From owner-nat@portmasters.com  Fri Jan  5 13:08:01 2001
Received: from linux2.portmasters.com (root@[12.96.1.99])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id NAA18404
	for <nat-archive@odin.ietf.org>; Fri, 5 Jan 2001 13:07:59 -0500 (EST)
Received: (from majordomo@localhost)
	by linux2.portmasters.com (8.11.1/8.11.1) id f05Hxml03270
	for nat-outgoing; Fri, 5 Jan 2001 11:59:48 -0600
X-Authentication-Warning: linux2.portmasters.com: majordomo set sender to owner-nat@portmasters.com using -f
Message-Id: <5.0.2.1.2.20010105094921.01e979d0@pop3.ipverse.com>
X-Sender: matt@ipverse.com@pop3.ipverse.com
X-Mailer: QUALCOMM Windows Eudora Version 5.0.2
Date: Fri, 05 Jan 2001 09:53:54 -0800
To: iesg-secretary@ietf.org, sob@harvard.edu, mankin@isi.edu
From: Matt Holdrege <matt@ipverse.com>
Subject: (NAT) RSIP documents finish last call
Cc: nat@portmasters.com
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Sender: owner-nat@portmasters.com
Precedence: bulk
Reply-To: Matt Holdrege <matt@ipverse.com>

The following drafts have passed WG last call. The NAT WG recommends 
publishing them as RFC's in the Experiemental classification.

draft-ietf-nat-rsip-framework-05.txt
draft-ietf-nat-rsip-protocol-07.txt
draft-ietf-nat-rsip-ipsec-04.txt
draft-ietf-nat-rsip-slp-00.txt

-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.


From owner-nat@portmasters.com  Sat Jan  6 14:34:21 2001
Received: from linux2.portmasters.com (root@[12.96.1.99])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id OAA17175
	for <nat-archive@odin.ietf.org>; Sat, 6 Jan 2001 14:34:20 -0500 (EST)
Received: (from majordomo@localhost)
	by linux2.portmasters.com (8.11.1/8.11.1) id f06JUr018168
	for nat-outgoing; Sat, 6 Jan 2001 13:30:53 -0600
X-Authentication-Warning: linux2.portmasters.com: majordomo set sender to owner-nat@portmasters.com using -f
Message-ID: <20010106192521.3187.qmail@web1406.mail.yahoo.com>
Date: Sat, 6 Jan 2001 11:25:21 -0800 (PST)
From: Pyda Srisuresh <srisuresh@yahoo.com>
Subject: (NAT) Re: Comments on RSIP
To: nat@portmasters.com
Cc: matt@ipverse.com, sob@harvard.edu, mankin@east.isi.edu,
        patrick@netzuno.com, srisuresh@yahoo.com
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: owner-nat@portmasters.com
Precedence: bulk
Reply-To: Pyda Srisuresh <srisuresh@yahoo.com>

Folks,
Folks,

There has apparently been a problem with the NAT mailing list
(nat@livingston.com) and people havent been able to post messages
to the list. My apologies. I will send an e-mail out to 
the list administrator about the problem and hopefully have it 
fixed. In the mean time, if any of you have had faced this 
problem with posting, please submit your message to 
nat@portmasters.com. Thanks.

Below, I am reposting a message concerning some issues raised 
with the RSIP documents. The message had apparently not showed 
up on the mailing list. Please review the message and add your
input to the authors and the list. Thanks.

cheers,
suresh

--- Patrick Bihan-Faou <patrick@netzuno.com> wrote:
> From: "Patrick Bihan-Faou" <patrick@netzuno.com>
> To: <jlo@ccrl.sj.nec.com>,
> 	<mike_borella@3com.com>,
> 	<david_grabelsky@3com.com>,
> 	<gab@sun.com>,
> 	<taniguti@ccrl.sj.nec.com>
> Subject: Questions about RSIP
> Date: Wed, 13 Dec 2000 15:28:32 -0500
> 
> Hi,
> 
> 
> First of all, excuse me for sending you this email directly. I subscribed to
> the NAT IETF mailing list, and sent that message there originally, but it
> never showed up on the list. If you could tell me where is the best place to
> hold this discussion, I'd appreciate it.
> 
> 
> Now onto my questions/remarks about RSIP...
> 
> First, draft-ietf-nat-rsip-framework-05.txt indicates that the RSIP gateway
> address should be indicated using the "next-server" field of the DHCP
> packets. Why was such a decision made ? More precisely, wouldn't an DHCP
> option be better suited for that ? I feel that using the "next server" field
> would cause interaction problems with older DHCP implementations (which
> could understand the "next server" as being the one that will provide a boot
> file for example). Also it does not provide the same level of flexibility as
> an option would: with an option one could specify multiple possible RSIP
> gateways either by IP address of by name to be resolved using DNS.
> 
> 
> Second, I have a question related to the RSAP-IP mode as described in "".
> 
> In the RSIP protocol definition (using RSAP-IP), a host can request a set of
> ports. These ports are supposed to be the de-multiplexing information used
> by the RSIP gateway to forward the traffic to the proper destination (RSIP
> Host). The information used for multiplexing/demultiplexing IP traffic
> largely depends on the protocol used (TCP, UDP can use ports, ESP or AH
> could use the SPI, other protocols may rely on other "unique"
> identifications). However I am surprised that nowhere in the RSIP protocol,
> the IP protocol that will be used is indicated. Namely, the RSIP parameter 2
> is used to indicate what port numbers have been assigned but without
> specifying if these are for UDP or TCP. The way I understand this is that
> once a port number is assigned to a host, it is reserved on the RSIP gateway
> for both UDP and TCP. This seems to be a waste of resources on the RSIP
> gateway. Also it is definitely protocol dependent (it can be used only for
> UDP and TCP), and is not applicable for a ESP/AH connection (there may be
> other issues with ESP/AH, I just mention that as an example). Shouldn't
> there be a protocol identifier in the RSIP message indicating which protocol
> the multiplexing info is requested for. This identifier can take the
> protocol value for which the request is made (UDP, TCP, ESP, AH, whatever
> else) and would add semantic to the current "port" parameter (context for
> interpretation: UDP => port #, ESP => SPI #). This way the "port" parameter
> would be more general (i.e. "multiplexing ids"), and could be used in more
> applications than just UDP/TCP. Also it would limit the resource waste on
> the RSIP host (no need to tie UDP port 80, because a host requested usage of
> TCP port 80 for a web server).
> 
> If I missed an important fact, please enlight me!
> 
> Patrick.
> 


=====


__________________________________________________
Do You Yahoo!?
Yahoo! Photos - Share your holiday photos online!
http://photos.yahoo.com/
-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.


From owner-nat@portmasters.com  Sun Jan  7 09:42:35 2001
Received: from linux2.portmasters.com (root@ns2.jakes.org [12.96.1.99])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id JAA08239
	for <nat-archive@odin.ietf.org>; Sun, 7 Jan 2001 09:42:34 -0500 (EST)
Received: (from majordomo@localhost)
	by linux2.portmasters.com (8.11.1/8.11.1) id f07EhCN27289
	for nat-outgoing; Sun, 7 Jan 2001 08:43:12 -0600
X-Authentication-Warning: linux2.portmasters.com: majordomo set sender to owner-nat@portmasters.com using -f
From: "Bernard Aboba" <aboba@internaut.com>
To: "Pyda Srisuresh" <srisuresh@yahoo.com>, <nat@portmasters.com>
Cc: <matt@ipverse.com>, <sob@harvard.edu>, <mankin@east.isi.edu>,
        <patrick@netzuno.com>
Subject: RE: (NAT) Re: Comments on RSIP
Date: Sun, 7 Jan 2001 06:37:44 -0800
Message-ID: <OJEJKOMOEAKLMOILFCPJAEOJDNAA.aboba@internaut.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
In-Reply-To: <20010106192521.3187.qmail@web1406.mail.yahoo.com>
X-Mimeole: Produced By Microsoft MimeOLE V5.50.4133.2400
Importance: Normal
Sender: owner-nat@portmasters.com
Precedence: bulk
Reply-To: "Bernard Aboba" <aboba@internaut.com>
Content-Transfer-Encoding: 7bit

> First, draft-ietf-nat-rsip-framework-05.txt indicates that the RSIP
gateway
> address should be indicated using the "next-server" field of the DHCP
> packets.

Indeed, the use of the "next-server" field is completely inappropriate,
since it is re-defining use of an existing field in a way that many
DHCP servers cannot easily support.

> Shouldn't there be a protocol identifier in the RSIP message indicating
which protocol
> the multiplexing info is requested for. This identifier can take the
> protocol value for which the request is made (UDP, TCP, ESP, AH, whatever
> else) and would add semantic to the current "port" parameter (context for
> interpretation: UDP => port #, ESP => SPI #).

Yes, this is another obvious problem. It would be more appropriate
to provide the protocol, offset and length of the de-multiplexing
parameter. This would be more general and would allow RSIP to work with
a number of protocols.

-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.


From owner-nat@portmasters.com  Sun Jan  7 10:12:50 2001
Received: from linux2.portmasters.com (root@ns2.jakes.org [12.96.1.99])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id KAA08376
	for <nat-archive@odin.ietf.org>; Sun, 7 Jan 2001 10:12:50 -0500 (EST)
Received: (from majordomo@localhost)
	by linux2.portmasters.com (8.11.1/8.11.1) id f07FExD27648
	for nat-outgoing; Sun, 7 Jan 2001 09:14:59 -0600
X-Authentication-Warning: linux2.portmasters.com: majordomo set sender to owner-nat@portmasters.com using -f
Message-ID: <006f01c078bb$582aeec0$288dd818@advmed.com>
From: "Jake Messinger" <jake@portmasters.com>
To: <nat@portmasters.com>
Subject: (NAT) Archives of the NAT list
Date: Sun, 7 Jan 2001 09:05:58 -0600
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 5.00.2615.200
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2615.200
Sender: owner-nat@portmasters.com
Precedence: bulk
Reply-To: "Jake Messinger" <jake@portmasters.com>
Content-Transfer-Encoding: 7bit

I have added information that will be appended to the end of each
email sent to nat@portmasters.com giving the archive URL's.

You can go directly to the archives at:

http://www.portmasters.com/cgi-bin/lwgate/NAT/archives/

~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~
Jake Messinger, VP.     ph:713-772-6690         Visit:  portmasters.com
AMS, Inc.               fx:713-774-3498                 advmed.com
8300 Bissonnet  #400    jake@jakes.org                  disc.cba.uh.edu
Houston, Texas  77074   http://jakes.org                ICQ# 4403734

Adjunct Professor       University of Houston, CBA      jake@uh.edu
~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~

                INVENTOR OF the _.,-*~'^^'~*-,._ SQUIGGLES (c) 1978




From owner-nat@portmasters.com  Sun Jan  7 15:07:46 2001
Received: from linux2.portmasters.com (root@ns2.jakes.org [12.96.1.99])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id PAA09852
	for <nat-archive@odin.ietf.org>; Sun, 7 Jan 2001 15:07:45 -0500 (EST)
Received: (from majordomo@localhost)
	by linux2.portmasters.com (8.11.1/8.11.1) id f07JmIW29873
	for nat-outgoing; Sun, 7 Jan 2001 13:48:18 -0600
X-Authentication-Warning: linux2.portmasters.com: majordomo set sender to owner-nat@portmasters.com using -f
Message-ID: <003401c078e1$78d87400$288dd818@advmed.com>
From: "Jake Messinger" <jake@portmasters.com>
To: <nat@portmasters.com>
Subject: (NAT) Footer test
Date: Sun, 7 Jan 2001 13:38:54 -0600
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 5.00.2615.200
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2615.200
Sender: owner-nat@portmasters.com
Precedence: bulk
Reply-To: "Jake Messinger" <jake@portmasters.com>
Content-Transfer-Encoding: 7bit

This is a test of the footer information. There should be information 
at the bottom of this email abou the list.

~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~
Jake Messinger, VP.     ph:713-772-6690         Visit:  portmasters.com
AMS, Inc.               fx:713-774-3498                 advmed.com
8300 Bissonnet  #400    jake@jakes.org                  disc.cba.uh.edu
Houston, Texas  77074   http://jakes.org                ICQ# 4403734

Adjunct Professor       University of Houston, CBA      jake@uh.edu
~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~

                INVENTOR OF the _.,-*~'^^'~*-,._ SQUIGGLES (c) 1978


-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.
List archive: <URL:http://www.portmasters.com/archives/>
<<  END
-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.


From owner-nat@portmasters.com  Mon Jan  8 15:00:53 2001
Received: from linux2.portmasters.com (root@[12.96.1.99])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id PAA15379
	for <nat-archive@odin.ietf.org>; Mon, 8 Jan 2001 15:00:52 -0500 (EST)
Received: (from majordomo@localhost)
	by linux2.portmasters.com (8.11.1/8.11.1) id f08JwVD13448
	for nat-outgoing; Mon, 8 Jan 2001 13:58:31 -0600
X-Authentication-Warning: linux2.portmasters.com: majordomo set sender to owner-nat@portmasters.com using -f
X-Lotus-FromDomain: 3COM
From: Mike_Borella@3com.com
To: Pyda Srisuresh <srisuresh@yahoo.com>
cc: nat@portmasters.com, matt@ipverse.com, sob@harvard.edu,
        mankin@east.isi.edu, patrick@netzuno.com, srisuresh@yahoo.com
Message-ID: <882569CE.006D165E.00@hqoutbound.ops.3com.com>
Date: Mon, 8 Jan 2001 13:51:18 -0600
Subject: Re: (NAT) Re: Comments on RSIP
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Content-Disposition: inline
Sender: owner-nat@portmasters.com
Precedence: bulk
Reply-To: Mike_Borella@3com.com




The first issue is a misunderstanding (or perhaps a lack of clarity on our
part...).
The draft specifically states that an RSIP server address can be
passed to a client with the DHCP next server *option* which has been
defined in another (now expired) draft.  If anyone would like to see
the draft, I can forward a copy.

The second issue is one that we were aware of, but it never was
specifically brought to our attention, flagged by the WG, a showstopper
in any way, shape or form, nor (IMO) is it major.  To the best of my
knowledge, there are no well known ports for which one application
uses TCP and another uses UDP.  So I'm not sure of the point of
say, giving TCP port X to one host and UDP port X to another. Not saying
that it can't be done, but any advantage of doing so would be minimal.

Thanks for the comments,
-Mike





Pyda Srisuresh <srisuresh@yahoo.com> on 01/06/2001 01:25:21 PM

Please respond to Pyda Srisuresh <srisuresh@yahoo.com>

Sent by:  Pyda Srisuresh <srisuresh@yahoo.com>


To:   nat@portmasters.com
cc:   matt@ipverse.com, sob@harvard.edu, mankin@east.isi.edu,
      patrick@netzuno.com, srisuresh@yahoo.com (Mike Borella/MW/US/3Com)
Subject:  (NAT) Re: Comments on RSIP



Folks,
Folks,

There has apparently been a problem with the NAT mailing list
(nat@livingston.com) and people havent been able to post messages
to the list. My apologies. I will send an e-mail out to
the list administrator about the problem and hopefully have it
fixed. In the mean time, if any of you have had faced this
problem with posting, please submit your message to
nat@portmasters.com. Thanks.

Below, I am reposting a message concerning some issues raised
with the RSIP documents. The message had apparently not showed
up on the mailing list. Please review the message and add your
input to the authors and the list. Thanks.

cheers,
suresh

--- Patrick Bihan-Faou <patrick@netzuno.com> wrote:
> From: "Patrick Bihan-Faou" <patrick@netzuno.com>
> To: <jlo@ccrl.sj.nec.com>,
>    <mike_borella@3com.com>,
>    <david_grabelsky@3com.com>,
>    <gab@sun.com>,
>    <taniguti@ccrl.sj.nec.com>
> Subject: Questions about RSIP
> Date: Wed, 13 Dec 2000 15:28:32 -0500
>
> Hi,
>
>
> First of all, excuse me for sending you this email directly. I subscribed to
> the NAT IETF mailing list, and sent that message there originally, but it
> never showed up on the list. If you could tell me where is the best place to
> hold this discussion, I'd appreciate it.
>
>
> Now onto my questions/remarks about RSIP...
>
> First, draft-ietf-nat-rsip-framework-05.txt indicates that the RSIP gateway
> address should be indicated using the "next-server" field of the DHCP
> packets. Why was such a decision made ? More precisely, wouldn't an DHCP
> option be better suited for that ? I feel that using the "next server" field
> would cause interaction problems with older DHCP implementations (which
> could understand the "next server" as being the one that will provide a boot
> file for example). Also it does not provide the same level of flexibility as
> an option would: with an option one could specify multiple possible RSIP
> gateways either by IP address of by name to be resolved using DNS.
>
>
> Second, I have a question related to the RSAP-IP mode as described in "".
>
> In the RSIP protocol definition (using RSAP-IP), a host can request a set of
> ports. These ports are supposed to be the de-multiplexing information used
> by the RSIP gateway to forward the traffic to the proper destination (RSIP
> Host). The information used for multiplexing/demultiplexing IP traffic
> largely depends on the protocol used (TCP, UDP can use ports, ESP or AH
> could use the SPI, other protocols may rely on other "unique"
> identifications). However I am surprised that nowhere in the RSIP protocol,
> the IP protocol that will be used is indicated. Namely, the RSIP parameter 2
> is used to indicate what port numbers have been assigned but without
> specifying if these are for UDP or TCP. The way I understand this is that
> once a port number is assigned to a host, it is reserved on the RSIP gateway
> for both UDP and TCP. This seems to be a waste of resources on the RSIP
> gateway. Also it is definitely protocol dependent (it can be used only for
> UDP and TCP), and is not applicable for a ESP/AH connection (there may be
> other issues with ESP/AH, I just mention that as an example). Shouldn't
> there be a protocol identifier in the RSIP message indicating which protocol
> the multiplexing info is requested for. This identifier can take the
> protocol value for which the request is made (UDP, TCP, ESP, AH, whatever
> else) and would add semantic to the current "port" parameter (context for
> interpretation: UDP => port #, ESP => SPI #). This way the "port" parameter
> would be more general (i.e. "multiplexing ids"), and could be used in more
> applications than just UDP/TCP. Also it would limit the resource waste on
> the RSIP host (no need to tie UDP port 80, because a host requested usage of
> TCP port 80 for a web server).
>
> If I missed an important fact, please enlight me!
>
> Patrick.
>


=====


__________________________________________________
Do You Yahoo!?
Yahoo! Photos - Share your holiday photos online!
http://photos.yahoo.com/
-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.




-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.
List archive: <URL:http://www.portmasters.com/archives/>
<<  END
-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.


From owner-nat@portmasters.com  Tue Jan  9 00:18:22 2001
Received: from linux2.portmasters.com (root@ns2.jakes.org [12.96.1.99])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id AAA23663
	for <nat-archive@odin.ietf.org>; Tue, 9 Jan 2001 00:18:22 -0500 (EST)
Received: (from majordomo@localhost)
	by linux2.portmasters.com (8.11.1/8.11.1) id f095GTl18878
	for nat-outgoing; Mon, 8 Jan 2001 23:16:29 -0600
X-Authentication-Warning: linux2.portmasters.com: majordomo set sender to owner-nat@portmasters.com using -f
Message-Id: <5.0.2.1.2.20010108210622.02308370@pop3.ipverse.com>
X-Sender: matt@ipverse.com@pop3.ipverse.com
X-Mailer: QUALCOMM Windows Eudora Version 5.0.2
Date: Mon, 08 Jan 2001 21:10:28 -0800
To: nat@portmasters.com
From: Matt Holdrege <matt@ipverse.com>
Subject: RE: (NAT) Re: Comments on RSIP
Cc: patrick@netzuno.com
In-Reply-To: <OJEJKOMOEAKLMOILFCPJAEOJDNAA.aboba@internaut.com>
References: <20010106192521.3187.qmail@web1406.mail.yahoo.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Sender: owner-nat@portmasters.com
Precedence: bulk
Reply-To: Matt Holdrege <matt@ipverse.com>


> > Shouldn't there be a protocol identifier in the RSIP message indicating
>which protocol
> > the multiplexing info is requested for. This identifier can take the
> > protocol value for which the request is made (UDP, TCP, ESP, AH, whatever
> > else) and would add semantic to the current "port" parameter (context for
> > interpretation: UDP => port #, ESP => SPI #).

Seeing as how this is proposed as an experimental RFC, I think such input 
would be useful during the experimentation of this protocol. As for now, 
I'm inclined to letting the RSIP experimenters play with the protocol as 
is. Your suggestion is a good one and I wish we had heard it a year ago. 
But I expect many more useful comments as folks continue to experiment with 
RSIP.

-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.
List archive: <URL:http://www.portmasters.com/archives/>


From owner-nat@portmasters.com  Tue Jan  9 09:19:10 2001
Received: from linux2.portmasters.com (root@ns2.jakes.org [12.96.1.99])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id JAA12125
	for <nat-archive@odin.ietf.org>; Tue, 9 Jan 2001 09:19:10 -0500 (EST)
Received: (from majordomo@localhost)
	by linux2.portmasters.com (8.11.1/8.11.1) id f09EJmZ22910
	for nat-outgoing; Tue, 9 Jan 2001 08:19:48 -0600
X-Authentication-Warning: linux2.portmasters.com: majordomo set sender to owner-nat@portmasters.com using -f
Message-ID: <20010109141356.11737.qmail@web1406.mail.yahoo.com>
Date: Tue, 9 Jan 2001 06:13:56 -0800 (PST)
From: Pyda Srisuresh <srisuresh@yahoo.com>
Subject: (NAT) IETF NAT WG web page updated with new mailing list info
To: nat@portmasters.com
Cc: Allison Mankin <mankin@east.isi.edu>, Pyda Srisuresh <srisuresh@yahoo.com>,
        Scott Bradner <sob@harvard.edu>, Matt Holdrege <matt@ipverse.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: owner-nat@portmasters.com
Precedence: bulk
Reply-To: Pyda Srisuresh <srisuresh@yahoo.com>

Folks,

Please note the change in mailing list information on the NAT working 
group WEB page as follows. The change is prompted by the problems 
noticed with the previous list ID (nat@livingston.com) and per the  
advice of the list administrator. Sorry for the inconvenience. 
Thanks for your understanding.

          General Discussion:nat@portmasters.com 
          To Subscribe: nat-request@portmasters.com 
          Archive: http://www.portmasters.com/cgi-bin/lwgate/NAT/archives/

cheers,
suresh


=====


__________________________________________________
Do You Yahoo!?
Yahoo! Photos - Share your holiday photos online!
http://photos.yahoo.com/
-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.
List archive: <URL:http://www.portmasters.com/archives/>


From owner-nat@portmasters.com  Thu Jan 11 14:11:47 2001
Received: from linux2.portmasters.com (root@[12.96.1.99])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id OAA12709
	for <nat-archive@odin.ietf.org>; Thu, 11 Jan 2001 14:11:46 -0500 (EST)
Received: (from majordomo@localhost)
	by linux2.portmasters.com (8.11.1/8.11.1) id f0BJ3f729626
	for nat-outgoing; Thu, 11 Jan 2001 13:03:41 -0600
X-Authentication-Warning: linux2.portmasters.com: majordomo set sender to owner-nat@portmasters.com using -f
Message-ID: <003401c07c00$c491af20$1501a8c0@ttsdemo.net>
From: "Torbjorn Friberg" <onsite-admin@tts-demo.net>
To: <nat@portmasters.com>
Subject: (NAT) Digital Certificates and NAT
Date: Thu, 11 Jan 2001 20:00:30 +0100
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_NextPart_000_0031_01C07C09.2623BC80"
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 5.50.4133.2400
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400
Sender: owner-nat@portmasters.com
Precedence: bulk
Reply-To: "Torbjorn Friberg" <onsite-admin@tts-demo.net>

This is a multi-part message in MIME format.

------=_NextPart_000_0031_01C07C09.2623BC80
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Hi,=20

When trying to access a protected resource over the Internet I use a =
digital certificate to authenticate my selft (over SSL). But when on a =
network using NAT then it is not possible.=20

Why? and what is the solution to this?

Torbjorn Friberg

------=_NextPart_000_0031_01C07C09.2623BC80
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=3DContent-Type content=3D"text/html; =
charset=3Diso-8859-1">
<META content=3D"MSHTML 5.50.4134.600" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT face=3DVerdana size=3D2>Hi, </FONT></DIV>
<DIV><FONT face=3DVerdana size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3DVerdana size=3D2>When trying to access a protected =
resource over=20
the Internet I use a digital certificate to authenticate my selft (over =
SSL).=20
But when on a network using NAT then it is not possible. </FONT></DIV>
<DIV><FONT face=3DVerdana size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3DVerdana size=3D2>Why? and what is the solution to=20
this?</FONT></DIV>
<DIV><FONT face=3DVerdana size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3DVerdana size=3D2>Torbjorn =
Friberg</FONT></DIV></BODY></HTML>

------=_NextPart_000_0031_01C07C09.2623BC80--

-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.
List archive: <URL:http://www.portmasters.com/archives/>


From owner-nat@portmasters.com  Sat Jan 13 17:43:28 2001
Received: from linux2.portmasters.com (root@ns2.jakes.org [12.96.1.99])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id RAA20577
	for <nat-archive@odin.ietf.org>; Sat, 13 Jan 2001 17:43:26 -0500 (EST)
Received: (from majordomo@localhost)
	by linux2.portmasters.com (8.11.1/8.11.1) id f0DMTDp11168
	for nat-outgoing; Sat, 13 Jan 2001 16:29:13 -0600
X-Authentication-Warning: linux2.portmasters.com: majordomo set sender to owner-nat@portmasters.com using -f
Message-ID: <20010113223027.559.qmail@web1406.mail.yahoo.com>
Date: Sat, 13 Jan 2001 14:30:27 -0800 (PST)
From: Pyda Srisuresh <srisuresh@yahoo.com>
Subject: Re: (NAT) Digital Certificates and NAT
To: Torbjorn Friberg <onsite-admin@tts-demo.net>, nat@portmasters.com
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: owner-nat@portmasters.com
Precedence: bulk
Reply-To: Pyda Srisuresh <srisuresh@yahoo.com>

Hi Torbjorn,

Were you using SSL to authenticate yourself (the client) or the server?
I have used https in the past from behind a NAT and it works fine. But, 
that may have been likely just the server authentication. I dont think,
it would have failed even for client authentication.

SSL authentication can fail for several reasons. I dont think, NAT is
the reason. For example, some SSL servers may have disabled weaker ciphers 
(say 40-bit cyphers) to prevent SSL connections with weaker encryption. 
Due to U.S. government restrictions, this effectively restricts access 
to network browsers shipped by the US companies to firms outside the US
(unless, they were shipped with case-by-case approval from the govts.).

Have you tried this without NAT?

cheers,
suresh 
 
--- Torbjorn Friberg <onsite-admin@tts-demo.net> wrote:
> Hi, 
> 
> When trying to access a protected resource over the Internet I use a digital
> certificate to authenticate my selft (over SSL). But when on a network using
> NAT then it is not possible. 
> 
> Why? and what is the solution to this?
> 
> Torbjorn Friberg
> 


=====


__________________________________________________
Do You Yahoo!?
Get email at your own domain with Yahoo! Mail. 
http://personal.mail.yahoo.com/
-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.
List archive: <URL:http://www.portmasters.com/archives/>


From owner-nat@portmasters.com  Mon Jan 15 01:43:18 2001
Received: from linux2.portmasters.com (root@ns2.jakes.org [12.96.1.99])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id BAA20575
	for <nat-archive@odin.ietf.org>; Mon, 15 Jan 2001 01:43:17 -0500 (EST)
Received: (from majordomo@localhost)
	by linux2.portmasters.com (8.11.1/8.11.1) id f0F6Y0O29353
	for nat-outgoing; Mon, 15 Jan 2001 00:34:00 -0600
X-Authentication-Warning: linux2.portmasters.com: majordomo set sender to owner-nat@portmasters.com using -f
Date: Sun, 14 Jan 2001 23:48:37 -0700 (MST)
From: Randy Ridley <rridley@anr.net>
X-Sender: rridley@localhost.localdomain
To: nat@portmasters.com
Subject: (NAT) default route with nat problem
Message-ID: <Pine.LNX.4.10.10101142343470.3762-100000@localhost.localdomain>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-nat@portmasters.com
Precedence: bulk
Reply-To: Randy Ridley <rridley@anr.net>

Hello all :-)

I'm trying to recycle my old pm2 by using it as the dialin server for my
home business machines.

I am using the 10.0.0.0/24 network for my home machines with comos3.9-usa
on the portmaster.

setup is right out of the release notes under NAT Examples.

The problem I am having is that I have a static IP address assigned from
my ISP, but I am dialing into a pool of servers with varying IP's. I need
the default route of the Portmaster to follow whichever IP i am dialed
into.

The system works great if I telnet to the portmaster and manualy set the
route, but that's a real bummer when you want dial-on-demand :-(

Any help is appreciated.

Randy

-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.
List archive: <URL:http://www.portmasters.com/archives/>


From owner-nat@portmasters.com  Fri Jan 19 12:16:55 2001
Received: from linux2.portmasters.com (root@[12.96.1.99])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id MAA02757
	for <nat-archive@odin.ietf.org>; Fri, 19 Jan 2001 12:16:54 -0500 (EST)
Received: (from majordomo@localhost)
	by linux2.portmasters.com (8.11.1/8.11.1) id f0JGxP302755
	for nat-outgoing; Fri, 19 Jan 2001 10:59:25 -0600
X-Authentication-Warning: linux2.portmasters.com: majordomo set sender to owner-nat@portmasters.com using -f
Message-Id: <5.0.2.1.2.20010119085851.021c21d0@pop3.ipverse.com>
X-Sender: matt@ipverse.com@pop3.ipverse.com
X-Mailer: QUALCOMM Windows Eudora Version 5.0.2
Date: Fri, 19 Jan 2001 09:00:00 -0800
To: "Patrick Bihan-Faou" <patrick@netzuno.com>, <nat@portmasters.com>
From: Matt Holdrege <matt@ipverse.com>
Subject: Re: (NAT) Re: Comments on RSIP
In-Reply-To: <HJEEKLMFLKEOKHOKNPBMOEPECIAA.patrick@netzuno.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Sender: owner-nat@portmasters.com
Precedence: bulk
Reply-To: Matt Holdrege <matt@ipverse.com>

At 04:47 PM 1/18/2001, Patrick Bihan-Faou wrote:


>"Matt Holdrege" <matt@ipverse.com> wrote in message
>news:<5.0.2.1.2.20010108210622.02308370@pop3.ipverse.com>...
> >
> > > > Shouldn't there be a protocol identifier in the RSIP message
>indicating
> > >which protocol
> > > > the multiplexing info is requested for. This identifier can take the
> > > > protocol value for which the request is made (UDP, TCP, ESP, AH,
>whatever
> > > > else) and would add semantic to the current "port" parameter (context
>for
> > > > interpretation: UDP => port #, ESP => SPI #).
> >
> > Seeing as how this is proposed as an experimental RFC, I think such input
> > would be useful during the experimentation of this protocol. As for now,
> > I'm inclined to letting the RSIP experimenters play with the protocol as
> > is. Your suggestion is a good one and I wish we had heard it a year ago.
> > But I expect many more useful comments as folks continue to experiment
>with
> > RSIP.
> >
>
>Well since we can already identify these shortcommings is it completely
>reasonable to say "go ahead implement this" knowing very well that we will
>have to change this later.
>
>I can see people in the future saying "well we already have implementations
>out there, we can not break them, so we have to keep supporting this broken
>scheme". Do we really want to face that situation by publishing the current
>"broken" (in quotes) document as a RFC ?

This is precisely the reason that the IETF has a classification for 
Experimental RFC's. That way everyone knows that it is highly likely the 
protocol will change.

-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.
List archive: <URL:http://www.portmasters.com/archives/>


From owner-nat@portmasters.com  Fri Jan 26 22:30:08 2001
Received: from linux2.portmasters.com (root@ns2.jakes.org [12.96.1.99])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id WAA28220
	for <nat-archive@odin.ietf.org>; Fri, 26 Jan 2001 22:30:08 -0500 (EST)
Received: (from majordomo@localhost)
	by linux2.portmasters.com (8.11.1/8.11.1) id f0R3J6005884
	for nat-outgoing; Fri, 26 Jan 2001 21:19:06 -0600
X-Authentication-Warning: linux2.portmasters.com: majordomo set sender to owner-nat@portmasters.com using -f
Message-ID: <3A723E0E.23F0CC6A@eyematic.com>
Date: Fri, 26 Jan 2001 19:18:38 -0800
From: Tim Dorcey <tim@eyematic.com>
X-Mailer: Mozilla 4.51 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: nat@portmasters.com
Subject: (NAT) Comment on RFC 2663
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-nat@portmasters.com
Precedence: bulk
Reply-To: Tim Dorcey <tim@eyematic.com>
Content-Transfer-Encoding: 7bit

In trying to give some users guidance on if/how a particular application
would work with their particular firewall/IP
sharing/router/modem/masquerading/NAT product, I was hoping I might find
the terminology I needed in RFC 2663.  The distinction I was looking for
was between what I would call "connection-oriented" versus
"connection-less" NAT.  A connection-oriented NAT maintains a 1:1
mapping between (private addr,private port,remote addr,remote port) <->
(public addr,public port).  A connection-less NAT maintains a 1:1
mapping between (private addr,private port) <-> (public addr, public
port), regardless of how many remote end-points the (private
addr,private port) may be interacting with.

This issue is not the same as the one-directional versus bi-directional
distinction discussed in RFC 2663.  For example, a NAT could be
connection-less, but still enforce the rule that incoming packets are
rejected unless they originate at an address to which a packet has been
recently sent.  And, theoretically at least, a NAT could be
connection-oriented while still allowing the first packet of a session
to originate outside the private network.

The distinction is significant because "telephone-like" applications can
function easily in a connection-less NAT environment as follows.  When
the application wishes to accept calls, it can establish a session with
a directory server in the public address space, which can then provide
to other hosts the (public addr, public port) at which this application
is currently reachable.  If the NAT is configured to prohibit external
initiations, then 2 applications who wish to communicate with one
another can both initiate packets toward each other (coordinated by the
directory server or some other intermediary in the public address
space), signalling their NAT that incoming traffic from that end-point
is desired.

Of course, this won't work if the NAT is connection-oriented, since
"each other" is not defined until packets are initiated to each other.
One might say that connection-oriented NAT is therefore more secure,
since it prevents 2 parties from using coordinated simultaneous
initiations to get around a block on external initiations.  But, the
only extra security is that a host behind a connection-oriented NAT
could not be triggered into establishing a connection with another host
who is also behind a connection-oriented NAT (or any device that blocks
external initiations).  That doesn't sound like much of a security
feature.  It does sound like a major impediment to desired
communications.

Whether it is a good thing or a bad thing, we really need some accepted
terminology for this distinction.  E.g., recent releases of Linux IP
Masquerade will operate in "connection-less" mode for UDP traffic if you
enable the "LooseUDP" option.  Who knows what it is called (if it is
available) with other NAT implementations.   I do not believe the RFC
2663 discussion of "one-directional" versus "bi-directional" NAT is
adequate.  And, I don't believe an application that depends on 1:1
end-point mappings should need an ALG to operate through a NAT.  If
anything, applications that want connections mapped to end-points should
need an ALG.

Tim

-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.
List archive: <URL:http://www.portmasters.com/archives/>


From owner-nat@portmasters.com  Sun Jan 28 13:04:51 2001
Received: from linux2.portmasters.com (root@ns2.jakes.org [12.96.1.99])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id NAA09665
	for <nat-archive@odin.ietf.org>; Sun, 28 Jan 2001 13:04:50 -0500 (EST)
Received: (from majordomo@localhost)
	by linux2.portmasters.com (8.11.1/8.11.1) id f0SHv0T22566
	for nat-outgoing; Sun, 28 Jan 2001 11:57:00 -0600
X-Authentication-Warning: linux2.portmasters.com: majordomo set sender to owner-nat@portmasters.com using -f
Message-ID: <20010128175630.25786.qmail@web1405.mail.yahoo.com>
Date: Sun, 28 Jan 2001 09:56:30 -0800 (PST)
From: Pyda Srisuresh <srisuresh@yahoo.com>
Subject: Re: (NAT) Comment on RFC 2663
To: Tim Dorcey <tim@eyematic.com>, nat@portmasters.com
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: owner-nat@portmasters.com
Precedence: bulk
Reply-To: Pyda Srisuresh <srisuresh@yahoo.com>

Tim,

NAT is basically session oriented. Below is text from RFC 2663, 
section 9, concerning the security issue with UDP sessions.
        
   1. UDP sessions are inherently unsafe. Responses to a datagram
      could come from an address different from the target address
      used by sender ([Ref 4]). As a result, an incoming UDP packet
      might match the outbound session of a traditional NAT router
      only in part (the destination address and UDP port number of
      the packet match, but the source address and port number may
      not). In such a case, there is a potential security compromise
      for the NAT device in permitting inbound packets with partial
      match. This UDP security issue is also inherent to firewalls.

      Traditional NAT implementations that do not track datagrams on
      a per-session basis but lump states of multiple UDP sessions
      using the same address binding into a single unified session
      could compromise the security even further. This is because,
      the granularity of packet matching would be further limited to
      just the destination address of the inbound UDP packets.


RFC 3022 (Traditional NAT) says the following about address (and port)
binding for NAPT in section 3.1

   In the case of NAPT, where many private addresses are mapped to a
   single globally unique address, the binding would be from the tuple
   of (private address, private TU port) to the tuple of (assigned
   address, assigned TU port).  As with Basic NAT, this binding is
   determined when the first outgoing session is initiated by the tuple
   of (private address, private TU port) on the private host.  While not
   a common practice, it is possible to have an application on private
   host establish multiple simultaneous sessions originating from the
   same tuple of (private address, private TU port).  In such a case, a
   single binding for the tuple of (private address, private TU port)
   may be used for translation of packets pertaining to all sessions
   originating from the same tuple on a host.

As you can see, neither of the RFCs prohibit the use of UDP port 
binding in the manner you suggest. It is an implementation option 
as to how UDP session binding is done for individual UDP based 
applications. For example, the approach you suggest was pursued by 
Activision folks for their game protocol design (Ref: Section 5.1, 
RC3027).

Hope this helps.

cheers,
suresh

--- Tim Dorcey <tim@eyematic.com> wrote:
> In trying to give some users guidance on if/how a particular application
> would work with their particular firewall/IP
> sharing/router/modem/masquerading/NAT product, I was hoping I might find
> the terminology I needed in RFC 2663.  The distinction I was looking for
> was between what I would call "connection-oriented" versus
> "connection-less" NAT.  A connection-oriented NAT maintains a 1:1
> mapping between (private addr,private port,remote addr,remote port) <->
> (public addr,public port).  A connection-less NAT maintains a 1:1
> mapping between (private addr,private port) <-> (public addr, public
> port), regardless of how many remote end-points the (private
> addr,private port) may be interacting with.
> 
> This issue is not the same as the one-directional versus bi-directional
> distinction discussed in RFC 2663.  For example, a NAT could be
> connection-less, but still enforce the rule that incoming packets are
> rejected unless they originate at an address to which a packet has been
> recently sent.  And, theoretically at least, a NAT could be
> connection-oriented while still allowing the first packet of a session
> to originate outside the private network.
> 
> The distinction is significant because "telephone-like" applications can
> function easily in a connection-less NAT environment as follows.  When
> the application wishes to accept calls, it can establish a session with
> a directory server in the public address space, which can then provide
> to other hosts the (public addr, public port) at which this application
> is currently reachable.  If the NAT is configured to prohibit external
> initiations, then 2 applications who wish to communicate with one
> another can both initiate packets toward each other (coordinated by the
> directory server or some other intermediary in the public address
> space), signalling their NAT that incoming traffic from that end-point
> is desired.
> 
> Of course, this won't work if the NAT is connection-oriented, since
> "each other" is not defined until packets are initiated to each other.
> One might say that connection-oriented NAT is therefore more secure,
> since it prevents 2 parties from using coordinated simultaneous
> initiations to get around a block on external initiations.  But, the
> only extra security is that a host behind a connection-oriented NAT
> could not be triggered into establishing a connection with another host
> who is also behind a connection-oriented NAT (or any device that blocks
> external initiations).  That doesn't sound like much of a security
> feature.  It does sound like a major impediment to desired
> communications.
> 
> Whether it is a good thing or a bad thing, we really need some accepted
> terminology for this distinction.  E.g., recent releases of Linux IP
> Masquerade will operate in "connection-less" mode for UDP traffic if you
> enable the "LooseUDP" option.  Who knows what it is called (if it is
> available) with other NAT implementations.   I do not believe the RFC
> 2663 discussion of "one-directional" versus "bi-directional" NAT is
> adequate.  And, I don't believe an application that depends on 1:1
> end-point mappings should need an ALG to operate through a NAT.  If
> anything, applications that want connections mapped to end-points should
> need an ALG.
> 
> Tim
> 
> -
> To unsubscribe, email 'majordomo@portmasters.com' with
> 'unsubscribe nat' in the body of the message.
> List archive: <URL:http://www.portmasters.com/archives/>


=====


__________________________________________________
Do You Yahoo!?
Yahoo! Auctions - Buy the things you want at great prices. 
http://auctions.yahoo.com/
-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.
List archive: <URL:http://www.portmasters.com/archives/>


From owner-nat@portmasters.com  Sun Jan 28 17:45:46 2001
Received: from linux2.portmasters.com (root@ns2.jakes.org [12.96.1.99])
	by ietf.org (8.9.1a/8.9.1a) with SMTP id RAA10703
	for <nat-archive@odin.ietf.org>; Sun, 28 Jan 2001 17:45:46 -0500 (EST)
Received: (from majordomo@localhost)
	by linux2.portmasters.com (8.11.1/8.11.1) id f0SMfAN24729
	for nat-outgoing; Sun, 28 Jan 2001 16:41:10 -0600
X-Authentication-Warning: linux2.portmasters.com: majordomo set sender to owner-nat@portmasters.com using -f
Message-ID: <3A749FDE.D4DE9AAE@eyematic.com>
Date: Sun, 28 Jan 2001 14:40:30 -0800
From: Tim Dorcey <tim@eyematic.com>
X-Mailer: Mozilla 4.51 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: Pyda Srisuresh <srisuresh@yahoo.com>
CC: nat@portmasters.com
Subject: Re: (NAT) Comment on RFC 2663
References: <20010128175630.25786.qmail@web1405.mail.yahoo.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-nat@portmasters.com
Precedence: bulk
Reply-To: Tim Dorcey <tim@eyematic.com>
Content-Transfer-Encoding: 7bit

Pyda Srisuresh wrote:

> As you can see, neither of the RFCs prohibit the use of UDP port
> binding in the manner you suggest. It is an implementation option
> as to how UDP session binding is done for individual UDP based
> applications. For example, the approach you suggest was pursued by
> Activision folks for their game protocol design (Ref: Section 5.1,
> RC3027).

Yes, I realize that, but I am not a NAT implementor.  What I was looking for
was some terminology by which the needs of my application could be concisely
communicated to users and to NAT product vendors.  It would great if I could
say "this application will work if your NAT supports connection-less UDP port
binding, as defined in RFC 2663."  Then, the user could search for this in
their NAT documentation, or contact the NAT vendor.  I guess I can refer them
to "Activision protocol," but I really think this issue is more fundamental.

NAT would not be nearly so disruptive to the development of new applications
if it simply engaged in address remapping, rather than prohibiting multiple
sessions per address (where I take "address" in this context to mean (IP
address,TU port)).  That's what really destroys the concept of an address.  If
an application wants only one session per address, it can open up new ports on
the local host for each new session.  Allowing multiple sessions per address
does not prevent the NAT from requiring internal initiation of new sessions.
It just means that it must maintain a list of acceptable sessions for each
local address, rather than assigning a new local address to each acceptable
session.  With either variety of NAT, it is possible for a host to be
triggered into initiating a session toward a malicious host.  The only
difference is that connection-oriented NAT prevents this from occurring when
the malicious host also cannot accept external initiations.

I am personally interested in UDP, where connectionless NAT allows 2 hosts to
establish a session by simultaneous inititations (since the NAT cannot
distinguish a session initiation from a reply).  In the case of TCP, a NAT can
distinguish a reply from an initiation, but I wonder why it should care?  If I
send a SYN to a certain address, isn't that evidence that I would like a TCP
session with that end-point, so why should the NAT care whether it is an
outgoing or an incoming SYN that actually succeeds in establishing a session?
I guess there is some risk that an incoming SYN could be from a source address
that is spoofed to match a recent outgoing SYN destination.  I'm not sure this
is a major threat, but I am no security expert.

Again, I am not suggesting that NAT's should be mandated to maintain 1:1
address mappings (though my personal preferences are obviously evident).  I am
just asking for some terminology by which application developers, application
users, and NAT developers might communicate about this issue.

Tim







-
To unsubscribe, email 'majordomo@portmasters.com' with
'unsubscribe nat' in the body of the message.
List archive: <URL:http://www.portmasters.com/archives/>


