocol="application/pgp-signature"
Date: Mon, 24 Apr 2006 22:46:22 -0400
Message-Id: <20060425024623.1FD3F3FA23B@lawyers.icir.org>
X-Spam-Score: 0.1 (/)
X-Scan-Signature: 10d3e4e3c32e363f129e380e644649be
Cc: Pasi Sarolahti <pasi.sarolahti@iki.fi>, Sally Floyd <floyd@icir.org>,
	tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: mallman@icir.org
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

--=_bOundary
Content-Type: text/plain
Content-Disposition: inline


> 	Given our past discussions and some additional
> 	thinking about your concern for 
> 	slow speed (wireless/ GPRS (Generalized Packet Radio
> 	Service) / etc, support the following suggestion
> 	is offered.
> 
> 	Reference RFC 3150 : E2E Perf Implication of Slow Links
> 
> 	The suggestion is a DS (Downward Scale bit) against the
> 	specified bps scale. This should allow a finer spacing
> 	of bps bits for the higher speed links and still allow
> 	the support for slower speed links. The scale should
> 	use shift operation to minimze the overhead.
> 
> 	The bit is overloaded in that it SHOULD also indicate
> 	that:
> 
> 	-- The timeout for the initial request to establish a TCP
>            connection should be upwardly scaled.
>            Reason : Ludwig, A multi-layer Tracing of TCP over a Reliable
> 	   Wireless Link, ACM Sigmetrics 1999 -- publicly indicates that 
>            this is a issue to establish a number of low speed
> connections.
> 
> 	-- 3150: Due to the low bps scaled rate, the number of MTU sized
> 	   pkts is minimal. Section 2.3 Choosing MTU sizes of 3150 should
>            be consulted. This would also allow a larger number of
> 	   pkts to be initially placed on the wire, decrease queueing
> 	   delays, etc..
> 
> 	-- Due to the low speed environment, the coarse timeout should
> 	   be upwardly scaled to 30 secs.
> 
>         -- Lastly, which is probably the most controverisal, upon
> percieved
> 	   dropouts the standard TCP MD should be scaled back (if implimented
> 	   at all) because LOSS in wireless / slow links are NOT an indication
> 	   of congestion.

First, I disagree that it would ever be prudent engineering to use a
single bit to indicate everything in this disparate list.  Ever.

Second, about the notion of a finer-grained rate request...

  + I understand that you believe in the need for finer grainer rate
    requests.  That's fine.

  + In discussions with various authors of this document, you have not
    compelled us that a change in this area is needed.

  + I have not seen this idea get a bunch of traction on the tsvwg
    mailing list.

  + This document has been around for awhile and has received detailed
    review from many folks.  As far as I can recall, this issue has
    never come up in a big way.

  + Therefore, it seems to me that the current document represents
    consensus (which does not mean everyone agrees).

  + And, further, if that read is wrong and there are folks who think we
    need finer-grained requests then those folk ought to speak up now.
    We will clearly follow the consensus of the group.

Maybe the chairs have some wisdom to add here, but I think your point is
taken and we need to see input from others.

allman




--=_bOundary
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (Darwin)

iD8DBQFETY1+WyrrWs4yIs4RAuIDAKCdPB/VD/yrSfweVqEtUmTdmtpaMACgn3rr
ddGsTBguIj8Nips976XmtZE=
=bISo
-----END PGP SIGNATURE-----
--=_bOundary--




From tsvwg-bounces@ietf.org Mon Apr 24 23:58:05 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FYEga-00052u-Ph; Mon, 24 Apr 2006 23:58:00 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FYEgZ-0004si-4P
	for tsvwg@ietf.org; Mon, 24 Apr 2006 23:57:59 -0400
Received: from test-iport-2.cisco.com ([171.71.176.105])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FYEgX-0005c0-Ql
	for tsvwg@ietf.org; Mon, 24 Apr 2006 23:57:59 -0400
Received: from sj-core-2.cisco.com ([171.71.177.254])
	by test-iport-2.cisco.com with ESMTP; 24 Apr 2006 20:57:57 -0700
Received: from xbh-sjc-231.amer.cisco.com (xbh-sjc-231.cisco.com
	[128.107.191.100])
	by sj-core-2.cisco.com (8.12.10/8.12.6) with ESMTP id k3P3vvh0014269;
	Mon, 24 Apr 2006 20:57:57 -0700 (PDT)
Received: from xfe-sjc-212.amer.cisco.com ([171.70.151.187]) by
	xbh-sjc-231.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Mon, 24 Apr 2006 20:57:57 -0700
Received: from jmpolk-wxp.cisco.com ([10.89.20.31]) by
	xfe-sjc-212.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.211);
	Mon, 24 Apr 2006 20:57:56 -0700
Message-Id: <4.3.2.7.2.20060424224811.02a12488@email.cisco.com>
X-Sender: jmpolk@email.cisco.com
X-Mailer: QUALCOMM Windows Eudora Version 4.3.2
Date: Mon, 24 Apr 2006 22:57:54 -0500
To: mallman@icir.org, Erblichs <erblichs@earthlink.net>
From: "James M. Polk" <jmpolk@cisco.com>
Subject: Re: [Tsvwg] WG LC fordraft-ietf-tsvwg-quickstart-03.txt? DS
	bit Suggestion 
In-Reply-To: <20060425024623.1FD3F3FA23B@lawyers.icir.org>
References: <444D03F8.2A9CB6D0@earthlink.net>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
X-OriginalArrivalTime: 25 Apr 2006 03:57:56.0647 (UTC)
	FILETIME=[6F8C3370:01C6681C]
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 9182cfff02fae4f1b6e9349e01d62f32
Cc: Pasi Sarolahti <pasi.sarolahti@iki.fi>, Sally Floyd <floyd@icir.org>,
	tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

At 10:46 PM 4/24/2006 -0400, Mark Allman wrote:

>Maybe the chairs have some wisdom to add here, but I think your point is
>taken and we need to see input from others.

"rough consensus" is just that: "rough"

I too have not seen these points gain any traction, so far.

I have been on the minority side of a point many times, and had to agree to 
disagree, and move on.

Barring any others having strong differing beliefs they wish to express 
here, I believe TSVWG will move forward with the rough consensus model here.


>allman


Cheers,
James
IETF TSVWG co-chair

              ***********************************
     "It should NEVER be inconvenient to do the right thing"




From tsvwg-bounces@ietf.org Tue Apr 25 08:33:15 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FYMis-0005hS-Ie; Tue, 25 Apr 2006 08:32:54 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FYMir-0005hJ-T2
	for tsvwg@ietf.org; Tue, 25 Apr 2006 08:32:53 -0400
Received: from mailer1.psc.edu ([128.182.58.100])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FYMiq-0004NM-Ki
	for tsvwg@ietf.org; Tue, 25 Apr 2006 08:32:53 -0400
Received: from [192.168.1.4] (ipl-67-013.pppoe.stargate.net [206.210.67.13])
	(authenticated bits=0)
	by mailer1.psc.edu (8.13.5.20060308/8.13.3) with ESMTP id
	k3PCWpBd028497
	(version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO)
	for <tsvwg@ietf.org>; Tue, 25 Apr 2006 08:32:52 -0400 (EDT)
Message-ID: <444E16F2.1070704@psc.edu>
Date: Tue, 25 Apr 2006 08:32:50 -0400
From: John Heffner <jheffner@psc.edu>
User-Agent: Thunderbird 1.5.0.2 (Macintosh/20060308)
MIME-Version: 1.0
To: tsvwg@ietf.org
Subject: Re: [Tsvwg] Seeking comments on draft-heffner-frag-harmful-00
References: <443181B1.9030107@psc.edu>
In-Reply-To: <443181B1.9030107@psc.edu>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.0 (/)
X-Scan-Signature: d17f825e43c9aed4fd65b7edddddec89
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

I've updated the draft reflecting the discussion on the list.  Are there 
any other comments?

The new draft can be found at:
http://www.ietf.org/internet-drafts/draft-heffner-frag-harmful-01.txt

A "changebar" version can be found at:
http://www.psc.edu/~jheffner/drafts/draft-heffner-frag-harmful-01.txt.bar

Thanks,
   -John




From tsvwg-bounces@ietf.org Tue Apr 25 08:40:31 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FYMqE-00011q-Nn; Tue, 25 Apr 2006 08:40:30 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FYMqD-00011k-Ag
	for tsvwg@ietf.org; Tue, 25 Apr 2006 08:40:29 -0400
Received: from smtp5.smtp.bt.com ([217.32.164.139])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FYMqB-0005BP-65
	for tsvwg@ietf.org; Tue, 25 Apr 2006 08:40:29 -0400
Received: from i2kc06-ukbr.domain1.systemhost.net ([193.113.197.70]) by
	smtp5.smtp.bt.com with Microsoft SMTPSVC(6.0.3790.1830); 
	Tue, 25 Apr 2006 13:40:26 +0100
Received: from cbibipnt05.iuser.iroot.adidom.com ([147.149.196.177]) by
	i2kc06-ukbr.domain1.systemhost.net with Microsoft
	SMTPSVC(6.0.3790.211); Tue, 25 Apr 2006 13:40:25 +0100
Received: From bagheera.jungle.bt.co.uk ([132.146.168.158]) by
	cbibipnt05.iuser.iroot.adidom.com (WebShield SMTP v4.5 MR1a
	P0803.399); id 1145968825244; Tue, 25 Apr 2006 13:40:25 +0100
Received: from mut.jungle.bt.co.uk ([10.215.130.24])
	by bagheera.jungle.bt.co.uk (8.13.5/8.12.8) with ESMTP id
	k3PCeLbI029846; Tue, 25 Apr 2006 13:40:23 +0100
Message-Id: <5.2.1.1.2.20060425090103.037e75b8@pop3.jungle.bt.co.uk>
X-Sender: rbriscoe@pop3.jungle.bt.co.uk
X-Mailer: QUALCOMM Windows Eudora Version 5.2.1
Date: Tue, 25 Apr 2006 13:40:53 +0100
To: ken carlberg <carlberg@g11.org.uk>
From: Bob Briscoe <rbriscoe@jungle.bt.co.uk>
Subject: Re: [Tsvwg] draft-briscoe-tsvwg-cl-architecture-02.txt
In-Reply-To: <C454E278-F94B-410C-92E0-D0C24692134B@g11.org.uk>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
X-Spam-Score: -1.36 () ALL_TRUSTED
X-Scanned-By: MIMEDefang 2.56 on 132.146.168.158
X-OriginalArrivalTime: 25 Apr 2006 12:40:25.0866 (UTC)
	FILETIME=[6D2366A0:01C66865]
X-Spam-Score: 0.1 (/)
X-Scan-Signature: e16ce0269ccb2f59707d16700199d13b
Cc: tsvwg <tsvwg@ietf.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Ken,

Responses embedded to your very helpful, carefully thought through points. 
Many thanks.

But I have a feeling we've made something ambiguous that underlies all your 
questions. Or perhaps you are approaching the draft with some 
pre-conception on how it works that we haven't managed to demolish. I can't 
quite work out what it is, but we're interested in fixing it...

At 15:38 18/04/2006, ken carlberg wrote:
>Bob (and gang :-),
>
>here are some comments/questions about your draft.
>
>-ken
>===================================
>
>Page 4, Figure 1: Overall QoS Architecture.  Would there be a problem
>in extending the target architecture from that shown in Figure 1 to
>one that potentially could include two or more "islands" of CL- 
>regions?  The reason I ask is that the CL-region over the internet is
>only as large as one can extend diff-serv to multiple autonomous
>domains, which is difficult at best.  But if one could accomplish
>what you set out to do between two or more leaf ISPs with non-diff- serv 
>transit backbones, then it would seem that your design is more
>flexible, and possibly more tractable.

We considered an edge-to-edge region would be the most useful deployment 
model, but you are right that we have unnecessarily constrained ourselves. 
We should alter the draft so that the end-to-end architecture covers any 
number of big hops across PCN regions requiring admission control.

But I think your question is about islands separated by nodes doing *no* 
admission control, rather than *other* admission control. This could only 
work if you could be sure that the transit network between the leafs would 
never be congested. This would be difficult, given the motivation for flow 
pre-emption is to cope with failure scenarios. If an operator chose to take 
this risk, it could, but I believe we have a better solution...

The original reason we brought this CL over PCN work to the IETF was 
because we believed we had a way of getting it to work across multiple 
trust domains. But that aspect was a bit researchy at the time, so we 
decided to write the cl-architecture draft around a deployment model that 
addressed just one trust domain. Then, in Feb 06, I posted 
draft-briscoe-tsvwg-re-ecn-border-cheat-00.txt to document the ideas on how 
multiple domains can use this technology even if they don't trust each 
other not to cheat.

I am in the process of completing draft-01 of that. It is almost complete 
barring a fix to an attack Joe Babiarz pointed out (off-list). The latest 
sub-version is at:
<http://www.cs.ucl.ac.uk/staff/B.Briscoe/pubs.html#repcn>. [The fix is 
relatively simple -- I had unnecessarily over-penalised an attacker, which 
could be used by others as an attack.] I shall write in the correction when 
I have some time in the next couple of weeks, but the draft at the above 
URL is a better read than the one posted to internet-drafts.

I would be interested to hear if this addresses your disquiet about 
extending diffserv to multiple autonomous domains, given this draft 
proposes a different (mostly passive) approach to that used in classic 
Diffserv for inter-domain policing.


>page 5, 3'rd paragraph.  This is definitely just a nit, but could you
>add some text to be more precise when you state that PCN "can
>supplement" any PHB?  The word "can" leaves the reader with a measure
>of doubt, and the possibility that the word "supplement" may involve
>more than just a superset of functionality.

How about changing this sentence and moving it to the end of the para as 
follows?:

"  In order to support CL traffic we would expect it to supplement the 
existing Expedited Forwarding (EF). Within the controlled edge-to- edge 
region, a particular packet receives the Pre-Congestion Notification 
behaviour if the packet's DSCP (differentiated services codepoint) is set 
to EF and also the ECN field indicates ECN Capable Transport. However, PCN 
is not only intended to supplement EF. The PCN draft has been written as a 
building block so that authors of other PHBs can choose to add the PCN 
marking behaviour to their scheduling behaviour."


>page 6, 4'th paragraph.  When you mention the dependency on
>measurement, I was wondering if relatively short transaction 
>flows  becomes an issue with your design.  Phone calls today still last just
>~4.5 minutes in duration, while on the other side of the spectrum,
>SMS over telco's obviously are just brief blips.  If we extend these
>to the thousands and on, will a need arise for more than one
>measurement algorithm?  Note, this is just food for thought and
>doesn't really require an answer in your draft.  But this does lead
>to the next point....

Any admission control scheme will have a signalling etc effort which
may not be worthwhile for a very short flow. I'm not sure SMS is a good 
example, but there certainly will be cases where short flows still might 
need admission control. We envisage one measurement algorithm for short and 
long flows. That is, one marking algorithm on interior routers and one 
measurement alg in egress gateways. Having found an algorithm that gets a 
fast answer with minimal cost, there would be little point in having a 
second, slower alg.


>page 6, 5'th paragraph.  Here, and in other parts of your draft, you
>state the use of EWMA to dampen and smoothen the measurements of CL
>traffic.  In reality, what you want is a moving average of some type
>-- it could be simple, weighted, non-weighted, etc.  But given that
>this is an architecture document, is it necessary to pin it down to
>one specific type of moving average?  perhaps EWMA could be the
>default, but even then one would eventually need to state the best
>common practice regarding its parameters.

You're probably right that we have over-constrained implementation. We are 
now discussing how to specify the minimum necessary constraint on the 
algorithms for the next draft. Exponential decay is generally what is most 
desirable and it is also extremely cheap to implement, so I don't often 
consider other algs, but I'm sure we can allow any moving average.


>page 7, 1'st paragraph.  In this segment, and throughout the draft,
>you state that interior nodes within the CL-region do not hold
>"state".  Isn't it more accurate to say that they do not hold "per- flow 
>state"?  The diff-serv capable interior nodes still hold
>aggregate state, correct?

You're right that they do hold some state. But it is per outgoing interface 
card, not per aggregate (assuming the term aggregate means a 
CL-region-aggregate [ie the set of microflows between a specific pair of 
ingress and egress gateways]). We will clarify this fully in the Key 
Benefits section (2.3). Elsewhere we tried to say "no per-flow state", but 
we may have missed a few. We will check.


>Small nit.  You use the term "nodes" throughout the draft.  When I
>hear the term node, I tend to view it as an IP capable object that
>could be either a router or a host/server.  And yet in that first
>paragraph of page 7, the "node" you speak of is just a router.  I
>think it would be better if you could use the term routers and hosts
>when it is clear that is what you are speaking about, instead of
>adding ambiguity when using the term "node".

OK


>Section 1.1.2 Flow Pre-emption.  When I first came across this
>section, the questions that came to mind were: where does this occur
>-- at the edge, or interior to the CL-region?...and how does it occur
>-- through the use of signaling, or in-band state via DCSP?  These
>questions didn't get (somewhat) answered until a hint is given on the
>last paragraph of Page 24 and Section 7 on Security.  I think you
>need to address these questions in Section 1.1.2

On re-reading 1.1.2, it looks pretty clear to us which node is doing what. 
But it's obviously not clear to you, so we need to understand what 
ambiguity has crept in. Would the following change to the second para of 
1.1.2 help?


"   The solution involves four steps:

    First, triggering the ingress gateway
    to test whether pre-emption may be needed. An interior router enhanced 
with
    Pre-Congestion Notification may optionally include an algorithm that
    sets packets into the Pre-emption Marked state. Such a packet alerts
    the egress that pre-emption may be needed, which in turn sends a Pre-
    emption Alert message to the ingress.

    Secondly, calculating the right
    amount of traffic to drop. This involves the egress gateway
    measuring, and reporting to the ingress gateway, the current amount
    of CL traffic received from that particular ingress gateway. The
    ingress gateway compares this measurement (which is the amount that
    the network can actually support, and which we thus call the
    Sustainable-Aggregate-Rate) with the rate that it is sending and
    hence determines how much traffic needs to be pre-empted.

    Third, choosing which flows to shed in order to drop the traffic 
calculated in the second step. Information on the priority of flows may be 
held by the ingress gateway, or by some out of band policy decision point. 
How these systems co-ordinate to determine which flows to drop is outside 
the scope of this document, but between them they have all the information 
necessary to make the decision.

    Fourth, tearing down the chosen flows. The ingress gateway triggers 
standard tear-down messages for the signalling protocol in use.

>Also, when capacity constraints occur, would it be possible to
>introduce another model besides pre-emption that focuses on
>probability using engineered spare capacity?  Instead of a lengthy
>explanation written here, let me point you to
>the following draft:
>
>ftp://ftp.rfc-editor.org/in-notes/internet-drafts/draft-lefaucheur-emergency-rsvp-01.txt
>
>and specifically, Section 2.1.3.  If you have further questions on
>the draft or my request, we can discuss that off-line.

I guess Francois will want to comment on this part! I believe a similar 
effect can be engineered by allowing emergency calls in without admission 
control, having appropriately set the configured admission and pre-emption 
rates on interior routers. Then emergency calls can use the capacity 
between these two rates.

Alternatively, an ingress could pre-empt other flows already admitted to 
the same CL-region-aggregate whenever it admits an emergency call while 
admission control is telling it not to.

However, the purpose of this informational draft was to give one clear 
working example. We were trying not to give a load of options.

Francois, can we add some text & a ref in section 1.1.3 [which talks about 
whether both adm ctrl and prmpt are needed]?



>Page 8, 1'st paragraph.  Small nit, but the first sentence is unclear.

Assume this is the sentence "The solution has two aspects." See above - it 
is now: "   The solution involves four steps. " Better?

>Page 12, 4'th Item.  The timeline of 1-2 seconds really applies to
>voice.  User expectations for video would not seem to be under the
>same constraints.

How about this alternative bullet in the key goals:

"   o It needs to be possible to complete flow pre-emption within 1-2 
seconds.  Operators will have varying requirements but, at least for voice, 
it has been estimated that after a few seconds many affected users will 
start to hang up, making the flow pre-emption mechanism redundant and 
possibly even counter- productive. Until flow pre-emption kicks in, other 
applications using CL (e.g. video) and lower priority traffic (e.g. AF) 
could be receiving reduced service. Therefore an even faster flow 
pre-emption mechanism would be desirable (even if, in practice, operators 
have to add a deliberate pause to ride out a transient while the natural 
rate of call tear down or lower layer protection mechanisms kick in). "



>Page 13, 1'st Item.  When you speak of "session priorities", are you
>referring to the priorities one may see from signaling protocols like
>SIP and the Resource Priority header?  Or are you referring to the
>session as a flow identified by RSVP?


How about this alternative text:

"   o The CL service should support emergency services ([EMERG-RQTS],
       [EMERG-TEL]) as well as the Assured Service which is the IP
       implementation of the existing ITU-T/NATO/DoD telephone system
       architecture known as Multi-Level Pre-emption and Precedence
       [ITU.MLPP.1990] [ANSI.MLPP.Spec][ANSI.MLPP.Supplement], or MLPP.
       In particular, this involves admitting new flows that are part of 
high priority sessions even when admission control would reject other 
routine flows. Similarly, when having to choose which flows to pre-empt, 
this involves taking into account the priorities and properties of the 
sessions that flows are part of. "



>Page 15, 2'nd Item.  You should either substitute "NSIS" with "GIST",
>or expand the current text to say "signaling protocol produced by the
>NSIS working group".

Suggest the latter. I'm not sure that GIST is correct anyway, surely
QOS-NSLP?


>Page 21, 1'st Item & 1'st para.  You state "packets are pre-emption
>marked".  My question is, how and where does this occur?  Is this a
>specific (potentially new?) DSCP marking?  If so, it would be clearer
>to state this explicitly as opposed to referring to a new behavior.

Just before the phrase "packets are pre-emption marked", we say "...each 
node in the CL-region runs an algorithm...". Is this sufficient to say how 
and where this algorithm runs?

As to how pre-emption marking (and indeed admission marking) are encoded in 
a pkt, this is under discussion in the cl-phb draft (so we didn't want to 
pre-judge the outcome in this informational draft). New DSCP is one option, 
but there are others using the ECN field. Please see Appendix C 
"Pre-Congestion Notification marking" draft-briscoe-tsvwg-cl-phb-01.txt - 
we also had some discussion at the "bar BoF" in Dallas.


>Section 4.6 Tunneling.  Do you foresee marking the DSCP field in the
>outer packet header?

Whatever encoding is used, whether DSCP or ECN, we envisage it being in the 
outer header.


>Section 6.1.  I ended up getting confused by what you are trying to
>say.  It may be helpful to break up the first sentence into 2 or more
>smaller ones.

OK


Hope this all helps.



Bob Briscoe & Phil Eardley (& team)



____________________________________________________________________________
Bob Briscoe, <bob.briscoe@bt.com>      Networks Research Centre, BT Research
B54/77 Adastral Park,Martlesham Heath,Ipswich,IP5 3RE,UK.    +44 1473 645196 






From tsvwg-bounces@ietf.org Tue Apr 25 10:07:21 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FYOC7-0000J1-W3; Tue, 25 Apr 2006 10:07:11 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FYOC7-0000Iw-6Q
	for tsvwg@ietf.org; Tue, 25 Apr 2006 10:07:11 -0400
Received: from adsl-070-155-160-098.sip.cae.bellsouth.net ([70.155.160.98]
	helo=lakerest.net) by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FYOC6-00020Q-4A
	for tsvwg@ietf.org; Tue, 25 Apr 2006 10:07:11 -0400
Received: from [IPv6:::1] (localhost [IPv6:::1])
	by lakerest.net (8.13.6/8.13.4) with ESMTP id k3PE71Lf044627;
	Tue, 25 Apr 2006 10:07:03 -0400 (EDT)
	(envelope-from randall@lakerest.net)
Message-ID: <444E2D06.8000105@lakerest.net>
Date: Tue, 25 Apr 2006 10:07:02 -0400
From: Randall Stewart <randall@lakerest.net>
User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060223
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: ash kat <ashwani_groups@yahoo.com>
Subject: Re: [Tsvwg] ambiguity in handling invalid cookie in COOKIE ECHO chunks
References: <20060424113224.75593.qmail@web37402.mail.mud.yahoo.com>
In-Reply-To: <20060424113224.75593.qmail@web37402.mail.mud.yahoo.com>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: quoted-printable
X-MIME-Autoconverted: from 8bit to quoted-printable by lakerest.net id
	k3PE71Lf044627
X-Spam-Score: 0.1 (/)
X-Scan-Signature: 43317e64100dd4d87214c51822b582d1
Cc: tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org


Ahh.. now I see your point .. sorry I missed it
before... :-D

Basically this is EXACTLY the type of thing we need
to find when I get the next rev of the BIS out... (as soon
as RFC 4460 issues).. Places where we changed X (in this
case 5.1) but we forgot to change Y (in this case 4).

I will add this to the "oversight" list of the I-G.. good catch!!

R

ash kat wrote:
> Hi
>   =20
>   Thanks for your response.
>   =20
>   I agree with you that the above section 5.1 deals with the lack of re=
sources but it also says about the invalid parameter value.=20
>   =20
>   However, there is still an ambiguity in section 4 and modified sectio=
n 5.1 (section 2.47 of draft-ietf-tsvwg-sctpimpguide-16). Both deal with =
the handling of =93invalid cookie (parameter) value in the COOKIE ECHO ch=
unk=94, but suggest different actions for the same.
>   =20
>   Section 4:
>   =3D=3D=3D=3D=3D=3D=3D
>   1) If the State Cookie in the received COOKIE ECHO is invalid (i.e.,
>       failed to pass the integrity check), the receiver MUST silently
>       discard the packet. =20
>   =20
>   =20
>   =20
>   Modified section 5.1 (section 2.47 of draft-ietf-tsvwg-sctpimpguide-1=
6)
>   =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D
>         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,=20
>   ****invalid parameter values,****
>   **** or lack of local resources,****
>    it SHOULD respond with an ABORT chunk.
>   =20
>   Please note that, in case the corrupted cookie in COOKIE ECHO chunk i=
t will be treated as the invalid parameter value. Hence the line=20
>   ****invalid parameter values,**** of the above section 5.1 creates am=
biguity.
>   =20
>   --K.  =C4sh
>  =20
>=20
> Randall Stewart <randall@lakerest.net> wrote:
>   ash kat wrote:
>=20
>>Hi All
>>
>>I was looking at the behavior of SCTP stack when the COOKIE ECHO chunk =
contains the invalid cookie.
>>
>>The following section of RFC2960 are conflicting each other:
>>
>>Section 4 (Page# 50)
>>Notes:
>>1) If the State Cookie in the received COOKIE ECHO is invalid (i.e.,
>>failed to pass the integrity check), the receiver MUST silently
>>discard the packet. Or, if the received State Cookie is expired
>>(see Section 5.1.5), the receiver MUST send back an ERROR chunk.
>>In either case, the receiver stays in the CLOSED state.
>>
>>Section 5.1 (Page# 52)
>>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.
>>
>>Section 5.1.5 (Page# 56)
>>
>>2) Authenticate the State Cookie as one that it previously generated
>>by comparing the computed MAC against the one carried in the State
>>Cookie. If this comparison fails, the SCTP packet, including the
>>COOKIE ECHO and any DATA chunks, should be silently discarded.
>>
>>
>>While section 2.47 of draft-ietf-tsvwg-sctpimpguide-16 has some modific=
ations on this:
>>
>>2.47. Sending of ABORT chunks
>>2.47.1. Description of the problem
>>Section 5.1 of RFC2960 [7] requires that an ABORT chunk is sent in
>>response to an INIT chunk when there is no listening end point. To
>>make port scanning harder someone might not want these ABORTs to be
>>received by the sender of the INIT chunks. Currently the only way to
>>enforce this is by using a firewall discarding the packets containing
>>the INIT chunks or the packets containing the ABORT chunks. It is
>>desirable that the same can be done without a middle box.
>>2.47.2. Text changes to the document
>>---------
>>Old text: (Section 5.1)
>>---------
>>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.
>>---------
>>New text: (Section 5.1)
>>---------
>>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 SHOULD respond
>>with an ABORT chunk.
>>
>>
>>But still the conflict is there.
>>
>>In my point of view there should be only one of these three actions, wh=
ich should be specified when the SCTP receive the COOKIE ECHO chunk with =
an invalid cookie.
>>i) MUST silently discard (section 4 of RFC2960)
>>ii) SHOULD respond with an ABORT (section 2.47 of draft-ietf-tsvwg-sctp=
impguide-16)
>>iii) should be silently discarded. (section 5.1.5 of RFC2960)
>>
>>Please help me resolving this conflict.
>>
>>
>>-K. =C4sh
>>
>>
>>-----------------------------------------------------------------------=
----------------
>>Sooner or later, those who win are those who think they can.
>>-----------------------------------------------------------------------=
----------------
>>
>>
>>---------------------------------
>>How low will we go? Check out Yahoo! Messenger=92s low PC-to-Phone call=
 rates.
>=20
> K. Ash:
>=20
> Here I disagree with you..
>=20
> In the case of a COOKIE that has been modified, the integrity
> check will fail... i.e. most likely you are being attacked... In this
> case, the only logical thing to do is silently discard the packet.
>=20
> In the case of being short on resources... or NOT having a listener the=
n
> it really helps if an ABORT is sent back.... i.e. if an innocent
> guy trys to do 'sftp xxx.com' (where sftp is an SCTP based ftp)... it i=
s
> nice if an ABORT comes back instead of him hanging around waiting for
> INIT time-outs... but some, for security reasons, don't like the idea
> of giving clues or hints...
>=20
>=20
> These are two different classes of problems... IMO.. and thus they
> are not in conflict.. they would be in conflict if you can site
> two places that have the SAME thing that tell you to do different
> things... i.e. if the two places that mention a corrupted cookie, not
> passing the security check, told you to in one place abort and in
> the other silently discard....
>=20
> The document does not do that...
>=20
> R
>=20
> --
> Randall Stewart
> 803-345-0369 815-342-5222(cell)
>=20
>=20
> 		=09
> ---------------------------------
> Celebrate Earth Day everyday!  Discover 10 things you can do to help sl=
ow climate change. Yahoo! Earth Day


--=20
Randall Stewart
803-345-0369 <or> 815-342-5222(cell)




From tsvwg-bounces@ietf.org Tue Apr 25 15:07:39 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FYSsb-0000Ew-0z; Tue, 25 Apr 2006 15:07:21 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FYSrh-0007wN-Td
	for tsvwg@ietf.org; Tue, 25 Apr 2006 15:06:25 -0400
Received: from pop-canoe.atl.sa.earthlink.net ([207.69.195.66])
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FYSht-0000Tt-Ow
	for tsvwg@ietf.org; Tue, 25 Apr 2006 14:56:17 -0400
Received: from h-68-164-86-13.snvacaid.dynamic.covad.net ([68.164.86.13]
	helo=earthlink.net)
	by pop-canoe.atl.sa.earthlink.net with esmtp (Exim 3.36 #10)
	id 1FYShm-0007Di-00; Tue, 25 Apr 2006 14:56:10 -0400
Message-ID: <444E5D95.FDACD8CA@earthlink.net>
Date: Tue, 25 Apr 2006 10:34:13 -0700
From: Erblichs <erblichs@earthlink.net>
X-Sender: "Erblichs" <@smtp.earthlink.net> (Unverified)
X-Mailer: Mozilla 4.72 [en]C-gatewaynet  (Win98; I)
X-Accept-Language: en
MIME-Version: 1.0
To: mallman@icir.org
Subject: Re: [Tsvwg] WG LC fordraft-ietf-tsvwg-quickstart-03.txt? DS 
	bitSuggestion
References: <20060425024623.1FD3F3FA23B@lawyers.icir.org>
Content-Type: text/plain; charset=iso-8859-2
Content-Transfer-Encoding: 7bit
X-Spam-Score: 0.1 (/)
X-Scan-Signature: 32b73d73e8047ed17386f9799119ce43
Cc: Pasi Sarolahti <pasi.sarolahti@iki.fi>, Sally Floyd <floyd@icir.org>,
	tsvwg@ietf.org
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Mark Allman, et al,

	FYI, 

	The bit suggestion is really to indicate that the EXPECTATION is a
	slow-speed / wireless environment and thus the 3150 RFC Slow-speed Best
	Common Practices doc, et al should be considered with this
	connection.

	Their were very strong reasons to support the lower N values.

	Thus, if the doc is respun, CAN I suggest that a reference to 
	the 3150 RFC be made if the request is with low N values.

	Mitchell Erblich
	-----------------------------
	

Mark Allman wrote:
> 
> >       Given our past discussions and some additional
> >       thinking about your concern for
> >       slow speed (wireless/ GPRS (Generalized Packet Radio
> >       Service) / etc, support the following suggestion
> >       is offered.
> >
> >       Reference RFC 3150 : E2E Perf Implication of Slow Links
> >
> >       The suggestion is a DS (Downward Scale bit) against the
> >       specified bps scale. This should allow a finer spacing
> >       of bps bits for the higher speed links and still allow
> >       the support for slower speed links. The scale should
> >       use shift operation to minimze the overhead.
> >
> >       The bit is overloaded in that it SHOULD also indicate
> >       that:
> >
> >       -- The timeout for the initial request to establish a TCP
> >            connection should be upwardly scaled.
> >            Reason : Ludwig, A multi-layer Tracing of TCP over a Reliable
> >          Wireless Link, ACM Sigmetrics 1999 -- publicly indicates that
> >            this is a issue to establish a number of low speed
> > connections.
> >
> >       -- 3150: Due to the low bps scaled rate, the number of MTU sized
> >          pkts is minimal. Section 2.3 Choosing MTU sizes of 3150 should
> >            be consulted. This would also allow a larger number of
> >          pkts to be initially placed on the wire, decrease queueing
> >          delays, etc..
> >
> >       -- Due to the low speed environment, the coarse timeout should
> >          be upwardly scaled to 30 secs.
> >
> >         -- Lastly, which is probably the most controverisal, upon
> > percieved
> >          dropouts the standard TCP MD should be scaled back (if implimented
> >          at all) because LOSS in wireless / slow links are NOT an indication
> >          of congestion.
> 
> First, I disagree that it would ever be prudent engineering to use a
> single bit to indicate everything in this disparate list.  Ever.
> 
> Second, about the notion of a finer-grained rate request...
> 
>   + I understand that you believe in the need for finer grainer rate
>     requests.  That's fine.
> 
>   + In discussions with various authors of this document, you have not
>     compelled us that a change in this area is needed.
> 
>   + I have not seen this idea get a bunch of traction on the tsvwg
>     mailing list.
> 
>   + This document has been around for awhile and has received detailed
>     review from many folks.  As far as I can recall, this issue has
>     never come up in a big way.
> 
>   + Therefore, it seems to me that the current document represents
>     consensus (which does not mean everyone agrees).
> 
>   + And, further, if that read is wrong and there are folks who think we
>     need finer-grained requests then those folk ought to speak up now.
>     We will clearly follow the consensus of the group.
> 
> Maybe the chairs have some wisdom to add here, but I think your point is
> taken and we need to see input from others.
> 
> allman
> 
>   ------------------------------------------------------------------------
>    Part 1.2Type: application/pgp-signature




From tsvwg-bounces@ietf.org Thu Apr 27 21:19:59 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FZHeA-0007sY-MO; Thu, 27 Apr 2006 21:19:50 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FZHe7-0007oB-GC; Thu, 27 Apr 2006 21:19:47 -0400
Received: from nit.isi.edu ([128.9.160.116])
	by ietf-mx.ietf.org with esmtp (Exim 4.43)
	id 1FZHe7-0001Q7-55; Thu, 27 Apr 2006 21:19:47 -0400
Received: from nit.isi.edu (loopback [127.0.0.1])
	by nit.isi.edu (8.12.11.20060308/8.12.11) with ESMTP id k3S1JkHm002465; 
	Thu, 27 Apr 2006 18:19:46 -0700
Received: (from apache@localhost)
	by nit.isi.edu (8.12.11.20060308/8.12.11/Submit) id k3S1Jkwk002464;
	Thu, 27 Apr 2006 18:19:46 -0700
Date: Thu, 27 Apr 2006 18:19:46 -0700
Message-Id: <200604280119.k3S1Jkwk002464@nit.isi.edu>
To: ietf-announce@ietf.org, rfc-dist@rfc-editor.org
From: rfc-editor@rfc-editor.org
X-Spam-Score: -14.8 (--------------)
X-Scan-Signature: 25620135586de10c627e3628c432b04a
Cc: tsvwg@ietf.org, rfc-editor@rfc-editor.org
Subject: [Tsvwg] RFC 4460 on Stream Control Transmission Protocol (SCTP)
	Specification Errata and Issues
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org


A new Request for Comments is now available in online RFC libraries.

        
        RFC 4460

        Title:      Stream Control Transmission Protocol (SCTP) 
                    Specification Errata and Issues 
        Author:     R. Stewart, I. Arias-Rodriguez,
                    K. Poon, A. Caro,
                    M. Tuexen
        Status:     Informational
        Date:       April 2006
        Mailbox:    rrs@cisco.com, 
                    ivan.arias-rodriguez@nokia.com, 
                    kacheong.poon@sun.com,  acaro@bbn.com, 
                    tuexen@fh-muenster.de
        Pages:      109
        Characters: 215405
        Updates/Obsoletes/SeeAlso:   None

        I-D Tag:    draft-ietf-tsvwg-sctpimpguide-16.txt

        URL:        http://www.rfc-editor.org/rfc/rfc4460.txt

This document is a compilation of issues found during six
interoperability events and 5 years of experience with implementing,
testing, and using SCTP along with the suggested fixes.  This
document provides deltas to RFC 2960 and is organized in a time-based
way.  The issues are listed in the order they were brought up.
Because some text is changed several times, the last delta in the text
is the one that should be applied.  In addition to the delta, a
description of the problem and the details of the solution are also
provided.  This memo provides information for the Internet community.

This document is a product of the Transport Area Working Group
Working Group of the IETF.

INFORMATIONAL: This memo provides information for the Internet community. 
It does not specify an Internet standard of any kind. Distribution
of this memo is unlimited.

This announcement is sent to the IETF list and the RFC-DIST list.
Requests to be added to or deleted from the IETF distribution list
should be sent to IETF-REQUEST@IETF.ORG.  Requests to be
added to or deleted from the RFC-DIST distribution list should
be sent to RFC-DIST-REQUEST@RFC-EDITOR.ORG.

Details on obtaining RFCs via FTP or EMAIL may be obtained by sending
an EMAIL message to rfc-info@RFC-EDITOR.ORG with the message body 

help: ways_to_get_rfcs. For example:

        To: rfc-info@RFC-EDITOR.ORG
        Subject: getting rfcs

        help: ways_to_get_rfcs

Requests for special distribution should be addressed to either the
author of the RFC in question, or to RFC-Manager@RFC-EDITOR.ORG.  Unless
specifically noted otherwise on the RFC itself, all RFCs are for
unlimited distribution.

Submissions for Requests for Comments should be sent to
RFC-EDITOR@RFC-EDITOR.ORG.  Please consult RFC 2223, Instructions to RFC
Authors, for further information.


Joyce K. Reynolds and Sandy Ginoza
USC/Information Sciences Institute

...






From tsvwg-bounces@ietf.org Fri Apr 28 11:58:22 2006
Received: from [127.0.0.1] (helo=stiedprmman1.va.neustar.com)
	by megatron.ietf.org with esmtp (Exim 4.43)
	id 1FZVML-00036b-7h; Fri, 28 Apr 2006 11:58:21 -0400
Received: from [10.91.34.44] (helo=ietf-mx.ietf.org)
	by megatron.ietf.org with esmtp (Exim 4.43) id 1FZVMJ-00036W-9p
	for tsvwg@ietf.org; Fri, 28 Apr 2006 11:58:19 -0400
Received: from hermes.hosts.co.uk ([212.84.175.24] helo=minerva.hosts.co.uk)
	by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1FZVMH-0001Se-Ai
	for tsvwg@ietf.org; Fri, 28 Apr 2006 11:58:19 -0400
Received: from [69.138.71.89] (helo=[192.168.1.2])
	by minerva.hosts.co.uk with esmtpsa (TLSv1:RC4-SHA:128) (Exim 4.60)
	(envelope-from <carlberg@g11.org.uk>)
	id 1FZVM1-0000l9-0E; Fri, 28 Apr 2006 16:58:04 +0100
In-Reply-To: <5.2.1.1.2.20060425090103.037e75b8@pop3.jungle.bt.co.uk>
References: <5.2.1.1.2.20060425090103.037e75b8@pop3.jungle.bt.co.uk>
Mime-Version: 1.0 (Apple Message framework v749.3)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <0D94CB68-6047-4D96-B9A9-60B95A0DDCFE@g11.org.uk>
Content-Transfer-Encoding: 7bit
From: ken carlberg <carlberg@g11.org.uk>
Subject: Re: [Tsvwg] draft-briscoe-tsvwg-cl-architecture-02.txt
Date: Fri, 28 Apr 2006 11:58:08 -0400
To: Bob Briscoe <rbriscoe@jungle.bt.co.uk>
X-Mailer: Apple Mail (2.749.3)
X-Spam-Score: -1.4 (-)
X-Spam-Score: 0.0 (/)
X-Scan-Signature: f8184d7d4d1b986353eb58ea3e887935
Cc: tsvwg <tsvwg@ietf.org>
X-BeenThere: tsvwg@ietf.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Transport Area Working Group <tsvwg.ietf.org>
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=unsubscribe>
List-Post: <mailto:tsvwg@ietf.org>
List-Help: <mailto:tsvwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/tsvwg>,
	<mailto:tsvwg-request@ietf.org?subject=subscribe>
Errors-To: tsvwg-bounces@ietf.org

Bob,

Sorry about my tardy response.

> We considered an edge-to-edge region would be the most useful  
> deployment model, but you are right that we have unnecessarily  
> constrained ourselves. We should alter the draft so that the end-to- 
> end architecture covers any number of big hops across PCN regions  
> requiring admission control.

I think this would be helpful in gaining wider acceptance for your  
proposed architecture.

> But I think your question is about islands separated by nodes doing  
> *no* admission control, rather than *other* admission control. This  
> could only work if you could be sure that the transit network  
> between the leafs would never be congested. This would be  
> difficult, given the motivation for flow pre-emption is to cope  
> with failure scenarios. If an operator chose to take this risk, it  
> could, but I believe we have a better solution...

understood.  the information I tend to come across is that more times  
than not, its the Tier-2 & 3 ISPs that have a higher propensity to  
suffer congestion in comparison to the Tier-1's.  And that many of  
the Tier-1 ISPs have over-provisioned to such a degree that several  
simultaneous failures would need to occur to be subject to  
congestion.  On the other hand "many" does not equal all, so I'm  
sympathetic to your approach.

I guess what I had in mind is that if one had disjoint islands (eg,  
leaf ISPs) cooperating with each other using your proposed  
architecture, then there wouldn't be a dependency of a continual path  
of ISPs using some form of admission control.  In a sense, one would  
follow the old "smart edge, dumb core" approach and move the concept  
of the edge to the leaf ISP and the core to the Tier-1 ISP.

> The original reason we brought this CL over PCN work to the IETF  
> was because we believed we had a way of getting it to work across  
> multiple trust domains. But that aspect was a bit researchy at the  
> time, so we decided to write the cl-architecture draft around a  
> deployment model that addressed just one trust domain. Then, in Feb  
> 06, I posted draft-briscoe-tsvwg-re-ecn-border-cheat-00.txt to  
> document the ideas on how multiple domains can use this technology  
> even if they don't trust each other not to cheat.
>
> I am in the process of completing draft-01 of that. It is almost  
> complete barring a fix to an attack Joe Babiarz pointed out (off- 
> list). The latest sub-version is at:
> <http://www.cs.ucl.ac.uk/staff/B.Briscoe/pubs.html#repcn>. [The fix  
> is relatively simple -- I had unnecessarily over-penalised an  
> attacker, which could be used by others as an attack.] I shall  
> write in the correction when I have some time in the next couple of  
> weeks, but the draft at the above URL is a better read than the one  
> posted to internet-drafts.
>
> I would be interested to hear if this addresses your disquiet about  
> extending diffserv to multiple autonomous domains, given this draft  
> proposes a different (mostly passive) approach to that used in  
> classic Diffserv for inter-domain policing.

I'll need to go over the ecn-border-cheat-01 draft more carefully  
before I can say that the proposed supplemental solution would  
address concerns about extending your architecture across multiple  
administrative/trust domains.  But my gut feeling is that I think its  
easier to gain deployment/usage when an approach such as yours that  
extends over the Internet can work more as an overlay-with-gaps  
instead of natively from leaf-ISP to leaf-ISP.  Also, I wouldn't  
treat my concern as something to hold up your document, since I'm  
just speaking from a personal preference.

>> page 5, 3'rd paragraph.  This is definitely just a nit, but could you
>> add some text to be more precise when you state that PCN "can
>> supplement" any PHB?  The word "can" leaves the reader with a measure
>> of doubt, and the possibility that the word "supplement" may involve
>> more than just a superset of functionality.
>
> How about changing this sentence and moving it to the end of the  
> para as follows?:
>
> "  In order to support CL traffic we would expect it to supplement  
> the existing Expedited Forwarding (EF). Within the controlled edge- 
> to- edge region, a particular packet receives the Pre-Congestion  
> Notification behaviour if the packet's DSCP (differentiated  
> services codepoint) is set to EF and also the ECN field indicates  
> ECN Capable Transport. However, PCN is not only intended to  
> supplement EF. The PCN draft has been written as a building block  
> so that authors of other PHBs can choose to add the PCN marking  
> behaviour to their scheduling behaviour."

that's fine.

> Any admission control scheme will have a signalling etc effort which
> may not be worthwhile for a very short flow. I'm not sure SMS is a  
> good example, but there certainly will be cases where short flows  
> still might need admission control. We envisage one measurement  
> algorithm for short and long flows. That is, one marking algorithm  
> on interior routers and one measurement alg in egress gateways.  
> Having found an algorithm that gets a fast answer with minimal  
> cost, there would be little point in having a second, slower alg.

understood.  I wasn't so much advocating a "slower" algorithm, but  
rather contemplating the benefit of different measurement algorithms  
whose selection were perhaps dependent on the type of data that was  
being measures.  But again, this was just food for thought.

> You're probably right that we have over-constrained implementation.  
> We are now discussing how to specify the minimum necessary  
> constraint on the algorithms for the next draft. Exponential decay  
> is generally what is most desirable and it is also extremely cheap  
> to implement, so I don't often consider other algs, but I'm sure we  
> can allow any moving average.

ok.  I was mostly curious about leaving that door open.

> You're right that they do hold some state. But it is per outgoing  
> interface card, not per aggregate (assuming the term aggregate  
> means a CL-region-aggregate [ie the set of microflows between a  
> specific pair of ingress and egress gateways]). We will clarify  
> this fully in the Key Benefits section (2.3). Elsewhere we tried to  
> say "no per-flow state", but we may have missed a few. We will check.

ok.

> On re-reading 1.1.2, it looks pretty clear to us which node is  
> doing what. But it's obviously not clear to you, so we need to  
> understand what ambiguity has crept in. Would the following change  
> to the second para of 1.1.2 help?
>
> "   The solution involves four steps:
>
>    First, triggering the ingress gateway
>    to test whether pre-emption may be needed. An interior router  
> enhanced with
>    Pre-Congestion Notification may optionally include an algorithm  
> that
>    sets packets into the Pre-emption Marked state. Such a packet  
> alerts
>    the egress that pre-emption may be needed, which in turn sends a  
> Pre-
>    emption Alert message to the ingress.
>
>    Secondly, calculating the right
>    amount of traffic to drop. This involves the egress gateway
>    measuring, and reporting to the ingress gateway, the current amount
>    of CL traffic received from that particular ingress gateway. The
>    ingress gateway compares this measurement (which is the amount that
>    the network can actually support, and which we thus call the
>    Sustainable-Aggregate-Rate) with the rate that it is sending and
>    hence determines how much traffic needs to be pre-empted.
>
>    Third, choosing which flows to shed in order to drop the traffic  
> calculated in the second step. Information on the priority of flows  
> may be held by the ingress gateway, or by some out of band policy  
> decision point. How these systems co-ordinate to determine which  
> flows to drop is outside the scope of this document, but between  
> them they have all the information necessary to make the decision.
>
>    Fourth, tearing down the chosen flows. The ingress gateway  
> triggers standard tear-down messages for the signalling protocol in  
> use.

that's helpful.  thanks.

> I guess Francois will want to comment on this part! I believe a  
> similar effect can be engineered by allowing emergency calls in  
> without admission control, having appropriately set the configured  
> admission and pre-emption rates on interior routers. Then emergency  
> calls can use the capacity between these two rates.
>
> Alternatively, an ingress could pre-empt other flows already  
> admitted to the same CL-region-aggregate whenever it admits an  
> emergency call while admission control is telling it not to.
>
> However, the purpose of this informational draft was to give one  
> clear working example. We were trying not to give a load of options.

ok, understood.

>> Page 8, 1'st paragraph.  Small nit, but the first sentence is  
>> unclear.
>
> Assume this is the sentence "The solution has two aspects." See  
> above - it is now: "   The solution involves four steps. " Better?

yes, thanks.

>> Page 12, 4'th Item.  The timeline of 1-2 seconds really applies to
>> voice.  User expectations for video would not seem to be under the
>> same constraints.
>
> How about this alternative bullet in the key goals:
>
> "   o It needs to be possible to complete flow pre-emption within  
> 1-2 seconds.  Operators will have varying requirements but, at  
> least for voice, it has been estimated that after a few seconds  
> many affected users will start to hang up, making the flow pre- 
> emption mechanism redundant and possibly even counter- productive.  
> Until flow pre-emption kicks in, other applications using CL (e.g.  
> video) and lower priority traffic (e.g. AF) could be receiving  
> reduced service. Therefore an even faster flow pre-emption  
> mechanism would be desirable (even if, in practice, operators have  
> to add a deliberate pause to ride out a transient while the natural  
> rate of call tear down or lower layer protection mechanisms kick  
> in). "

That's fine.

>> Page 13, 1'st Item.  When you speak of "session priorities", are you
>> referring to the priorities one may see from signaling protocols like
>> SIP and the Resource Priority header?  Or are you referring to the
>> session as a flow identified by RSVP?
>
> How about this alternative text:
>
> "   o The CL service should support emergency services ([EMERG-RQTS],
>       [EMERG-TEL]) as well as the Assured Service which is the IP
>       implementation of the existing ITU-T/NATO/DoD telephone system
>       architecture known as Multi-Level Pre-emption and Precedence
>       [ITU.MLPP.1990] [ANSI.MLPP.Spec][ANSI.MLPP.Supplement], or MLPP.
>       In particular, this involves admitting new flows that are  
> part of high priority sessions even when admission control would  
> reject other routine flows. Similarly, when having to choose which  
> flows to pre-empt, this involves taking into account the priorities  
> and properties of the sessions that flows are part of. "

Sorry to be a pain about this, but again, when you bring up the term  
"session", I (the novice reader) will wonder the context you are  
talking about.  My assumption is the session/flow that  was conveyed  
by a signaling protocol like RSVP.

>> Page 15, 2'nd Item.  You should either substitute "NSIS" with "GIST",
>> or expand the current text to say "signaling protocol produced by the
>> NSIS working group".
>
> Suggest the latter. I'm not sure that GIST is correct anyway, surely
> QOS-NSLP?

ok

>> Page 21, 1'st Item & 1'st para.  You state "packets are pre-emption
>> marked".  My question is, how and where does this occur?  Is this a
>> specific (potentially new?) DSCP marking?  If so, it would be clearer
>> to state this explicitly as opposed to referring to a new behavior.
>
> Just before the phrase "packets are pre-emption marked", we say  
> "...each node in the CL-region runs an algorithm...". Is this  
> sufficient to say how and where this algorithm runs?

ok, that helps.

> As to how pre-emption marking (and indeed admission marking) are  
> encoded in a pkt, this is under discussion in the cl-phb draft (so  
> we didn't want to pre-judge the outcome in this informational  
> draft). New DSCP is one option, but there are others using the ECN  
> field. Please see Appendix C "Pre-Congestion Notification marking"  
> draft-briscoe-tsvwg-cl-phb-01.txt - we also had some discussion at  
> the "bar BoF" in Dallas.

ok.  I missed the bar BoF

>> Section 4.6 Tunneling.  Do you foresee marking the DSCP field in the
>> outer packet header?
>
> Whatever encoding is used, whether DSCP or ECN, we envisage it  
> being in the outer header.

ok, it may be helpful to state this in the draft.

cheers,

-ken





