From exim@www1.ietf.org  Thu Aug 21 03:55:31 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA21652
	for <pmtud-archive@odin.ietf.org>; Thu, 21 Aug 2003 03:55:30 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pkHg-0001fq-VP
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 03:55:05 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7L7t4k5006435
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 03:55:04 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pkHg-0001fd-Nz
	for pmtud-web-archive@optimus.ietf.org; Thu, 21 Aug 2003 03:55:04 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA21621
	for <pmtud-web-archive@ietf.org>; Thu, 21 Aug 2003 03:55:00 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19pkHe-0007ZL-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 03:55:02 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19pkHd-0007ZH-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 03:55:01 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pkHe-0001ej-MD; Thu, 21 Aug 2003 03:55:02 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pkHS-0001bR-9K
	for pmtud@optimus.ietf.org; Thu, 21 Aug 2003 03:54:50 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA21605
	for <pmtud@ietf.org>; Thu, 21 Aug 2003 03:54:45 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19pkHP-0007Yv-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 03:54:47 -0400
Received: from lmr1.uibk.ac.at ([138.232.1.142] helo=smtp.uibk.ac.at)
	by ietf-mx with esmtp (Exim 4.12)
	id 19pkHO-0007Yj-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 03:54:47 -0400
Received: from lap10-c703.uibk.ac.at (lap10-c703.uibk.ac.at [138.232.65.57])
	by smtp.uibk.ac.at (8.12.9/8.12.9/F1) with ESMTP id h7L7sBWj005432;
	Thu, 21 Aug 2003 09:54:11 +0200
From: Michael Welzl <michael.welzl@uibk.ac.at>
To: pmtud@ietf.org, mtu@psc.edu
Content-Type: text/plain
Organization: University of Innsbruck
Message-Id: <1061452416.1733.104.camel@lap10-c703>
Mime-Version: 1.0
X-Mailer: Ximian Evolution 1.2.2 (1.2.2-4) 
Content-Transfer-Encoding: 7bit
X-Spam-Score: -5.4 () RCV_UIBK,USER_AGENT_XIMIAN
X-Scanned-By: MIMEDefang 2.35 at uibk.ac.at
Content-Transfer-Encoding: 7bit
Subject: [pmtud] MTU IP Option (reviving RFC 1063)
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: 21 Aug 2003 09:53:37 +0200
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit

Hi all,

I have a proposal for PMTUD which could be integrated
in the "new" as well as the "old" approach. The idea
is to revive RFC 1063 (the MTU Discovery Option, which
requires a router to update the MTU value in the option
if its MTU is smaller than the value already in there).
Back when this option was conceived, there was no
PMTUD - so PMTUD was clearly a better idea because the
option would require all routers to participate.

My idea is to combine the two. Routers MAY participate,
hosts SHOULD initially use the option when doing PTMUD.
If you receive something in the option, the Path MTU is
certainly not larger than the received value. If applied
properly, you have a method of PMTUD where a packet is
delayed (slow path processing), but, with some luck, you
lose no packets:
* With regular PMTUD, you use it before actually doing
  PMTUD and use the value as a starting point.
* With the new PMTUD, if you receive something, you
  should not increase the MTU beyond this point.

Is it a good idea?

Router perspective:
===================
Normally, routers send ICMP error messages back if
the MTU is exceeded. Router X which has a very small
MTU must generate a large number of such packets. Once
this new option is deployed, participating in the
mechanism can help router X to reduce its processing
overhead because option processing is certainly less
costly than generating an ICMP packet.

Source perspective:
===================
Is this worth having a packet delayed via slow path
processing? I think so, because it's a delayed vs. a
lost packet (well ... delayed in any case, vs. a chance
of not losing a packet). A precise answer requires
finding out the impact of slow path processing on
TCP performance. The first part of this sentence
(finding out the impact of slow path processing)
has already been done at our site; we are now
beginning a statistical analysis of our measured data.

What do you think?

Cheers,
Michael


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



From exim@www1.ietf.org  Thu Aug 21 05:39:28 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA25086
	for <pmtud-archive@odin.ietf.org>; Thu, 21 Aug 2003 05:39:28 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pluJ-0006BG-So
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 05:39:04 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7L9d3YD023752
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 05:39:03 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pluJ-0006B1-N0
	for pmtud-web-archive@optimus.ietf.org; Thu, 21 Aug 2003 05:39:03 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA25072
	for <pmtud-web-archive@ietf.org>; Thu, 21 Aug 2003 05:38:57 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19pluG-0000lR-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 05:39:00 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19pluF-0000lO-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 05:38:59 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pluH-0006Aa-LT; Thu, 21 Aug 2003 05:39:01 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pltm-00069z-Qj
	for pmtud@optimus.ietf.org; Thu, 21 Aug 2003 05:38:30 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA25043
	for <pmtud@ietf.org>; Thu, 21 Aug 2003 05:38:24 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19pltj-0000kd-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 05:38:27 -0400
Received: from brmea-mail-2.sun.com ([192.18.98.43])
	by ietf-mx with esmtp (Exim 4.12)
	id 19plti-0000kX-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 05:38:26 -0400
Received: from bu-ewat02-01.uk.sun.com ([129.156.199.2])
	by brmea-mail-2.sun.com (8.12.9/8.12.9) with ESMTP id h7L9ZdAh022557;
	Thu, 21 Aug 2003 03:35:40 -0600 (MDT)
Received: from sun.com (watford-109.UK.Sun.COM [129.156.199.109])
	by bu-ewat02-01.uk.sun.com (8.12.9+Sun/8.12.2/ENSMAIL,v2.2) with ESMTP id h7L9Zavm015374;
	Thu, 21 Aug 2003 10:35:38 +0100 (BST)
Message-ID: <3F449268.8020908@sun.com>
From: Jeremy Harris <jeremy.harris@sun.com>
User-Agent: Mozilla/5.0 (X11; U; SunOS sun4u; en-US; rv:1.3) Gecko/20030314
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Michael Welzl <michael.welzl@uibk.ac.at>
CC: pmtud@ietf.org, mtu@psc.edu
Subject: Re: [pmtud] MTU IP Option (reviving RFC 1063)
References: <1061452416.1733.104.camel@lap10-c703>
In-Reply-To: <1061452416.1733.104.camel@lap10-c703>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Thu, 21 Aug 2003 10:35:36 +0100
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit

Michael Welzl wrote:
> * With the new PMTUD, if you receive something, you
>   should not increase the MTU beyond this point.

Never?  What about route changes during a connection?
-Jeremy


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



From exim@www1.ietf.org  Thu Aug 21 05:51:27 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA25415
	for <pmtud-archive@odin.ietf.org>; Thu, 21 Aug 2003 05:51:27 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pm5u-0006fj-TN
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 05:51:02 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7L9p2Nv025641
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 05:51:02 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pm5u-0006fQ-Me
	for pmtud-web-archive@optimus.ietf.org; Thu, 21 Aug 2003 05:51:02 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA25407
	for <pmtud-web-archive@ietf.org>; Thu, 21 Aug 2003 05:50:56 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19pm5r-0000rB-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 05:50:59 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19pm5q-0000r8-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 05:50:58 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pm5t-0006eo-G5; Thu, 21 Aug 2003 05:51:01 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pm5M-0006cu-0H
	for pmtud@optimus.ietf.org; Thu, 21 Aug 2003 05:50:28 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA25384
	for <pmtud@ietf.org>; Thu, 21 Aug 2003 05:50:21 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19pm5I-0000qY-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 05:50:24 -0400
Received: from lmr1.uibk.ac.at ([138.232.1.142] helo=smtp.uibk.ac.at)
	by ietf-mx with esmtp (Exim 4.12)
	id 19pm5H-0000qF-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 05:50:23 -0400
Received: from lap10-c703.uibk.ac.at (lap10-c703.uibk.ac.at [138.232.65.57])
	by smtp.uibk.ac.at (8.12.9/8.12.9/F1) with ESMTP id h7L9nnWj017736;
	Thu, 21 Aug 2003 11:49:49 +0200
Subject: Re: [pmtud] MTU IP Option (reviving RFC 1063)
From: Michael Welzl <michael.welzl@uibk.ac.at>
To: Jeremy Harris <jeremy.harris@sun.com>
Cc: pmtud@ietf.org, mtu@psc.edu
In-Reply-To: <3F449268.8020908@sun.com>
References: <1061452416.1733.104.camel@lap10-c703>
	 <3F449268.8020908@sun.com>
Content-Type: text/plain
Organization: University of Innsbruck
Message-Id: <1061459354.1937.113.camel@lap10-c703>
Mime-Version: 1.0
X-Mailer: Ximian Evolution 1.2.2 (1.2.2-4) 
Content-Transfer-Encoding: 7bit
X-Spam-Score: -7.4 () EMAIL_ATTRIBUTION,IN_REP_TO,RCV_UIBK,REFERENCES,REPLY_WITH_QUOTES,USER_AGENT_XIMIAN
X-Scanned-By: MIMEDefang 2.35 at uibk.ac.at
Content-Transfer-Encoding: 7bit
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: 21 Aug 2003 11:49:14 +0200
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit

Hi,

Route Changes make no difference. The whole idea is to use
the IP Option when doing Path MTU Discovery, which implicitly
assumes that the path remains the same during the process.

Since a path may change indeed, the MTU should be increased
every once in a while (this is specified in RFC 1191).
So, if you generally use the MTU option for PMTUD, you
use it once again - simply each time you probe for the
Path MTU. The option must be seen in the context of a
single PMTUD process, no matter if it is by means of
increasing or decreasing the packet size until the right
one is found.

Cheers,
Michael

On Thu, 2003-08-21 at 11:35, Jeremy Harris wrote:
> Michael Welzl wrote:
> > * With the new PMTUD, if you receive something, you
> >   should not increase the MTU beyond this point.
> 
> Never?  What about route changes during a connection?
> -Jeremy


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



From exim@www1.ietf.org  Thu Aug 21 12:12:30 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA12236
	for <pmtud-archive@odin.ietf.org>; Thu, 21 Aug 2003 12:12:30 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19ps2f-0006Ok-2H
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 12:12:05 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7LGC5xm024590
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 12:12:05 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19ps2e-0006OX-Ux
	for pmtud-web-archive@optimus.ietf.org; Thu, 21 Aug 2003 12:12:04 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA12223
	for <pmtud-web-archive@ietf.org>; Thu, 21 Aug 2003 12:11:59 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19ps2d-0005BN-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 12:12:03 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19ps2d-0005BJ-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 12:12:03 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19ps2a-0006OC-Is; Thu, 21 Aug 2003 12:12:00 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19ps2M-0006Ns-4a
	for pmtud@optimus.ietf.org; Thu, 21 Aug 2003 12:11:46 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA12210
	for <pmtud@ietf.org>; Thu, 21 Aug 2003 12:11:40 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19ps2K-0005B5-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 12:11:44 -0400
Received: from oetest.freeswan.org ([205.150.200.166] helo=noxmail.sandelman.ottawa.on.ca)
	by ietf-mx with esmtp (Exim 4.12)
	id 19ps2I-0005Au-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 12:11:42 -0400
Received: from sandelman.ottawa.on.ca (marajade.sandelman.ottawa.on.ca [205.150.200.247])
	by noxmail.sandelman.ottawa.on.ca (8.11.6p2/8.11.6) with ESMTP id h7LGB9d06084
	(using TLSv1/SSLv3 with cipher EDH-RSA-DES-CBC3-SHA (168 bits) verified OK);
	Thu, 21 Aug 2003 12:11:15 -0400 (EDT)
Received: from marajade.sandelman.ottawa.on.ca (mcr@localhost)
	by sandelman.ottawa.on.ca (8.12.3/8.12.3/Debian -4) with ESMTP id h7LGB7T0002948;
	Thu, 21 Aug 2003 12:11:07 -0400
To: pmtud@ietf.org, mtu@psc.edu
Subject: Re: [pmtud] MTU IP Option (reviving RFC 1063) 
In-reply-to: Your message of "21 Aug 2003 09:53:37 +0200."
             <1061452416.1733.104.camel@lap10-c703> 
Mime-Version: 1.0 (generated by tm-edit 1.8)
Content-Type: text/plain; charset=US-ASCII
Message-ID: <2946.1061482267@marajade.sandelman.ottawa.on.ca>
From: Michael Richardson <mcr@sandelman.ottawa.on.ca>
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Thu, 21 Aug 2003 12:11:07 -0400

-----BEGIN PGP SIGNED MESSAGE-----


>>>>> "Michael" == Michael Welzl <michael.welzl@uibk.ac.at> writes:
    Michael> Is it a good idea?

    Michael> Router perspective: =================== Normally, routers send
    Michael> ICMP error messages back if the MTU is exceeded. Router X which
    Michael> has a very small MTU must generate a large number of such
    Michael> packets. Once this new option is deployed, participating in the

  The routers who are causing the MTU constraint clearly have some
obligation, and as an IPsec implementor, I'm happy to update my code like
this.
  So, this is a good idea. It may not be sufficient to deal with the problem.

]      Out and about in Ottawa.    hmmm... beer.                |  firewalls  [
]   Michael Richardson, Sandelman Software Works, Ottawa, ON    |net architect[
] mcr@sandelman.ottawa.on.ca http://www.sandelman.ottawa.on.ca/ |device driver[
] panic("Just another Debian/notebook using, kernel hacking, security guy");  [

  
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)
Comment: Finger me for keys - custom hacks make this fully PGP2 compat

iQCVAwUBP0TvGoqHRg3pndX9AQHPzAP+M82eZ9WOz6CmnGdwG2gQ2rWru+X8o2q4
MYtTBAt/wUmnVmv6kxB/rGfwpeOI03nWfaCr0hatQS/sbrZGDTAWq0iSH5XPezFh
rmXPk7kbElw1CNONwUUQVrL1vBt9GzydjneZt0ZN3awhlragbZ0fIFZpINdUjYy2
23vN6/LO+eA=
=8rTo
-----END PGP SIGNATURE-----

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



From exim@www1.ietf.org  Thu Aug 21 12:26:29 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA13023
	for <pmtud-archive@odin.ietf.org>; Thu, 21 Aug 2003 12:26:29 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19psGC-00074h-Kj
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 12:26:04 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7LGQ4OU027189
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 12:26:04 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19psGC-000749-FI
	for pmtud-web-archive@optimus.ietf.org; Thu, 21 Aug 2003 12:26:04 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA12983
	for <pmtud-web-archive@ietf.org>; Thu, 21 Aug 2003 12:25:57 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19psG9-0005OK-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 12:26:01 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19psG9-0005OH-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 12:26:01 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19psGA-00073S-0l; Thu, 21 Aug 2003 12:26:02 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19psG5-00072Z-9F
	for pmtud@optimus.ietf.org; Thu, 21 Aug 2003 12:25:57 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA12974
	for <pmtud@ietf.org>; Thu, 21 Aug 2003 12:25:51 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19psG3-0005Ny-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 12:25:55 -0400
Received: from viefep13-int.chello.at ([213.46.255.15])
	by ietf-mx with esmtp (Exim 4.12)
	id 19psG2-0005Md-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 12:25:54 -0400
Received: from fun ([80.109.142.58]) by viefep13-int.chello.at
          (InterMail vM.5.01.05.17 201-253-122-126-117-20021021) with SMTP
          id <20030821162521.FFNM17577.viefep13-int.chello.at@fun>;
          Thu, 21 Aug 2003 18:25:21 +0200
Message-ID: <000501c36801$404cb210$0200a8c0@fun>
From: "Michael Welzl" <michael.welzl@uibk.ac.at>
To: <pmtud@ietf.org>, <mtu@psc.edu>,
        "Michael Richardson" <mcr@sandelman.ottawa.on.ca>
References: <2946.1061482267@marajade.sandelman.ottawa.on.ca>
Subject: Re: [pmtud] MTU IP Option (reviving RFC 1063) 
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 6.00.2800.1106
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106
Content-Transfer-Encoding: 7bit
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Thu, 21 Aug 2003 18:28:27 +0200
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit

>   The routers who are causing the MTU constraint clearly have some
> obligation, and as an IPsec implementor, I'm happy to update my code like
> this.
>   So, this is a good idea. It may not be sufficient to deal with the
problem.

Let me make it clear to everybody that this is by no means intended
to be a sufficient solution to the problem, or to replace the original idea
of Matt (which I think is a very good one). It is an additional extension
which might give us an extra bonus (less loss, if lucky).

Cheers,
Michael



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



From exim@www1.ietf.org  Thu Aug 21 18:07:29 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA03702
	for <pmtud-archive@odin.ietf.org>; Thu, 21 Aug 2003 18:07:29 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pxaC-0004MW-Lv
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 18:07:05 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7LM73V7016761
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 18:07:03 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pxaB-0004MG-DU
	for pmtud-web-archive@optimus.ietf.org; Thu, 21 Aug 2003 18:07:03 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA03635
	for <pmtud-web-archive@ietf.org>; Thu, 21 Aug 2003 18:06:56 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19pxa8-0002XK-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 18:07:00 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19pxa7-0002XH-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 18:06:59 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pxa9-0004Ko-0T; Thu, 21 Aug 2003 18:07:01 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pxZX-0004CA-I8
	for pmtud@optimus.ietf.org; Thu, 21 Aug 2003 18:06:23 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA03500
	for <pmtud@ietf.org>; Thu, 21 Aug 2003 18:06:16 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19pxZU-0002Vw-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 18:06:20 -0400
Received: from basie.internet2.edu ([207.75.164.22])
	by ietf-mx with esmtp (Exim 4.12)
	id 19pxZT-0002VK-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 18:06:19 -0400
Received: from localhost (localhost.localdomain [127.0.0.1])
	by basie.internet2.edu (Postfix) with ESMTP
	id 781077B504; Thu, 21 Aug 2003 18:05:48 -0400 (EDT)
Received: from wlan238.internet2.edu (slip-32-101-183-168.mi.us.prserv.net [32.101.183.168])
	by basie.internet2.edu (Postfix) with ESMTP
	id 91A267B500; Thu, 21 Aug 2003 18:05:41 -0400 (EDT)
From: Matthew J Zekauskas <matt@internet2.edu>
To: pmtud@ietf.org
Cc: Matt Zekauskas <matt@internet2.edu>, Matt Mathis <mathis@psc.edu>
Message-ID: <27046290.1061489137@localhost>
X-Mailer: Mulberry/2.2.1 (Win32)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
X-Virus-Scanned: by mail.internet2.edu virus scanner
Content-Transfer-Encoding: 7bit
Subject: [pmtud] pmtud ietf57 vienna minutes, version 0
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Thu, 21 Aug 2003 18:05:37 -0400
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit

Due to a misunderstanding on my part, these are a little late...
Let us know if you have comments (stuff with XXX, in particular,
indicates areas where none of the note-takers captured the
conversation).  I expect these were actually due last week,
although I haven't had any of the usual reminder mail from
the proceedings editor; reading sooner rather than later would
be helpful, however.)

Matt M's original slides are here
http://www.psc.edu/~mathis/papers/pmtud200307/

A hacked version of the pdf file to try and get a clearer xplot
slide is here:
<http://people.internet2.edu/~matt/pmtud/meetings/ietf57/pmtud200307-mz.pdf>
==================================================================
Path Maximum Transmission Unit Discovery Pre-WG (pmtud)
Thursday, 17-July-2003 from 13:00 to 15:00
=======================================================

The meeting was chaired by Matt Mathis and Matt Zekauskas.
Al Morton, Itojun Hagino, and Matt Z. took notes, which were
assembled into these minutes by the chairs.

Agenda
------
 * Preliminaries: Blue sheets, Note takers, etc
 * WG Status
 * Short history and work to date
 * Robustness Issues
 * Other Stakeholders
 * Plans

Matt Mathis led off the meeting presenting the new co-chair, the
agenda, the changes to the proposed charter, and the aggressive
milestones.

The group status is that some parts of the administrative preparation
did not get done, but IESG has approved, hence "Pre WG".  This will be
a fast development, silence will be acceptance (at the start as sections
are integrated).  PMTUD is a re-activation of path MTU WG, which was
a very similar effort.  To participate, you must subscribe to IETF list
(pmtud@ietf.org).  Matt Zekauskas volunteered to Co-chair the group.
The charter was broadened so as not to restrict to a single method.
Milestones are aggressive, and need for implementation to test is clear.

No one disagreed, nor were there suggestions for other methods to
study.  Matt Mathis will make another editorial pass at the document
(which was a rewrite, starting from RFC 1981, instead of an update to
the previous BOF input).  Sections will be added based on mailing list
comments and any input from stakeholder communities.

Matt Mathis sketched the previous algorithm and noted some of the
problems.  He then sketched the new algorithm, noting that there
is just a small amount of MUST/SHOULD language: under what circumstances
can losses be ignored as a congestion signal.  The rest is heuristics;
it doesn't need to be the same for every application, and permits
vendor diversity.

Question: when can the algorithm be used by TCP?  Just after 3-way
handshake, or before real communication?  Matt responded that it
uses live payload data, and the draft has a recommendation not
to attempt the algorithm unless the congestion window is at least
twenty packets, so the connection is well established before the
algorithm starts.  Thus, this could slow down tiny files -- the
exact algorithm is a heuristic, so you could choose to perform
it differently.  There are tradeoffs.

The request for collaborators in the IPsec & security area led to a
big discussion on tunnel issues.  People were positive about the
method, but there are corner cases to consider.  Input has been
promised to the mailing list.  (In the IPv4 world, lack of PMTUD is
noted as a major problem with IPsec VPNs and providing services.)

This started around slide 11, "Plans for the Next Draft".  One
of the collaborators was folks from the multicast area; one possibility
is a generalization of the algorithm for reliable delivery.
This would solve an ICMP implosion problem if the current MTU
discovery technique was used.  Someone noted that the behavior
as specified in IPv4 and IPv6 was different, in  6 you respond,
in v4 you don't. [XXX]

Another collaborating group would be IPsec; currently the security
architecture document has major sections dealing with the interaction
of MTU discovery and IPsec (because tunnels are created); the new
technique might obsolete many of those sections.

One audience member [Itojun, I think -XXX] noted that the interaction
between IPsec and and TCP depends on if the TCP stack is aware of IPsec.
If the TCP stack does not take care of the IPsec header size, the
algorithm would need to be revised.

Matt M. responded that the detail in the draft needs to be resolved
in a consistent way.  You can count the IPsec header as part of the
IP header or TCP header.  The really nasty cases involve additional
layers, for example IPsec on a VPN, ICMP messages could go back to
the wrong place.

Michael Richardson expanded on this as an IPsec implementer.  The
worst case is common at meetings such as this -- you have a corporate
address on your laptop, and a VPN back to the corporate space, so
all traffic goes back to HQ.  Try to visit a bank, and they have
ICMP filters.  Your gateway is sending out ICMP messages to the
bank, and they drop them.  This proposed algorithm should work
really well.  Many times VPNs are blamed, when the problem is really
a bad ICMP filter.

However, there is a problem, if you raise the MTU and the tunnels do
not toss large messages but fragment them anyway, you will end up
always fragmenting.  Someone working on a Linux version mentioned that
he did not honor the DF bit; having a poor(er) performing
implementation was better than one that didn't work at all.  ("Poor
performance is better than no security.") Perhaps there could be a
heuristic that worked for a short term solution so these mechanisms
don't interact badly... the endpoint would need to be updated for this
algorithm, so IPsec tunneling could be updated at the same time.  This
behavior is often a kernel option, too.  Perhaps you could fragment
into tunnel, but retain the DF bit, and if set don't do anything
weird.  Itojun related KAME experience; there they ended up not
setting the DF bit on output header when IPsec tunnels are created.

Another point was that IPv6 (on IPv4?) tunnels have the same issue.
IPv6 tunnels should have a MTU of 1280 by default so a minimum MTU
can be maintained.

Matt M. mentioned that he's aware that a large number of tunneling
implementations don't copy the DF bit from inner packet to outer
header.  He's not yet sure if the document needs a specific section
covering tunnels and tunnel migration; an intermediate ground that
works is to let tunnels behave this way in the interim, and discourage
a mode where end systems ignore can't fragment messages.

Someone else pointed out that mobility might add additional headers;
so a tunnel MTU of 1280 might not be enough.

Itojun stated that he was thinking of configured tunnels and not mobile
IPv6.  If you send a packet with mobile headers, the TCP stack needs
to be aware of the size of the mobile IP headers and reduce MSS
appropriately -- maintain the total MTU size.

Another person agreed that 1280 is particularly bad, 1380 would be
better.

Matt Mathis noted that there was definitely a subgroup interested in
considering tunneling and MTU discovery; he encouraged folks to
join the mailing list and contribute the various circumstances where
there are potential problems.

Lars Eggert mentioned that RFC2003 specifies some things related
to MTU discovery, and RFC2401 specifically prohibits some of the
mechansims in RFC2003 for security reasons.

As to other transport protocols, Matt Z. reported that he had
quickly skimmed SCTP and DCCP documents, and that SCTP looked
possible, but DCCP says specifically that the MTU can't be raised.
No one that claimed to be an SCTP expert was in the room (or at
least didn't comment negatively on the applicability to SCTP).
Eddie Kohler noted that this behavior was revised in the
DCCP WG meeting this week.   Matt Z. prompted him to send some DCCP text.

Matt M. emphasized that the point with getting a draft done early
is to encourage implementation as soon as possible.  The algorithm
will use specific details of other protocols, and we're dependent
on the uniformity of implementation of certain features.  We need
to learn what implementations really do; ideally get a custom
implementation run on servers and real field data to feed back
into the document.

Matt M also noted two cases that he's worried about (although
these are just examples; others are encouraged to consider other
cases, or report back implementation experience).  First, what
happens if a path is striped across multiple links, and the MTU
is not the same across the stripes?  You can require that the
MTU is not raised until a certain number of segments are received
successfully.  You need to understand the interaction between random
losses and whether the MTU is or is not raised.  Second, what happens
if there is a parametric failure -- when raising the MTU causes the
error rate to increase?  An actual case is one particular 10G gbic;
it was error free with 1500 byte packets, but not with 9000 byte
packets.  There is an opportiunity for different heuristics here,
for example use a smaller MSS if you cannot fill a window.

For hard, repeated, timeouts the first thing you want to do is
reduce congestion variables, then reduce MTU.  At some point want
to restart the checks to increase MTU.

There are other possible protocol interactions, too: for example,
SCTP can use multiple endpoints.  What if it changes addresses,
and the new path has a smaller MTU?

Michael felt that it was important to focus where the production
environments hurt most with the current MTU scheme.  Matt M noted
that different things hurt in different environments.  Michael
expanded that the most frequent case will likely be large port 80
responses to a client.  And it's the client that would decide that
the path is stupid or broken, and other A record should be tried.
The web server is getting the timeouts, not the client.  This won't
deploy if we can't solve the web server case.

Matt M noted another case he had thought about, but not seen: what
happens if raising the MTU causes link stability problems (as opposed
to hard failures) -- say the link "goes away" for 10 seconds and then
returns.  He's thought about using a state machine to catch this case...
the link is broken, and we don't necessarily want to fix it with
an MTU discovery algorithm.

On ignoring DF bits, so that a tunnel fragments large packets: Matt M
contended it was worse for a 1500 tunnel fragmenting a 9000 packet
than a tunnel fragmenting a 1500 packet by the tunnel overhead.
Michael didn't understand this at first; Matt explained that
the problem is that with many fragments the odds are greater that
you lose a fragment, and hence the whole packet than if there
are only two fragments.

In thinking about other stakeholders, one person felt that the
algorithm would work for RTP over UDP; there you have some
extensions to share information -- reliable instrumentation of
losses. [XXX]

Itojun said that we should contact rrs@cisco.com for SCTP.

In the multicast case, it was argued that this algorithm might cause
an ACK implosion that is worse than a ICMP-message-too-big implosion.

[something about DSS/MPP? bridges & MSS flapping hack that I didn't catch.
delete if you don't know anything about it.
mr:
dss/mpp bridges doing is MSS flapping hack.  maybe should write it down
and say please don't do it, do it instead
XXX]

Someone commented that the document as written is very TCP specific.
The algorithm should be better separated from actual deployment.
Matt M said that's the intention.

Another question was what, exactly, is the definition of MTU?  End-to-end
or link-specific.  Matt M said that we were talking about IP MTU
when using a particular link layer; how the IETF uses it, not what
hardware specifications say.

It was also argued that IPv6 might not need this at all; the algorithm
could arguably make performance worse (since the MSS size would ramp
up instead of being decided once by ICMP-message-too-big); However,
the new algorithm would prevent against implementation or
configuration bugs and also work in the cases where L2 MTUs were
different on a switch.

One audience member said that this should be documented in detail.
Since v6 has no DF bit, people that block ICMPv6 won't get connectivity,
so people won't block it.

Matt M noted that some stacks have IPv4 mimic IPv6 -- they always set
DF, even on fragments,  and attempt to fragment only at the endpoints.

However, there's no requirement that routers send the too big messages
in v4, but there is a requirement in v6.

Another comment was tha the MTU in Router Advertisement messages should
solve this problem.  If operational experience says this isn't happening,
it should be reflected to the v6 working group.

Matt M said that in all cases the problems with path MTU discovery
are bugs.  There are a large set of problems.



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



From exim@www1.ietf.org  Thu Aug 21 19:51:32 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA09701
	for <pmtud-archive@odin.ietf.org>; Thu, 21 Aug 2003 19:51:31 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pzCr-0000Af-Sw
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 19:51:06 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7LNp51L000653
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 19:51:05 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pzCr-0000AS-PK
	for pmtud-web-archive@optimus.ietf.org; Thu, 21 Aug 2003 19:51:05 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA09689
	for <pmtud-web-archive@ietf.org>; Thu, 21 Aug 2003 19:51:01 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19pzCq-0003m6-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 19:51:04 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19pzCp-0003m3-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 19:51:03 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pzCn-00009j-GA; Thu, 21 Aug 2003 19:51:01 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19pzCL-00007L-1B
	for pmtud@optimus.ietf.org; Thu, 21 Aug 2003 19:50:33 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA09673
	for <pmtud@ietf.org>; Thu, 21 Aug 2003 19:50:28 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19pzCJ-0003lu-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 19:50:31 -0400
Received: from oetest.freeswan.org ([205.150.200.166] helo=noxmail.sandelman.ottawa.on.ca)
	by ietf-mx with esmtp (Exim 4.12)
	id 19pzCH-0003lr-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 19:50:30 -0400
Received: from sandelman.ottawa.on.ca (marajade.sandelman.ottawa.on.ca [205.150.200.247])
	by noxmail.sandelman.ottawa.on.ca (8.11.6p2/8.11.6) with ESMTP id h7LNoOd08151
	(using TLSv1/SSLv3 with cipher EDH-RSA-DES-CBC3-SHA (168 bits) verified OK)
	for <pmtud@ietf.org>; Thu, 21 Aug 2003 19:50:30 -0400 (EDT)
Received: from marajade.sandelman.ottawa.on.ca (mcr@localhost)
	by sandelman.ottawa.on.ca (8.12.3/8.12.3/Debian -4) with ESMTP id h7LNoNF2013398
	for <pmtud@ietf.org>; Thu, 21 Aug 2003 19:50:23 -0400
To: pmtud@ietf.org
Subject: Re: [pmtud] pmtud ietf57 vienna minutes, version 0 
In-reply-to: Your message of "Thu, 21 Aug 2003 18:05:37 EDT."
             <27046290.1061489137@localhost> 
Mime-Version: 1.0 (generated by tm-edit 1.8)
Content-Type: text/plain; charset=US-ASCII
Message-ID: <13396.1061509823@marajade.sandelman.ottawa.on.ca>
From: Michael Richardson <mcr@sandelman.ottawa.on.ca>
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Thu, 21 Aug 2003 19:50:23 -0400

-----BEGIN PGP SIGNED MESSAGE-----


>>>>> "Matthew" == Matthew J Zekauskas <matt@internet2.edu> writes:
    Matthew> Michael Richardson expanded on this as an IPsec implementer.  The
    Matthew> worst case is common at meetings such as this -- you have a
    Matthew> corporate address on your laptop, and a VPN back to the
    Matthew> corporate space, so 
    Matthew> all traffic goes back to HQ.  Try to visit a bank, and they have
    Matthew> ICMP filters.  Your gateway is sending out ICMP messages to the
    Matthew> bank, and they drop them.  This proposed algorithm should work
    Matthew> really well.  Many times VPNs are blamed, when the problem is
    Matthew> really a bad ICMP filter.

  But, the key thing is that the "new" item is the VPN, not the ICMP filter.

    Matthew> However, there is a problem, if you raise the MTU and the
    Matthew> tunnels do 
    Matthew> not toss large messages but fragment them anyway, you will end up
    Matthew> always fragmenting.  Someone working on a Linux version
    Matthew> mentioned that 
    Matthew> he did not honor the DF bit; having a poor(er) performing
    Matthew> implementation was better than one that didn't work at all.

  That was me. I am the FreeS/WAN technical lead.
 
    Matthew> ("Poor performance is better than no security.") Perhaps there
    Matthew> could be a heuristic that worked for a short term solution so
    Matthew> these mechanisms don't interact badly... the endpoint would need
    Matthew> to be updated for this algorithm, so IPsec tunneling could be
    Matthew> updated at the same time.  This 

  The key point is that I, the IPsec developer can't control:
      1) the ICMP filter.
      2) the TCP on the machine behind it.

  I can have *some* influence on the TCP at the receiving end of the flow,
but not a lot. *If* the IPsec tunnel terminates on the same machine as the 
TCP, then in theory, the TCP can learn about the reduced MTU, and set the
MSS appropriately. In practice, probably only KAME, Microsoft and Sun are
well enough integrated to do this right now. Probably Linux 2.6 will be able
to do so as well. 
  We do have an option to hack the MSS on the encapsulator's side already,
alas. 
 
  So, I need a mechanism that I, as the guy who is about to send the ICMP
can do to help things along. 

http://www.freeswan.ca/cgi-bin/viewcvs.cgi/linux/net/ipsec/ipsec_xmit.c?rev=1.2&cvsroot=FreeS%2FWAN&content-type=text/vnd.viewcvs-markup

	/* 
	   If the sender is doing PMTU discovery, and the
	   packet doesn't fit within ixs->prv->mtu, notify him
	   (unless it was an ICMP packet, or it was not the
	   zero-offset packet) and send it anyways.

	   Note: buggy firewall configuration may prevent the
	   ICMP packet from getting back.
	*/
	if(sysctl_ipsec_icmp
	   && ixs->prv->mtu < ntohs(ixs->iph->tot_len)
	   && (ixs->iph->frag_off & __constant_htons(IP_DF)) ) {
		int notify = ixs->iph->protocol != IPPROTO_ICMP
			&& (ixs->iph->frag_off & __constant_htons(IP_OFFSET)) == 0;
			
#ifdef IPSEC_obey_DF
		KLIPS_PRINT(debug_tunnel & DB_TN_CROUT,
			    "klips_debug:ipsec_xmit_encap_bundle: "
			    "fragmentation needed and DF set; %sdropping packet\n",
			    notify ? "sending ICMP and " : "");
		if (notify)
			ICMP_SEND(ixs->skb,
				  ICMP_DEST_UNREACH,
				  ICMP_FRAG_NEEDED,
				  ixs->prv->mtu,
				  ixs->physdev);
		ixs->stats->tx_errors++;
		bundle_stat = IPSEC_XMIT_CANNOTFRAG;
		goto cleanup;
#else /* IPSEC_obey_DF */
		KLIPS_PRINT(debug_tunnel & DB_TN_CROUT,
			    "klips_debug:ipsec_xmit_encap_bundle: "
			    "fragmentation needed and DF set; %spassing packet\n",
			    notify ? "sending ICMP and " : "");
		if (notify)
			ICMP_SEND(ixs->skb,
				  ICMP_DEST_UNREACH,
				  ICMP_FRAG_NEEDED,
				  ixs->prv->mtu,
				  ixs->physdev);
#endif /* IPSEC_obey_DF */
	}


    Matthew> behavior is often a kernel option, too.  Perhaps you could
    Matthew> fragment into tunnel, but retain the DF bit, and if set don't do
    Matthew> anything weird.  Itojun related KAME experience; there they
    Matthew> ended up not 
    Matthew> setting the DF bit on output header when IPsec tunnels are
    Matthew> created. 

  This refers to the question of whether or not to set DF bit on resulting
packet. That's a problem, but a different problem.

  If we go with the proposed algorithm, that we send probes with DF set, and
if they don't make progress, assume that they were broken, then we MUST have
all tunnel protocols copy the DF bit to the outside.

  The problem with this is that the tunnel system doesn't really know if the
sender is using the new or old algorithm, so it really can't adapt.

    Matthew> Someone else pointed out that mobility might add additional
    Matthew> headers; 
    Matthew> so a tunnel MTU of 1280 might not be enough.

  1500-1280 = 220 bytes, which is 11 layers of IP/IP tunnels, and 5 layers
of IPsec. 4 layers of NAT-IPsec.

    Matthew> happens if a path is striped across multiple links, and the MTU
    Matthew> is not the same across the stripes?  You can require that the
    Matthew> MTU is not raised until a certain number of segments are received
    Matthew> successfully.  You need to understand the interaction between
    Matthew> random 
    Matthew> losses and whether the MTU is or is not raised.  Second, what

  I think that one has to have hysteris in the MTU discovery. If it keeps
bouncing, then leave it at the lower value longer.

    Matthew> On ignoring DF bits, so that a tunnel fragments large packets:
    Matthew> Matt M 
    Matthew> contended it was worse for a 1500 tunnel fragmenting a 9000 packet
    Matthew> than a tunnel fragmenting a 1500 packet by the tunnel overhead.
    Matthew> Michael didn't understand this at first; Matt explained that
    Matthew> the problem is that with many fragments the odds are greater that
    Matthew> you lose a fragment, and hence the whole packet than if there
    Matthew> are only two fragments.

  btw, I'm happy to learn more about the problems of dealing with 9000 byte
packets, I'll provide the gigE port, and my address, if you'll pay for a
connection to a GigE backbone :-)

    Matthew> Someone commented that the document as written is very TCP
    Matthew> specific. 

  Is there anything we can do about non-TCP?
  For instance, ESP is definitely *not-TCP*, but it isn't really a transport
protocol in that it doesn't have a backflow channel.

    Matthew> Matt M said that in all cases the problems with path MTU discovery
    Matthew> are bugs.  There are a large set of problems.

  I think that we won't make progress in getting the bugs fixed without a
good testbench. If there are Cisco PIX people around, their customers seem 
to be the most clueless about ICMP. Maybe that's because the PIX is way too
easy to use? :-)

]      Out and about in Ottawa.    hmmm... beer.                |  firewalls  [
]   Michael Richardson, Sandelman Software Works, Ottawa, ON    |net architect[
] mcr@sandelman.ottawa.on.ca http://www.sandelman.ottawa.on.ca/ |device driver[
] panic("Just another Debian/notebook using, kernel hacking, security guy");  [
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)
Comment: Finger me for keys - custom hacks make this fully PGP2 compat

iQCVAwUBP0VavoqHRg3pndX9AQEmbAP+N884rGkEPpFXiii9XAsyh5Ivl+/UoYGb
ml2MVbBpNRPcjP6D/UfkiCwqtNR3rmK90OeDkntJv8V/HFo6TV/qQW7VaKT7582A
qcgTQJeZbjMGRy+DfDZ/tggRwer/mFSP3icmGhl+/0hv3SHqe7wOdngE+JcPGPUe
FMoUkCtHVCc=
=3Y98
-----END PGP SIGNATURE-----

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



From exim@www1.ietf.org  Thu Aug 21 21:57:10 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA13847
	for <pmtud-archive@odin.ietf.org>; Thu, 21 Aug 2003 21:57:09 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q18u-0004WB-AE
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 21:55:16 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7M1t8IJ017361
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 21:55:08 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q18u-0004Vw-61
	for pmtud-web-archive@optimus.ietf.org; Thu, 21 Aug 2003 21:55:08 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA13789
	for <pmtud-web-archive@ietf.org>; Thu, 21 Aug 2003 21:55:02 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19q18r-0004lR-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 21:55:05 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19q18q-0004lO-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 21:55:04 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q18o-0004VO-DJ; Thu, 21 Aug 2003 21:55:02 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q18N-0004Uw-Qa
	for pmtud@optimus.ietf.org; Thu, 21 Aug 2003 21:54:35 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA13772
	for <pmtud@ietf.org>; Thu, 21 Aug 2003 21:54:29 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19q18K-0004lI-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 21:54:32 -0400
Received: from coconut.itojun.org ([219.101.47.130])
	by ietf-mx with esmtp (Exim 4.12)
	id 19q18J-0004lC-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 21:54:31 -0400
Received: by coconut.itojun.org (Postfix, from userid 1001)
	id 5B29199; Fri, 22 Aug 2003 10:54:29 +0900 (JST)
To: matt@internet2.edu
Cc: pmtud@ietf.org, mathis@psc.edu
Subject: Re: [pmtud] pmtud ietf57 vienna minutes, version 0
In-Reply-To: Your message of "Thu, 21 Aug 2003 18:05:37 -0400"
	<27046290.1061489137@localhost>
References: <27046290.1061489137@localhost>
X-Mailer: Cue version 0.6 (030716-1832/itojun)
Mime-Version: 1.0
Content-Type: Text/Plain; charset=us-ascii
Message-Id: <20030822015429.5B29199@coconut.itojun.org>
From: itojun@itojun.org (Jun-ichiro itojun Hagino)
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Fri, 22 Aug 2003 10:54:29 +0900 (JST)

> One audience member [Itojun, I think -XXX] noted that the interaction
> between IPsec and and TCP depends on if the TCP stack is aware of IPsec.
> If the TCP stack does not take care of the IPsec header size, the
> algorithm would need to be revised.

	it's me i guess.

> However, there is a problem, if you raise the MTU and the tunnels do
> not toss large messages but fragment them anyway, you will end up
> always fragmenting.  Someone working on a Linux version mentioned that
> he did not honor the DF bit; having a poor(er) performing
> implementation was better than one that didn't work at all.  ("Poor
> performance is better than no security.") Perhaps there could be a
> heuristic that worked for a short term solution so these mechanisms
> don't interact badly... the endpoint would need to be updated for this
> algorithm, so IPsec tunneling could be updated at the same time.  This
> behavior is often a kernel option, too.  Perhaps you could fragment
> into tunnel, but retain the DF bit, and if set don't do anything
> weird.  Itojun related KAME experience; there they ended up not
> setting the DF bit on output header when IPsec tunnels are created.

	(sorry, this is not a comment to the minute itself)
	draft-ietf-v6ops-mech-v2-00.txt provides two choices for tunnels:
	- tunnel ingress router doesn't set outer DF bit set.  tunnel
	  ingress/egress router does not need to keep track of MTU on the
	  tunnel path.  tunnel MTU is static.  simpler to implement,
	  no wacky translation from inner ICMP message to outer ICMP message.
	- tunnel ingress router keeps track of tunnel path MTU (= outer DF bit
	  is set).  tunnel MTU is dynamic.  wacky translation from inner ICMP
	  message to outer ICMP message needed.

> Someone else pointed out that mobility might add additional headers;
> so a tunnel MTU of 1280 might not be enough.

> Another person agreed that 1280 is particularly bad, 1380 would be
> better.

	this two comments were from Dave Thaler @ microsoft, IIRC.

itojun

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



From exim@www1.ietf.org  Thu Aug 21 21:58:27 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA13900
	for <pmtud-archive@odin.ietf.org>; Thu, 21 Aug 2003 21:58:27 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q1Bi-0004ZX-N8
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 21:58:02 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7M1w2ro017569
	for pmtud-archive@odin.ietf.org; Thu, 21 Aug 2003 21:58:02 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q1Bi-0004ZE-GA
	for pmtud-web-archive@optimus.ietf.org; Thu, 21 Aug 2003 21:58:02 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA13882
	for <pmtud-web-archive@ietf.org>; Thu, 21 Aug 2003 21:57:56 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19q1Bf-0004mp-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 21:57:59 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19q1Be-0004mm-00
	for pmtud-web-archive@ietf.org; Thu, 21 Aug 2003 21:57:58 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q1Bg-0004YE-Qn; Thu, 21 Aug 2003 21:58:00 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q1BN-0004Xa-8v
	for pmtud@optimus.ietf.org; Thu, 21 Aug 2003 21:57:41 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA13867
	for <pmtud@ietf.org>; Thu, 21 Aug 2003 21:57:35 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19q1BK-0004mZ-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 21:57:38 -0400
Received: from oetest.freeswan.org ([205.150.200.166] helo=noxmail.sandelman.ottawa.on.ca)
	by ietf-mx with esmtp (Exim 4.12)
	id 19q1BJ-0004mW-00
	for pmtud@ietf.org; Thu, 21 Aug 2003 21:57:37 -0400
Received: from sandelman.ottawa.on.ca (marajade.sandelman.ottawa.on.ca [205.150.200.247])
	by noxmail.sandelman.ottawa.on.ca (8.11.6p2/8.11.6) with ESMTP id h7M1vVd08630
	(using TLSv1/SSLv3 with cipher EDH-RSA-DES-CBC3-SHA (168 bits) verified OK)
	for <pmtud@ietf.org>; Thu, 21 Aug 2003 21:57:37 -0400 (EDT)
Received: from marajade.sandelman.ottawa.on.ca (mcr@localhost)
	by sandelman.ottawa.on.ca (8.12.3/8.12.3/Debian -4) with ESMTP id h7M1vVew016052
	for <pmtud@ietf.org>; Thu, 21 Aug 2003 21:57:31 -0400
To: pmtud@ietf.org
In-reply-to: Your message of "Thu, 21 Aug 2003 18:05:37 EDT."
             <27046290.1061489137@localhost> 
Mime-Version: 1.0 (generated by tm-edit 1.8)
Content-Type: text/plain; charset=US-ASCII
Message-ID: <16050.1061517451@marajade.sandelman.ottawa.on.ca>
From: Michael Richardson <mcr@sandelman.ottawa.on.ca>
Subject: [pmtud] the DF-bit honouring problem
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Thu, 21 Aug 2003 21:57:31 -0400

-----BEGIN PGP SIGNED MESSAGE-----


As we discussed at the meeting last month, the plpmtud proposal breaks
when there are tunnelling systems that do not obey the DF bit. A second
order effect is when tunnelling systems do not copy the DF bit to the
outer header when they encapsulate.

The problem is that the DF-ignoring code is there for a reason - because
otherwise the network breaks.  Unfortunately, any proposal that requires
that the sending end-point upgrade fails - in 99% of the situations
where classic PMTUD is broken, it is due to actions that the administrators
of that end point have done.  Getting the end point to upgrade to plpmtud
is nice, and perhaps some end points will get upgraded as part of regular
OS-upgrades, but I doubt any will be motivated to upgrade just for that
"feature".

If the IETF wants to do something at the political layer, then there is
plenty that we can do in the document and then embarass route. A la ECN.
(Which still doesn't work everywhere, btw)

My impression is that many of the problems seem to be caused by a single
popular product.  I could be wrong here - but I've pursued this problem with
multiple organizations. Some of them, for instance, called me to complain
that I was attacking them with ICMP messages! This has happened multiple
times. Some complained about ICMP Fragneeded, others about
port-unreachable. {In each case they had a comprised mail server being
used to send spam... which generally involves large packets)

The remainder is by ISPs who have filtered out all of ICMP because they have,
e.g. DSL-aggregators from dead companies that are vulnerable to
directed-broadcast ICMP echo requests.   

I believe that most of the problems occur between web clients and web
servers, with the large packets going from server to client.  

Ideally, we would be able to document a change to TCP that would work if
only the receiver was upgraded. The receiver generally wants to receive the
information (the web page), and so has some motivation to fix things.

The only tool that the recipient has right now is setting a lower MSS
when initiating the connection. We aren't allowed to use that option again
later on. That's a shame, because it would otherwise provide an easy way
for a receiver to possibly recover. The receiver could run an algorithm 
which permitted the MSS to grow.

- --- back to DF-ignoring

The risk to pmtumtud of DF-ignoring systems is that they cause the 
algorithm to fail, since all probes would get through. The result is more
fragments, and possible much higher changes of failure to communicate.

But please understand that "fixing" the DF-ignoring systems causes 100% 
communication failure. 

=== variant proposal

I suggest that in addition to the process described in plmtud, that we
also define a new TCP option which the system using the algorithm would
add to its packet with the SYN set. In the web client/server situation,
this means that the web server adds this when it is upgraded. We have learnt
with ECN, that adding things on the client is just a loss with same above
mentioned popular product.

We also create a new TCP option to set a new MSS or relax re-use of
TCP MSS option to permit it to be used at any time.

The receiver (web client), if it is aware of pmtud-prime, seeing that option,
would observe the maximal size of fragments received (if any). It would be
particularly interested in any that reassemble with the DF bit set. The 
receiver would then use the new TCP option to let the sender know about the
maximal MSS.

The option part of the algorithm is specific to TCP. SCTP would need
something similar. UDP based protocols would have to do their own thing.
(NFS/UDP for instance, might change the read size that it asks for).

]      Out and about in Ottawa.    hmmm... beer.                |  firewalls  [
]   Michael Richardson, Sandelman Software Works, Ottawa, ON    |net architect[
] mcr@sandelman.ottawa.on.ca http://www.sandelman.ottawa.on.ca/ |device driver[
] panic("Just another Debian/notebook using, kernel hacking, security guy");  [










-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)
Comment: Finger me for keys - custom hacks make this fully PGP2 compat

iQCVAwUBP0V4gYqHRg3pndX9AQGFnQP+OJV71M4evxlECHnRbLZCioGNJm+cWWpL
zbvEy6jDfRVhG4ANWqFgXI0a1/cZrz61ePWB922/pJQxcS2aqkMgLeLikfE3wFFS
BQ11WYGLuKLml2jJQDgq49FVGyfjUX46OaLQfHYwIMQgdHYDegvkkbyf5+Jtde5I
KoCjmIS1p2A=
=DCm2
-----END PGP SIGNATURE-----

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



From exim@www1.ietf.org  Fri Aug 22 04:46:28 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA11410
	for <pmtud-archive@odin.ietf.org>; Fri, 22 Aug 2003 04:46:28 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q7YY-0006oM-ND
	for pmtud-archive@odin.ietf.org; Fri, 22 Aug 2003 04:46:03 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7M8k2Rp026183
	for pmtud-archive@odin.ietf.org; Fri, 22 Aug 2003 04:46:02 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q7YY-0006oE-GN
	for pmtud-web-archive@optimus.ietf.org; Fri, 22 Aug 2003 04:46:02 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA11400
	for <pmtud-web-archive@ietf.org>; Fri, 22 Aug 2003 04:45:57 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19q7YV-0000t8-00
	for pmtud-web-archive@ietf.org; Fri, 22 Aug 2003 04:45:59 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19q7YV-0000t5-00
	for pmtud-web-archive@ietf.org; Fri, 22 Aug 2003 04:45:59 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q7YW-0006nh-AA; Fri, 22 Aug 2003 04:46:00 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q7YA-0006n9-8H
	for pmtud@optimus.ietf.org; Fri, 22 Aug 2003 04:45:38 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA11382
	for <pmtud@ietf.org>; Fri, 22 Aug 2003 04:45:33 -0400 (EDT)
From: Stephane.Antoine@panasonicmobile.co.uk
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19q7Y7-0000t0-00
	for pmtud@ietf.org; Fri, 22 Aug 2003 04:45:35 -0400
Received: from laurel.mci.co.uk ([195.144.128.50])
	by ietf-mx with esmtp (Exim 4.12)
	id 19q7Y6-0000sx-00
	for pmtud@ietf.org; Fri, 22 Aug 2003 04:45:34 -0400
Message-ID: <H000025e0096c10a.1061541919.contactserver1@MHS>
Subject: RE: [pmtud] pmtud ietf57 vienna minutes, version 0
MIME-Version: 1.0
To: itojun@itojun.org, matt@internet2.edu
cc: mathis@psc.edu, pmtud@ietf.org
X-WSS-ID: 135B079638985-01-01
Content-Type: text/plain;
 charset=us-ascii
Content-Disposition: inline;
 Creation-Date="Fri, 22 Aug 2003 09:45:20 +0100"
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Fri, 22 Aug 2003 09:45:20 +0100
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit



> Someone else pointed out that mobility might add additional headers;
> so a tunnel MTU of 1280 might not be enough.

=>However a fix MTU of 1280 bytes
does not prevent you from designing (dynamic) path MTUD algorithms.
In fact the presence of mobility header or security headers above a
static IPv6 tunnel will definitely cause fragmentation if the TCP 
segment 
size is not reduced appropriately....


> Another person agreed that 1280 is particularly bad, 1380 would be
> better.
	this two comments were from Dave Thaler @ microsoft, IIRC.

itojun

=>I don't understand why 1380 bytes would be much better than 1280?


stephane 


==============================================================================
Confidentiality Notice

The information contained in this E-mail, and any attachments, is intended for
the named recipients only. It may contain confidential and/or privileged
information.

If you are not the intended recipient, you must not copy, distribute, or take
any action in reliance on it. Any views expressed do not necessarily reflect
the views of Panasonic Mobile Communications Development of Europe Limited.

Please note that whilst the company takes steps to protect against viruses it
cannot accept any liability for any damage, whether direct or indirect,
sustained as a result of any software viruses being transmitted.

If you receive this E-mail by mistake, please advise the sender by using the
reply facility in your E-mail software and then delete it.
==============================================================================


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



From exim@www1.ietf.org  Fri Aug 22 05:22:28 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA12683
	for <pmtud-archive@odin.ietf.org>; Fri, 22 Aug 2003 05:22:28 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q87P-00088n-Tz
	for pmtud-archive@odin.ietf.org; Fri, 22 Aug 2003 05:22:03 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7M9M3i2031284
	for pmtud-archive@odin.ietf.org; Fri, 22 Aug 2003 05:22:03 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q87P-00088P-Ir
	for pmtud-web-archive@optimus.ietf.org; Fri, 22 Aug 2003 05:22:03 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA12656
	for <pmtud-web-archive@ietf.org>; Fri, 22 Aug 2003 05:21:57 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19q87M-0001F8-00
	for pmtud-web-archive@ietf.org; Fri, 22 Aug 2003 05:22:00 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19q87L-0001F5-00
	for pmtud-web-archive@ietf.org; Fri, 22 Aug 2003 05:21:59 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q87N-00087z-Rs; Fri, 22 Aug 2003 05:22:01 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19q86s-00086n-Jo
	for pmtud@optimus.ietf.org; Fri, 22 Aug 2003 05:21:30 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA12591
	for <pmtud@ietf.org>; Fri, 22 Aug 2003 05:21:24 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19q86p-0001EH-00
	for pmtud@ietf.org; Fri, 22 Aug 2003 05:21:27 -0400
Received: from penguin-ext.wise.edt.ericsson.se ([193.180.251.47])
	by ietf-mx with esmtp (Exim 4.12)
	id 19q86o-0001EE-00
	for pmtud@ietf.org; Fri, 22 Aug 2003 05:21:26 -0400
Received: from esealnt610.al.sw.ericsson.se ([153.88.254.120])
	by penguin-ext.wise.edt.ericsson.se (8.12.9/8.12.9/WIREfire-1.7) with ESMTP id h7M9LQ31026676;
	Fri, 22 Aug 2003 11:21:26 +0200 (MEST)
Received: from ericsson.com (research-nnng7k.ki.sw.ericsson.se [147.214.34.132]) by esealnt610.al.sw.ericsson.se with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2655.55)
	id QY1J65QT; Fri, 22 Aug 2003 11:22:04 +0200
Message-ID: <3F45E093.1070301@ericsson.com>
X-Sybari-Space: 00000000 00000000 00000000 00000000
From: Magnus Westerlund <magnus.westerlund@ericsson.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030624 Netscape/7.1 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Matthew J Zekauskas <matt@internet2.edu>
CC: pmtud@ietf.org, Matt Mathis <mathis@psc.edu>
Subject: Re: [pmtud] pmtud ietf57 vienna minutes, version 0
References: <27046290.1061489137@localhost>
In-Reply-To: <27046290.1061489137@localhost>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Fri, 22 Aug 2003 11:21:23 +0200
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit

Hi,

Some comments on the minutes.

Matthew J Zekauskas wrote:
> In thinking about other stakeholders, one person felt that the
> algorithm would work for RTP over UDP; there you have some
> extensions to share information -- reliable instrumentation of
> losses. [XXX]

I think this is my comments. With usage of RTCP extensions for packet 
loss vectors (there are two available) the method is possible to use.

As an outside of the minutes comment:
There is probably some more considerations to make when using it for 
RTP/UDP. However I think there is certainly need for something like 
this. There is some RTP payload formats that generate extremely high 
bit-rates, on the order of Gigabits, so there is definitely need to 
check for the path MTU.

> 
> Someone commented that the document as written is very TCP specific.
> The algorithm should be better separated from actual deployment.
> Matt M said that's the intention.
> 

This is also my comment.

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


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



From exim@www1.ietf.org  Fri Aug 22 15:35:40 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA15829
	for <pmtud-archive@odin.ietf.org>; Fri, 22 Aug 2003 15:35:40 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19qHgo-0004Nq-RJ
	for pmtud-archive@odin.ietf.org; Fri, 22 Aug 2003 15:35:14 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7MJZEKm016850
	for pmtud-archive@odin.ietf.org; Fri, 22 Aug 2003 15:35:14 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19qHgo-0004Nd-Gm
	for pmtud-web-archive@optimus.ietf.org; Fri, 22 Aug 2003 15:35:14 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA15809
	for <pmtud-web-archive@ietf.org>; Fri, 22 Aug 2003 15:35:10 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19qHgn-0000xy-00
	for pmtud-web-archive@ietf.org; Fri, 22 Aug 2003 15:35:13 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19qHgm-0000xj-00
	for pmtud-web-archive@ietf.org; Fri, 22 Aug 2003 15:35:12 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19qHgd-0004MR-CI; Fri, 22 Aug 2003 15:35:03 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19qHgK-0004LE-So
	for pmtud@optimus.ietf.org; Fri, 22 Aug 2003 15:34:44 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA15759
	for <pmtud@ietf.org>; Fri, 22 Aug 2003 15:34:40 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19qHgJ-0000wy-00
	for pmtud@ietf.org; Fri, 22 Aug 2003 15:34:43 -0400
Received: from pa-monroeville3a-31.pit.adelphia.net ([24.53.185.31] helo=zippy.psc.edu)
	by ietf-mx with esmtp (Exim 4.12)
	id 19qHgI-0000wC-00
	for pmtud@ietf.org; Fri, 22 Aug 2003 15:34:42 -0400
Received: from localhost (mathis@localhost)
	by zippy.psc.edu (8.11.6/8.11.6) with ESMTP id h7MJYC703720
	for <pmtud@ietf.org>; Fri, 22 Aug 2003 15:34:12 -0400
X-Authentication-Warning: zippy.psc.edu: mathis owned process doing -bs
From: Matt Mathis <mathis@psc.edu>
To: Path MTU Discovery WG <pmtud@ietf.org>
Subject: Re: [pmtud] the DF-bit honouring problem
In-Reply-To: <16050.1061517451@marajade.sandelman.ottawa.on.ca>
Message-ID: <Pine.LNX.4.44.0308221347120.1275-100000@zippy.psc.edu>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Fri, 22 Aug 2003 15:34:12 -0400 (EDT)

On Thu, 21 Aug 2003, Michael Richardson wrote:
> As we discussed at the meeting last month, the plpmtud proposal breaks
> when there are tunnelling systems that do not obey the DF bit. A second
> order effect is when tunnelling systems do not copy the DF bit to the
> outer header when they encapsulate.

... snip ...

> The risk to pmtumtud of DF-ignoring systems is that they cause the 
> algorithm to fail, since all probes would get through. The result is more
> fragments, and possible much higher changes of failure to communicate.

In this environment connections don't fail: the endsystem would raise the MTU,
the out-of-spec tunnels endpoint ignore DF and fragment, and the packet is
delivered at the cost of wasted resources - mostly CPU on the tunnel endpoint.
This is no worse than the current situation.  Plpmtud only failed to make it 
better.

During the meeting, I suggested that the tunnel endpoint might use some
additional criteria to minimize extra overhead, for example honoring DF (and
tossing the packet) if the upper fragment would be larger than a tunnel header.  
I believe that with this additional check, changing the mtu discovery algorithm
does not change the consequences/impact of non-DF compliant tunnels.

But I think there are ways to do even better.  For example if you have a
heuristic to detect probes (e.g. the first over sized packet after several more 
than half sized packets) you can discard suspected probes.  This has a slight
impact on the old algorithm for a large gain on the new.   As you sharpen the 
heuristic I bet you will be able to do really well.

Thanks,
--MM--
-------------------------------------------
Matt Mathis      http://www.psc.edu/~mathis
Work:412.268.3319    Home/Cell:412.654.7529
-------------------------------------------



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



From exim@www1.ietf.org  Wed Aug 27 19:38:31 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA19151
	for <pmtud-archive@odin.ietf.org>; Wed, 27 Aug 2003 19:38:31 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19s40y-0004tU-II
	for pmtud-archive@odin.ietf.org; Wed, 27 Aug 2003 13:23:24 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7RHNO3L018810
	for pmtud-archive@odin.ietf.org; Wed, 27 Aug 2003 13:23:24 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19s3Wb-00031D-3o
	for pmtud-web-archive@optimus.ietf.org; Wed, 27 Aug 2003 12:52:01 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA16526
	for <pmtud-web-archive@ietf.org>; Wed, 27 Aug 2003 12:51:55 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19s3WZ-000321-00
	for pmtud-web-archive@ietf.org; Wed, 27 Aug 2003 12:51:59 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19s3WY-00031w-00
	for pmtud-web-archive@ietf.org; Wed, 27 Aug 2003 12:51:58 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19s35R-00018k-Ps; Wed, 27 Aug 2003 12:23:57 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19rsbD-0006F3-17
	for pmtud@optimus.ietf.org; Wed, 27 Aug 2003 01:12:03 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA22523
	for <pmtud@ietf.org>; Wed, 27 Aug 2003 01:11:58 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19rsbA-0004wa-00
	for pmtud@ietf.org; Wed, 27 Aug 2003 01:12:00 -0400
Received: from basie.internet2.edu ([207.75.164.22])
	by ietf-mx with esmtp (Exim 4.12)
	id 19rsb9-0004wX-00
	for pmtud@ietf.org; Wed, 27 Aug 2003 01:11:59 -0400
Received: from localhost (localhost.localdomain [127.0.0.1])
	by basie.internet2.edu (Postfix) with ESMTP
	id 6E2D27B477; Wed, 27 Aug 2003 01:12:00 -0400 (EDT)
Received: from BMW (basie.internet2.edu [207.75.164.22])
	by basie.internet2.edu (Postfix) with ESMTP
	id EB1A87B46D; Wed, 27 Aug 2003 01:11:58 -0400 (EDT)
From: Matthew J Zekauskas <matt@internet2.edu>
To: pmtud@ietf.org
Cc: Matt Zekauskas <matt@internet2.edu>
Message-ID: <124800403.1061946314@localhost>
X-Mailer: Mulberry/2.2.1 (Win32)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
X-Virus-Scanned: by mail.internet2.edu virus scanner
Content-Transfer-Encoding: 7bit
Subject: [pmtud] The numbers 1280 and 1380
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Wed, 27 Aug 2003 01:05:14 -0400
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit

I asked Dave Thaler where these numbers came from, and here's
his reply.  Forwarded with permission.

--Matt

---------- Forwarded Message ----------
Date: Tuesday, August 26, 2003 4:17 PM -0700
From: Dave Thaler <dthaler@windows.microsoft.com>
To: Matthew J Zekauskas <matt@internet2.edu>
Cc: Erik Nordmark <Erik.Nordmark@sun.com>
Subject: RE: pmtud comments...

Matthew writes:
> I'm curious (and someone actually asked on the pmtud list) where
> the 1280/1380 numbers come from.

RFC 2460, section 5:
>  IPv6 requires that every link in the internet have an MTU of 1280
>  octets or greater.

1380 comes from the fact that a widely-deployed L2TP tunnel MTU is 1400.
So you end up with 1400 - IPv4 header size = 1380 bytes available for
the IPv4 payload.  (Aside: the number 1400 I believe was originally
obtained from Ethernet MTU - space for IPv4+AH+ESP+UDP+L2TP+PPP.)
Hence, when IPv6 is tunneled inside this tunnel (e.g. using ISATAP or
whatever other IPv6-in-IPv4 encapsulation method), the link MTU of that
tunnel could be anything 1280 or higher.  However, once you go over
1380, you'll cause fragmentation when going over an underlying link with
an MTU of 1400.

In a separate scenario, you have Mobile IPv6 in use where you're
tunneling
over IPv6.  If the MTU of the underlying IPv6 link is small (like 1280)
then you'll force Mobile IPv6 to be doing fragmentation when it
encapsulates, since the MTU of its tunnel has to be at least 1280, which
doesn't leave room for the encapsulation overhead.

Hence, when discussing a "default" tunnel MTU, anything over 1380 causes
gratuitous fragmentation in one scenario we expect to be common.
Anything much below 1380 will cause gratuitous fragmentation in a
different
scenario (Mobile IPv6).  And of course lower MTUs result in more
overhead to get the same throughput.

This is why I recommend the value 1380 as the best default tunnel MTU.
Those who argue to just say 1280 are essentially saying they don't care
about gratuitous fragmentation caused by Mobile IPv6 over the tunnel.

-Dave

P.S. I think I'm not yet on the PMTUD list, so feel free to forward this
if you want.



---------- End Forwarded Message ----------



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



From exim@www1.ietf.org  Wed Aug 27 23:36:30 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA10180
	for <pmtud-archive@odin.ietf.org>; Wed, 27 Aug 2003 23:36:29 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sByC-0007OH-RM
	for pmtud-archive@odin.ietf.org; Wed, 27 Aug 2003 21:53:05 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7S1r4ce028405
	for pmtud-archive@odin.ietf.org; Wed, 27 Aug 2003 21:53:04 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sBBB-0003fh-JD
	for pmtud-web-archive@optimus.ietf.org; Wed, 27 Aug 2003 21:02:25 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA25594
	for <pmtud-web-archive@ietf.org>; Wed, 27 Aug 2003 21:02:20 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sBB9-0004rd-00
	for pmtud-web-archive@ietf.org; Wed, 27 Aug 2003 21:02:23 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19sBB8-0004ra-00
	for pmtud-web-archive@ietf.org; Wed, 27 Aug 2003 21:02:22 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19s4u7-0000qV-KW; Wed, 27 Aug 2003 14:20:23 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19rmmS-00054b-SK
	for pmtud@optimus.ietf.org; Tue, 26 Aug 2003 18:59:16 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA18833;
	Tue, 26 Aug 2003 18:59:08 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19rmmP-00045c-00; Tue, 26 Aug 2003 18:59:13 -0400
Received: from basie.internet2.edu ([207.75.164.22])
	by ietf-mx with esmtp (Exim 4.12)
	id 19rmmN-00045Y-00; Tue, 26 Aug 2003 18:59:12 -0400
Received: from localhost (localhost.localdomain [127.0.0.1])
	by basie.internet2.edu (Postfix) with ESMTP
	id 240937B4F8; Tue, 26 Aug 2003 18:59:08 -0400 (EDT)
Received: from BMW (basie.internet2.edu [207.75.164.22])
	by basie.internet2.edu (Postfix) with ESMTP
	id 71AE87B47E; Tue, 26 Aug 2003 18:59:06 -0400 (EDT)
From: Matthew J Zekauskas <matt@internet2.edu>
To: minutes@ietf.org
Cc: pmtud@ietf.org, Matt Zekauskas <matt@internet2.edu>,
        Matt Mathis <mathis@psc.edu>
Message-ID: <102831143.1061924345@localhost>
X-Mailer: Mulberry/2.2.1 (Win32)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
X-Virus-Scanned: by mail.internet2.edu virus scanner
Content-Transfer-Encoding: 7bit
Subject: [pmtud] Minutes from the pmtud meeting at IETF57 - VIE
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Tue, 26 Aug 2003 18:59:05 -0400
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit

Path Maximum Transmission Unit Discovery Pre-WG (pmtud)
Thursday, 17-July-2003 from 13:00 to 15:00
=======================================================

The meeting was chaired by Matt Mathis and Matt Zekauskas.
Al Morton, Itojun Hagino and Matt Z. took notes that were
assembled into these minutes by the chairs.

Agenda
------
 * Preliminaries: Blue sheets, Note takers, etc
 * WG Status
 * Short history and work to date
 * Robustness Issues
 * Other Stakeholders
 * Plans

Matt Mathis led off the meeting presenting the new co-chair, the
agenda, the changes to the proposed charter, and the aggressive
milestones.

The group status is that some parts of the administrative preparation
did not get done, but IESG has approved, hence "Pre WG".  This will be
a fast development, silence will be acceptance (at the start as sections
are integrated).  PMTUD is a re-activation of path MTU WG, which was
a very similar effort.  To participate, you must subscribe to IETF list
(pmtud@ietf.org).  Matt Zekauskas volunteered to Co-chair the group.
The charter was broadened so as not to restrict to a single method.
Milestones are aggressive, and need for implementation to test is clear.

No one disagreed, nor were there suggestions for other methods to
study.  Matt Mathis will make another editorial pass at the document
(which was a rewrite, starting from RFC 1981, instead of an update to
the previous BOF input).  Sections will be added based on mailing list
comments and any input from stakeholder communities.

Matt Mathis sketched the previous algorithm and noted some of the
problems.  He then sketched the new algorithm, noting that there
is just a small amount of MUST/SHOULD language: under what circumstances
can losses be ignored as a congestion signal.  The rest is heuristics;
it doesn't need to be the same for every application, and permits
vendor diversity.

Question: when can the algorithm be used by TCP?  Just after 3-way
handshake, or before real communication?  Matt responded that it
uses live payload data, and the draft has a recommendation not
to attempt the algorithm unless the congestion window is at least
twenty packets, so the connection is well established before the
algorithm starts.  Thus, this could slow down tiny files -- the
exact algorithm is a heuristic, so you could choose to perform
it differently.  There are tradeoffs.

The request for collaborators in the IPsec & security area led to a
big discussion on tunnel issues.  People were positive about the
method, but there are corner cases to consider.  Input has been
promised to the mailing list.  (In the IPv4 world, lack of PMTUD is
noted as a major problem with IPsec VPNs and providing services.)

This started around slide 11, "Plans for the Next Draft".  One
of the collaborators was folks from the multicast area; one possibility
is a generalization of the algorithm for reliable delivery.
This would solve an ICMP implosion problem if the current MTU
discovery technique was used.  Dave Thaler noted that the behavior
as specified in IPv4 and IPv6 was different, in 6 you respond,
in v4 you don't.

Another collaborating group would be IPsec; currently the security
architecture document has major sections dealing with the interaction
of MTU discovery and IPsec (because tunnels are created); the new
technique might obsolete many of those sections.

Itojun Hagino noted that the interaction between IPsec and and TCP
depends on if the TCP stack is aware of IPsec.  If the TCP stack does
not take care of the IPsec header size, the algorithm would need to be
revised.

Matt M. responded that the detail in the draft needs to be resolved
in a consistent way.  You can count the IPsec header as part of the
IP header or TCP header.  The really nasty cases involve additional
layers, for example IPsec on a VPN, ICMP messages could go back to
the wrong place.

Michael Richardson expanded on this as an IPsec implementer.  The
worst case is common at meetings such as this -- you have a corporate
address on your laptop, and a VPN back to the corporate space, so
all traffic goes back to HQ.  Try to visit a bank, and they have
ICMP filters.  Your gateway is sending out ICMP messages to the
bank, and they drop them.  This proposed algorithm should work
really well.  Many times VPNs are blamed (since they are the newest
element in the path), when the problem is really a bad ICMP filter.

However, there is a problem, if you raise the MTU and the tunnels do
not toss large messages but fragment them anyway, you will end up
always fragmenting.  Michael noted that his Linux implementation (he's
the FreeS/WAN technical lead) did not honor the DF bit by default.
Having a poor(er) performing implementation was better than one that
didn't work at all.  ("Poor performance is better than no security.")
Perhaps there could be a heuristic that worked for a short term
solution so these mechanisms don't interact badly... the endpoint
would need to be updated for this algorithm, so IPsec tunneling could
be updated at the same time.  This behavior is often a kernel option,
too.  [In reviewing the minutes, Michael related that:
 "The key point is that I, the IPsec developer can't control:
      1) the ICMP filter.
      2) the TCP on the machine behind it.
 I can have *some* influence on the TCP at the receiving end of the flow,
 but not a lot. *If* the IPsec tunnel terminates on the same machine as the
 TCP, then in theory, the TCP can learn about the reduced MTU, and set the
 MSS appropriately. In practice, probably only KAME, Microsoft and Sun are
 well enough integrated to do this right now. Probably Linux 2.6 will be able
 to do so as well.
  We do have an option to hack the MSS on the encapsulator's side already,
 alas."]

Perhaps you could fragment into tunnel, but retain the DF bit, and if
set don't do anything weird.  Itojun related KAME experience; there
they ended up not setting the DF bit on output header when IPsec
tunnels are created.

Another point was that IPv6 on IPv4 tunnels have the same issue.
IPv6 tunnels should have a MTU of 1280 by default so a minimum MTU
can be maintained.

Matt M. mentioned that he's aware that a large number of tunneling
implementations don't copy the DF bit from inner packet to outer
header.  He's not yet sure if the document needs a specific section
covering tunnels and tunnel migration; an intermediate ground that
works is to let tunnels behave this way in the interim, and discourage
a mode where end systems ignore can't fragment messages.

Dave Thaler noted that mobility might add additional headers; so a
tunnel MTU of 1280 might not be enough; 1380 would be better.

Itojun stated that he was thinking of configured tunnels and not mobile
IPv6.  If you send a packet with mobile headers, the TCP stack needs
to be aware of the size of the mobile IP headers and reduce MSS
appropriately -- maintain the total MTU size.

Matt Mathis noted that there was definitely a subgroup interested in
considering tunneling and MTU discovery; he encouraged folks to
join the mailing list and contribute the various circumstances where
there are potential problems.

Lars Eggert mentioned that RFC2003 specifies some things related
to MTU discovery, and RFC2401 specifically prohibits some of the
mechanisms in RFC2003 for security reasons.  Joe Touch also has
a relevant draft: draft-touch-ipsec-vpn.

As to other transport protocols, Matt Z. reported that he had
quickly skimmed SCTP and DCCP documents, and that SCTP looked
possible, but DCCP says specifically that the MTU can't be raised.
No one that claimed to be an SCTP expert was in the room (or at
least didn't comment negatively on the applicability to SCTP).
Eddie Kohler noted that this behavior was revised in the
DCCP WG meeting this week.   Matt Z. prompted him to send some DCCP text.

Matt M. emphasized that the point with getting a draft done early
is to encourage implementation as soon as possible.  The algorithm
will use specific details of other protocols, and we're dependent
on the uniformity of implementation of certain features.  We need
to learn what implementations really do; ideally get a custom
implementation run on servers and real field data to feed back
into the document.

Matt M also noted two cases that he's worried about (although
these are just examples; others are encouraged to consider other
cases, or report back implementation experience).  First, what
happens if a path is striped across multiple links, and the MTU
is not the same across the stripes?  You can require that the
MTU is not raised until a certain number of segments are received
successfully.  You need to understand the interaction between random
losses and whether the MTU is or is not raised.  Second, what happens
if there is a parametric failure -- when raising the MTU causes the
error rate to increase?  An actual case is one particular 10G gbic;
it was error free with 1500 byte packets, but not with 9000 byte
packets.  There is an opportiunity for different heuristics here,
for example use a smaller MSS if you cannot fill a window.

Michael thought a "brokenness test page" was needed -- a good testbench.

For hard, repeated, timeouts the first thing you want to do is
reduce congestion variables, then reduce MTU.  At some point want
to restart the checks to increase MTU.

There are other possible protocol interactions, too: for example,
SCTP can use multiple endpoints.  What if it changes addresses,
and the new path has a smaller MTU?

Michael felt that it was important to focus where the production
environments hurt most with the current MTU scheme.  Matt M noted
that different things hurt in different environments.  Michael
expanded that the most frequent case will likely be large port 80
responses to a client.  And it's the client that would decide that
the path is stupid or broken, and other A record should be tried.
The web server is getting the timeouts, not the client.  This won't
deploy if we can't solve the web server case.

Matt M noted another case he had thought about, but not seen: what
happens if raising the MTU causes link stability problems (as opposed
to hard failures) -- say the link "goes away" for 10 seconds and then
returns.  He's thought about using a state machine to catch this case...
the link is broken, and we don't necessarily want to fix it with
an MTU discovery algorithm.

On ignoring DF bits, so that a tunnel fragments large packets: Matt M
contended it was worse for a 1500 tunnel fragmenting a 9000 packet
than a tunnel fragmenting a 1500 packet by the tunnel overhead.
Michael didn't understand this at first; Matt explained that
the problem is that with many fragments the odds are greater that
you lose a fragment, and hence the whole packet than if there
are only two fragments.

In thinking about other stakeholders, Magnus Westerlund felt that the
algorithm would work for RTP over UDP with the use of RTCP extensions
for packet loss vectors.

Itojun said that we should contact rrs@cisco.com for SCTP.

In the multicast case, Dave Thaler argued that this algorithm might
cause an ACK implosion that is worse than a ICMP-message-too-big
implosion, since there are typically far more receivers than there are
bottleneck-MTU links.

Magnus commented that the document as written is very TCP specific.
The algorithm should be better separated from actual deployment.
Matt M said that's the intention.

Another question was what, exactly, is the definition of MTU?  End-to-end
or link-specific.  Matt M said that we were talking about IP MTU
when using a particular link layer; how the IETF uses it, not what
hardware specifications say.

Dave Thaler argued that IPv6 might not need this at all; the algorithm
could arguably make performance worse (since the MSS size would ramp
up instead of being decided once by ICMP-message-too-big).  Matt M.
said the new algorithm would prevent against implementation or
configuration bugs and also work in the cases where L2 MTUs were
different on a switch.

One audience member said that this should be documented in detail.

Dave Thaler mentioned that filtering ICMPv6 has larger problems.  IPv6
neighbor discovery uses ICMPv6, so if ICMP is filtered you won't get
connectivity.  In addition, since v6 has no DF bit (but implied DF on
all packets) blocking ICMP definitely leads to black hole problems in
the network.  Thus, there is already a natural incentive to allow PMTU
using ICMP.

Matt M noted that some stacks have IPv4 mimic IPv6 -- they always set
DF, even on fragments,  and attempt to fragment only at the endpoints.

However, there's no requirement that routers send the too big messages
in v4, but there is a requirement in v6.

Another comment was tha the MTU in Router Advertisement messages should
solve this problem.  If operational experience says this isn't happening,
it should be reflected to the v6 working group.

Matt M said that in all cases the problems with path MTU discovery
are bugs.  There are a large set of problems.



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



From exim@www1.ietf.org  Thu Aug 28 03:07:53 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA09854
	for <pmtud-archive@odin.ietf.org>; Thu, 28 Aug 2003 03:07:53 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sENm-0001a9-IU
	for pmtud-archive@odin.ietf.org; Thu, 28 Aug 2003 00:27:39 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7S4Rc30006075
	for pmtud-archive@odin.ietf.org; Thu, 28 Aug 2003 00:27:38 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sCdt-0002Aw-FB
	for pmtud-web-archive@optimus.ietf.org; Wed, 27 Aug 2003 22:36:09 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA03994
	for <pmtud-web-archive@ietf.org>; Wed, 27 Aug 2003 22:36:03 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sCdq-0007Bs-00
	for pmtud-web-archive@ietf.org; Wed, 27 Aug 2003 22:36:06 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19sCdp-0007Bp-00
	for pmtud-web-archive@ietf.org; Wed, 27 Aug 2003 22:36:05 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sBhj-0006NW-BT; Wed, 27 Aug 2003 21:36:03 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sBTw-0005D6-BO
	for pmtud@optimus.ietf.org; Wed, 27 Aug 2003 21:21:48 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA27147
	for <pmtud@ietf.org>; Wed, 27 Aug 2003 21:21:42 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sBTs-0005J8-00
	for pmtud@ietf.org; Wed, 27 Aug 2003 21:21:44 -0400
Received: from oetest.freeswan.org ([205.150.200.166] helo=noxmail.sandelman.ottawa.on.ca)
	by ietf-mx with esmtp (Exim 4.12)
	id 19sBTr-0005Ix-00
	for pmtud@ietf.org; Wed, 27 Aug 2003 21:21:43 -0400
Received: from sandelman.ottawa.on.ca ([2002:cd96:c8ed::1])
	by noxmail.sandelman.ottawa.on.ca (8.11.6p2/8.11.6) with ESMTP id h7S1Lfd23414
	(using TLSv1/SSLv3 with cipher EDH-RSA-DES-CBC3-SHA (168 bits) verified OK);
	Wed, 27 Aug 2003 21:21:42 -0400 (EDT)
Received: from marajade.sandelman.ottawa.on.ca (mcr@localhost)
	by sandelman.ottawa.on.ca (8.12.3/8.12.3/Debian -4) with ESMTP id h7S1LfXd010075;
	Wed, 27 Aug 2003 21:21:41 -0400
To: mtu@psc.edu
cc: Michael Welzl <michael.welzl@uibk.ac.at>, pmtud@ietf.org
Subject: Re: [pmtud] MTU IP Option (reviving RFC 1063) 
In-reply-to: Your message of "Wed, 27 Aug 2003 17:34:28 PDT."
             <20030828003428.6078.qmail@patheticgeek.net> 
Mime-Version: 1.0 (generated by tm-edit 1.8)
Content-Type: text/plain; charset=US-ASCII
Message-ID: <10074.1062033700@marajade.sandelman.ottawa.on.ca>
From: Michael Richardson <mcr@sandelman.ottawa.on.ca>
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Wed, 27 Aug 2003 21:21:40 -0400

-----BEGIN PGP SIGNED MESSAGE-----


>>>>> "Kevin" == Kevin Lahey <kml@patheticgeek.net> writes:
    >> I have a proposal for PMTUD which could be integrated in the "new" as
    >> well as the "old" approach. The idea is to revive RFC 1063 (the MTU
    >> Discovery Option, which requires a router to update the MTU value in
    >> the option if its MTU is smaller than the value already in there).
    >> Back when this option was conceived, there was no PMTUD - so PMTUD was
    >> clearly a better idea because the option would require all routers to
    >> participate.

    Kevin> Whenever this is proposed, a wail seems to go up from the routing
    Kevin> folks.  The oft-repeated claim is that processing options on some
    Kevin> reasonable percentage of the packets would be incredably slow.

  Is there some way that we can put in a option that is not a hop-by-hop
option, but that things that have MTU-constraints might be willing to process?

]      Out and about in Ottawa.    hmmm... beer.                |  firewalls  [
]   Michael Richardson, Sandelman Software Works, Ottawa, ON    |net architect[
] mcr@sandelman.ottawa.on.ca http://www.sandelman.ottawa.on.ca/ |device driver[
] panic("Just another Debian/notebook using, kernel hacking, security guy");  [
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)
Comment: Finger me for keys - custom hacks make this fully PGP2 compat

iQCVAwUBP01ZI4qHRg3pndX9AQF9MgP9FFV5cgpp2MzfzSaDm98GBGd9DCuWooFQ
Hyox7z8KDyEQGtmjFXejAdv/f+PAWCJ3zaL06sRc3VS1Y1jTjKWcChT2lL7FIZk7
h55Qv4MR0Ctz5HLz7irr78eV5yfSGoxM9S2JLDMDtXXMFyypuxAs3WTwX4Esw8HX
wf+eYBCCfh4=
=JZBY
-----END PGP SIGNATURE-----

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



From exim@www1.ietf.org  Thu Aug 28 04:24:23 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id EAA16118
	for <pmtud-archive@odin.ietf.org>; Thu, 28 Aug 2003 04:24:22 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sCq8-0003Ri-1K
	for pmtud-archive@odin.ietf.org; Wed, 27 Aug 2003 22:48:48 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7S2mlpF013237
	for pmtud-archive@odin.ietf.org; Wed, 27 Aug 2003 22:48:47 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sA48-00082b-2F
	for pmtud-web-archive@optimus.ietf.org; Wed, 27 Aug 2003 19:51:04 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA19987
	for <pmtud-web-archive@ietf.org>; Wed, 27 Aug 2003 19:50:57 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sA46-0003O3-00
	for pmtud-web-archive@ietf.org; Wed, 27 Aug 2003 19:51:02 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19sA45-0003O0-00
	for pmtud-web-archive@ietf.org; Wed, 27 Aug 2003 19:51:01 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19s4vg-00015Y-JW; Wed, 27 Aug 2003 14:22:00 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19rwDr-0002lh-Nb
	for pmtud@optimus.ietf.org; Wed, 27 Aug 2003 05:04:11 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA08505
	for <pmtud@ietf.org>; Wed, 27 Aug 2003 05:04:06 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19rwDo-0001eP-00
	for pmtud@ietf.org; Wed, 27 Aug 2003 05:04:08 -0400
Received: from lmr1.uibk.ac.at ([138.232.1.142] helo=smtp.uibk.ac.at)
	by ietf-mx with esmtp (Exim 4.12)
	id 19rwDn-0001e3-00
	for pmtud@ietf.org; Wed, 27 Aug 2003 05:04:08 -0400
Received: from lap10-c703.uibk.ac.at (lap10-c703.uibk.ac.at [138.232.65.57])
	by smtp.uibk.ac.at (8.12.9/8.12.9/F1) with ESMTP id h7R93cWj000434;
	Wed, 27 Aug 2003 11:03:38 +0200
Subject: Re: [pmtud] MTU IP Option (reviving RFC 1063)
From: Michael Welzl <michael.welzl@uibk.ac.at>
To: mtu@psc.edu
Cc: pmtud@ietf.org
In-Reply-To: <2946.1061482267@marajade.sandelman.ottawa.on.ca>
References: <2946.1061482267@marajade.sandelman.ottawa.on.ca>
Content-Type: text/plain
Organization: University of Innsbruck
Message-Id: <1061974978.1737.130.camel@lap10-c703>
Mime-Version: 1.0
X-Mailer: Ximian Evolution 1.2.2 (1.2.2-4) 
Content-Transfer-Encoding: 7bit
X-Spam-Score: -7.9 () EMAIL_ATTRIBUTION,IN_REP_TO,QUOTED_EMAIL_TEXT,RCV_UIBK,REFERENCES,REPLY_WITH_QUOTES,USER_AGENT_XIMIAN
X-Scanned-By: MIMEDefang 2.35 at uibk.ac.at
Content-Transfer-Encoding: 7bit
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: 27 Aug 2003 11:02:58 +0200
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit

Hi all,

This mail from last week was the last response to my
proposal ... so what? Should I put it in writing?
Matt?

Cheers,
Michael


On Thu, 2003-08-21 at 18:11, Michael Richardson wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> 
> 
> >>>>> "Michael" == Michael Welzl <michael.welzl@uibk.ac.at> writes:
>     Michael> Is it a good idea?
> 
>     Michael> Router perspective: =================== Normally, routers send
>     Michael> ICMP error messages back if the MTU is exceeded. Router X which
>     Michael> has a very small MTU must generate a large number of such
>     Michael> packets. Once this new option is deployed, participating in the
> 
>   The routers who are causing the MTU constraint clearly have some
> obligation, and as an IPsec implementor, I'm happy to update my code like
> this.
>   So, this is a good idea. It may not be sufficient to deal with the problem.
> 
> ]      Out and about in Ottawa.    hmmm... beer.                |  firewalls  [
> ]   Michael Richardson, Sandelman Software Works, Ottawa, ON    |net architect[
> ] mcr@sandelman.ottawa.on.ca http://www.sandelman.ottawa.on.ca/ |device driver[
> ] panic("Just another Debian/notebook using, kernel hacking, security guy");  [
> 
>   
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.2 (GNU/Linux)
> Comment: Finger me for keys - custom hacks make this fully PGP2 compat
> 
> iQCVAwUBP0TvGoqHRg3pndX9AQHPzAP+M82eZ9WOz6CmnGdwG2gQ2rWru+X8o2q4
> MYtTBAt/wUmnVmv6kxB/rGfwpeOI03nWfaCr0hatQS/sbrZGDTAWq0iSH5XPezFh
> rmXPk7kbElw1CNONwUUQVrL1vBt9GzydjneZt0ZN3awhlragbZ0fIFZpINdUjYy2
> 23vN6/LO+eA=
> =8rTo
> -----END PGP SIGNATURE-----
> ----------------------------------------------------------------------
> This message was sent to mtu@psc.edu.
> Send questions about this list to: mtu-owner@psc.edu.
> For help using this service, send a blank message to: majordomo@psc.edu
-- 
Michael Welzl <michael.welzl@uibk.ac.at>
University of Innsbruck


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



From exim@www1.ietf.org  Thu Aug 28 07:03:26 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA00449
	for <pmtud-archive@odin.ietf.org>; Thu, 28 Aug 2003 07:03:26 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sIuI-0001hF-3X
	for pmtud-archive@odin.ietf.org; Thu, 28 Aug 2003 05:17:30 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7S9HUXY006510
	for pmtud-archive@odin.ietf.org; Thu, 28 Aug 2003 05:17:30 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sE9r-0000yP-Of
	for pmtud-web-archive@optimus.ietf.org; Thu, 28 Aug 2003 00:13:15 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA13222
	for <pmtud-web-archive@ietf.org>; Thu, 28 Aug 2003 00:13:09 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sE9p-0001so-00
	for pmtud-web-archive@ietf.org; Thu, 28 Aug 2003 00:13:13 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19sE9o-0001sl-00
	for pmtud-web-archive@ietf.org; Thu, 28 Aug 2003 00:13:12 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sCM5-0000Xc-25; Wed, 27 Aug 2003 22:17:45 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19s7i2-0000pl-Gc
	for pmtud@optimus.ietf.org; Wed, 27 Aug 2003 17:20:06 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA09463
	for <pmtud@ietf.org>; Wed, 27 Aug 2003 17:20:00 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19s7i0-0000ke-00
	for pmtud@ietf.org; Wed, 27 Aug 2003 17:20:04 -0400
Received: from darkstar.iprg.nokia.com ([205.226.5.69])
	by ietf-mx with esmtp (Exim 4.12)
	id 19s7hy-0000jw-00
	for pmtud@ietf.org; Wed, 27 Aug 2003 17:20:03 -0400
Received: (from root@localhost)
	by darkstar.iprg.nokia.com (8.11.0/8.11.0-DARKSTAR) id h7RLJUO18791;
	Wed, 27 Aug 2003 14:19:30 -0700
X-mProtect: <200308272119> Nokia Silicon Valley Messaging Protection
Received: from UNKNOWN (205.226.2.67, claiming to be "iprg.nokia.com")
	by darkstar.iprg.nokia.com smtpdNfiypi; Wed, 27 Aug 2003 14:19:29 PDT
Message-ID: <3F4D2274.7020608@iprg.nokia.com>
From: Fred Templin <ftemplin@iprg.nokia.com>
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.2) Gecko/20021120 Netscape/7.01
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: pmtud@ietf.org
CC: mtu@psc.edu
References: <2946.1061482267@marajade.sandelman.ottawa.on.ca> <1061974978.1737.130.camel@lap10-c703>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Subject: [pmtud] MTU-related drafts - any interest in pursuing here?
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Wed, 27 Aug 2003 14:28:20 -0700
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit

Earlier this year I prepared two interrelated drafts on Path MTU
support for IPv6-in-IPv4 tunnels, but their domain of applicability
may also extend to other use cases. The drafts can be found at:

  http://www.ietf.org/internet-drafts/draft-templin-tunnelmtu-00.txt
  http://www.ietf.org/internet-drafts/draft-templin-ndiscmtu-00.txt

Is there any interest in pursuing this work further in this group?

Fred
ftemplin@iprg.nokia.com


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



From exim@www1.ietf.org  Thu Aug 28 07:30:49 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id HAA03290
	for <pmtud-archive@odin.ietf.org>; Thu, 28 Aug 2003 07:30:49 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sJhn-0004N1-NR
	for pmtud-archive@odin.ietf.org; Thu, 28 Aug 2003 06:08:40 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7SA8drf016789
	for pmtud-archive@odin.ietf.org; Thu, 28 Aug 2003 06:08:39 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sFVE-0005HM-Av
	for pmtud-web-archive@optimus.ietf.org; Thu, 28 Aug 2003 01:39:24 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id BAA20565
	for <pmtud-web-archive@ietf.org>; Thu, 28 Aug 2003 01:39:19 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sFVB-0003qK-00
	for pmtud-web-archive@ietf.org; Thu, 28 Aug 2003 01:39:21 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19sFVA-0003qH-00
	for pmtud-web-archive@ietf.org; Thu, 28 Aug 2003 01:39:20 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sCGX-000055-Ke; Wed, 27 Aug 2003 22:12:01 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sAqC-0002L2-9m
	for pmtud@optimus.ietf.org; Wed, 27 Aug 2003 20:40:44 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA23356
	for <pmtud@ietf.org>; Wed, 27 Aug 2003 20:40:39 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sAqA-0004Fp-00
	for pmtud@ietf.org; Wed, 27 Aug 2003 20:40:42 -0400
Received: from adsl-216-101-184-46.dsl.snfc21.pacbell.net ([216.101.184.46] helo=patheticgeek.net)
	by ietf-mx with smtp (Exim 4.12)
	id 19sAq8-0004FP-00
	for pmtud@ietf.org; Wed, 27 Aug 2003 20:40:40 -0400
Received: (qmail 14393 invoked by uid 153); 28 Aug 2003 00:34:28 -0000
Message-ID: <20030828003428.6078.qmail@patheticgeek.net>
From: "Kevin Lahey" <kml@patheticgeek.net>
To: Michael Welzl <michael.welzl@uibk.ac.at>
cc: pmtud@ietf.org, mtu@psc.edu
Subject: Re: [pmtud] MTU IP Option (reviving RFC 1063) 
In-Reply-To: Message from Michael Welzl <michael.welzl@uibk.ac.at> 
   of "21 Aug 2003 09:53:37 +0200." <1061452416.1733.104.camel@lap10-c703> 
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Wed, 27 Aug 2003 17:34:28 -0700

In message <1061452416.1733.104.camel@lap10-c703>Michael Welzl writes
>I have a proposal for PMTUD which could be integrated
>in the "new" as well as the "old" approach. The idea
>is to revive RFC 1063 (the MTU Discovery Option, which
>requires a router to update the MTU value in the option
>if its MTU is smaller than the value already in there).
>Back when this option was conceived, there was no
>PMTUD - so PMTUD was clearly a better idea because the
>option would require all routers to participate.

Whenever this is proposed, a wail seems to go up from the 
routing folks.  The oft-repeated claim is that processing
options on some reasonable percentage of the packets would
be incredably slow.

Similar suggestions were floated in the IPng just a few years
ago, and shot down in flames.  OTOH, I thought I just read
something on e2e suggesting that current routers were 
up to the task, so maybe now is the time to revive the idea.

Good luck,

Kevin
kml@patheticgeek.net

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



From exim@www1.ietf.org  Thu Aug 28 12:22:39 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA27589
	for <pmtud-archive@odin.ietf.org>; Thu, 28 Aug 2003 12:22:39 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sMGh-0007CX-3e
	for pmtud-archive@odin.ietf.org; Thu, 28 Aug 2003 08:52:51 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7SCqpM7027671
	for pmtud-archive@odin.ietf.org; Thu, 28 Aug 2003 08:52:51 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sLiB-0005Fe-U9
	for pmtud-web-archive@optimus.ietf.org; Thu, 28 Aug 2003 08:17:12 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA07134
	for <pmtud-web-archive@ietf.org>; Thu, 28 Aug 2003 08:17:07 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sLiA-00055J-00
	for pmtud-web-archive@ietf.org; Thu, 28 Aug 2003 08:17:11 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19sLiA-00055G-00
	for pmtud-web-archive@ietf.org; Thu, 28 Aug 2003 08:17:10 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sK9c-0006Xj-5A; Thu, 28 Aug 2003 06:37:24 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sGSe-0001Ru-Mu
	for pmtud@optimus.ietf.org; Thu, 28 Aug 2003 02:40:48 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id CAA08032
	for <pmtud@ietf.org>; Thu, 28 Aug 2003 02:40:42 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sGSa-0005AX-00
	for pmtud@ietf.org; Thu, 28 Aug 2003 02:40:44 -0400
Received: from lmr1.uibk.ac.at ([138.232.1.142] helo=smtp.uibk.ac.at)
	by ietf-mx with esmtp (Exim 4.12)
	id 19sGSa-0005AG-00
	for pmtud@ietf.org; Thu, 28 Aug 2003 02:40:44 -0400
Received: from lap10-c703.uibk.ac.at (lap10-c703.uibk.ac.at [138.232.65.57])
	by smtp.uibk.ac.at (8.12.9/8.12.9/F1) with ESMTP id h7S6eCWj029964;
	Thu, 28 Aug 2003 08:40:13 +0200
Subject: Re: [pmtud] MTU IP Option (reviving RFC 1063)
From: Michael Welzl <michael.welzl@uibk.ac.at>
To: Michael Richardson <mcr@sandelman.ottawa.on.ca>
Cc: mtu@psc.edu, pmtud@ietf.org
In-Reply-To: <10074.1062033700@marajade.sandelman.ottawa.on.ca>
References: <10074.1062033700@marajade.sandelman.ottawa.on.ca>
Content-Type: text/plain
Organization: University of Innsbruck
Message-Id: <1062052772.1758.40.camel@lap10-c703>
Mime-Version: 1.0
X-Mailer: Ximian Evolution 1.2.2 (1.2.2-4) 
Content-Transfer-Encoding: 7bit
X-Spam-Score: -7.4 () IN_REP_TO,QUOTED_EMAIL_TEXT,RCV_UIBK,REFERENCES,REPLY_WITH_QUOTES,USER_AGENT_XIMIAN
X-Scanned-By: MIMEDefang 2.35 at uibk.ac.at
Content-Transfer-Encoding: 7bit
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: 28 Aug 2003 08:39:33 +0200
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit

> >>>>> "Kevin" == Kevin Lahey <kml@patheticgeek.net> writes:
>     >> I have a proposal for PMTUD which could be integrated in the "new" as
>     >> well as the "old" approach. The idea is to revive RFC 1063 (the MTU
>     >> Discovery Option, which requires a router to update the MTU value in
>     >> the option if its MTU is smaller than the value already in there).
>     >> Back when this option was conceived, there was no PMTUD - so PMTUD was
>     >> clearly a better idea because the option would require all routers to
>     >> participate.
> 
>     Kevin> Whenever this is proposed, a wail seems to go up from the routing
>     Kevin> folks.  The oft-repeated claim is that processing options on some
>     Kevin> reasonable percentage of the packets would be incredably slow.

As I said earlier, I have some data from measurements using
NOP options, so "incredibly" could be put in numbers. However,
a statistical analysis is missing - a student is currently
going through the data, so I can't tell how long it's going
to take. Then, it would be possible to calculate / estimate
whether it's worth the disadvantage of slow path processing
or not. I would have proposed this idea a little later if
Matt hadn't started the working group...

I decided to give you an idea of the results by putting
two preliminary diagrams online; please look at:
http://www.welzl.at/tmp/option.html
for more. As you can see, the impact of option processing
is considerable, but clearly an order of magnitude smaller
than the RTT itself.


>   Is there some way that we can put in a option that is not a hop-by-hop
> option, but that things that have MTU-constraints might be willing to process?

This would be great, but the only way I can think of (given that
ANY kind of option processing puts your packets in the slow path,
and there are no additional bits available in the IP header) is
by introducing an extra protocol (new protocol number in IP) that
would reside in between IP and the transport layer. And I have a
bad feeling about this.

Any better ideas?

Cheers,
Michael


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



From exim@www1.ietf.org  Thu Aug 28 17:48:44 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA21984
	for <pmtud-archive@odin.ietf.org>; Thu, 28 Aug 2003 17:48:44 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sR0o-0004fQ-Qb
	for pmtud-archive@odin.ietf.org; Thu, 28 Aug 2003 13:56:48 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7SHuk0G017940
	for pmtud-archive@odin.ietf.org; Thu, 28 Aug 2003 13:56:46 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sNXK-00035l-SW
	for pmtud-web-archive@optimus.ietf.org; Thu, 28 Aug 2003 10:14:06 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA18117
	for <pmtud-web-archive@ietf.org>; Thu, 28 Aug 2003 10:14:00 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sNXI-00002P-00
	for pmtud-web-archive@ietf.org; Thu, 28 Aug 2003 10:14:04 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19sNXI-00002M-00
	for pmtud-web-archive@ietf.org; Thu, 28 Aug 2003 10:14:04 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sKtm-0002Xo-61; Thu, 28 Aug 2003 07:25:06 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sHQk-00050s-Bu
	for pmtud@optimus.ietf.org; Thu, 28 Aug 2003 03:42:54 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id DAA12614
	for <pmtud@ietf.org>; Thu, 28 Aug 2003 03:42:49 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sHQh-0006N4-00
	for pmtud@ietf.org; Thu, 28 Aug 2003 03:42:52 -0400
Received: from lmr1.uibk.ac.at ([138.232.1.142] helo=smtp.uibk.ac.at)
	by ietf-mx with esmtp (Exim 4.12)
	id 19sHQh-0006Lu-00
	for pmtud@ietf.org; Thu, 28 Aug 2003 03:42:51 -0400
Received: from lap10-c703.uibk.ac.at (lap10-c703.uibk.ac.at [138.232.65.57])
	by smtp.uibk.ac.at (8.12.9/8.12.9/F1) with ESMTP id h7S7gLWj002615;
	Thu, 28 Aug 2003 09:42:21 +0200
Subject: Re: [pmtud] MTU IP Option (reviving RFC 1063)
From: Michael Welzl <michael.welzl@uibk.ac.at>
To: mtu@psc.edu
Cc: Michael Richardson <mcr@sandelman.ottawa.on.ca>, pmtud@ietf.org
In-Reply-To: <1062052772.1758.40.camel@lap10-c703>
References: <10074.1062033700@marajade.sandelman.ottawa.on.ca>
	 <1062052772.1758.40.camel@lap10-c703>
Content-Type: text/plain
Organization: University of Innsbruck
Message-Id: <1062056501.1759.75.camel@lap10-c703>
Mime-Version: 1.0
X-Mailer: Ximian Evolution 1.2.2 (1.2.2-4) 
Content-Transfer-Encoding: 7bit
X-Spam-Score: -7.4 () IN_REP_TO,QUOTED_EMAIL_TEXT,RCV_UIBK,REFERENCES,REPLY_WITH_QUOTES,USER_AGENT_XIMIAN
X-Scanned-By: MIMEDefang 2.35 at uibk.ac.at
Content-Transfer-Encoding: 7bit
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: 28 Aug 2003 09:41:41 +0200
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit

> >   Is there some way that we can put in a option that is not a hop-by-hop
> > option, but that things that have MTU-constraints might be willing to process?
> 
> This would be great, but the only way I can think of (given that
> ANY kind of option processing puts your packets in the slow path,
> and there are no additional bits available in the IP header) is
> by introducing an extra protocol (new protocol number in IP) that
> would reside in between IP and the transport layer. And I have a
> bad feeling about this.
> 
> Any better ideas?

... come to think of it, once the receiver gets some valuable
information from the option, we ought to find a way to transmit
the information back to the sender. RFC1063 has a reply option;
there should be a better way to do it (slow path processing
once again, just for the feedback? nay ...)

We could perhaps put the feedback directly in the transport
layer - attach it to TCP ACKs...

Cheers,
Michael


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



From exim@www1.ietf.org  Thu Aug 28 21:07:21 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id VAA07670
	for <pmtud-archive@odin.ietf.org>; Thu, 28 Aug 2003 21:07:21 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sTgx-0004cd-VB
	for pmtud-archive@odin.ietf.org; Thu, 28 Aug 2003 16:48:28 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7SKmR3c017757
	for pmtud-archive@odin.ietf.org; Thu, 28 Aug 2003 16:48:27 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sSWH-0000Mr-D6
	for pmtud-web-archive@optimus.ietf.org; Thu, 28 Aug 2003 15:33:21 -0400
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id PAA11594
	for <pmtud-web-archive@ietf.org>; Thu, 28 Aug 2003 15:33:16 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sQJb-000232-Ip; Thu, 28 Aug 2003 13:12:07 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sOU9-0005Gv-W5
	for pmtud@optimus.ietf.org; Thu, 28 Aug 2003 11:14:54 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA23072
	for <pmtud@ietf.org>; Thu, 28 Aug 2003 11:14:47 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sOU7-00014t-00
	for pmtud@ietf.org; Thu, 28 Aug 2003 11:14:52 -0400
Received: from oetest.freeswan.org ([205.150.200.166] helo=noxmail.sandelman.ottawa.on.ca)
	by ietf-mx with esmtp (Exim 4.12)
	id 19sOU7-00014q-00
	for pmtud@ietf.org; Thu, 28 Aug 2003 11:14:51 -0400
Received: from sandelman.ottawa.on.ca (marajade.sandelman.ottawa.on.ca [205.150.200.247])
	by noxmail.sandelman.ottawa.on.ca (8.11.6p2/8.11.6) with ESMTP id h7SFEid27469
	(using TLSv1/SSLv3 with cipher EDH-RSA-DES-CBC3-SHA (168 bits) verified OK);
	Thu, 28 Aug 2003 11:14:50 -0400 (EDT)
Received: from marajade.sandelman.ottawa.on.ca (mcr@localhost)
	by sandelman.ottawa.on.ca (8.12.3/8.12.3/Debian -4) with ESMTP id h7SFEiVo018867;
	Thu, 28 Aug 2003 11:14:44 -0400
To: mtu@psc.edu, pmtud@ietf.org
Subject: Re: [pmtud] MTU IP Option (reviving RFC 1063) 
In-reply-to: Your message of "28 Aug 2003 09:41:41 +0200."
             <1062056501.1759.75.camel@lap10-c703> 
Mime-Version: 1.0 (generated by tm-edit 1.8)
Content-Type: text/plain; charset=US-ASCII
Message-ID: <18866.1062083684@marajade.sandelman.ottawa.on.ca>
From: Michael Richardson <mcr@sandelman.ottawa.on.ca>
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Thu, 28 Aug 2003 11:14:44 -0400


>>>>> "Michael" == Michael Welzl <michael.welzl@uibk.ac.at> writes:
    Michael> ... come to think of it, once the receiver gets some valuable
    Michael> information from the option, we ought to find a way to transmit
    Michael> the information back to the sender. RFC1063 has a reply option;
    Michael> there should be a better way to do it (slow path processing
    Michael> once again, just for the feedback? nay ...)

    Michael> We could perhaps put the feedback directly in the transport
    Michael> layer - attach it to TCP ACKs...

  Yes, that's what I proposed in another message.

  Perhaps we can agree that the receiver has some knowledge of the MTU
and can provide it to the sender.

]      Out and about in Ottawa.    hmmm... beer.                |  firewalls  [
]   Michael Richardson, Sandelman Software Works, Ottawa, ON    |net architect[
] mcr@sandelman.ottawa.on.ca http://www.sandelman.ottawa.on.ca/ |device driver[
] panic("Just another Debian/notebook using, kernel hacking, security guy");  [


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



From exim@www1.ietf.org  Thu Aug 28 23:52:12 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA20910
	for <pmtud-archive@odin.ietf.org>; Thu, 28 Aug 2003 23:52:11 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sXyu-0000RF-5q
	for pmtud-archive@odin.ietf.org; Thu, 28 Aug 2003 21:23:16 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7T1NG7K001673
	for pmtud-archive@odin.ietf.org; Thu, 28 Aug 2003 21:23:16 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sUFy-0006Fp-Ng
	for pmtud-web-archive@optimus.ietf.org; Thu, 28 Aug 2003 17:24:38 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id RAA20343
	for <pmtud-web-archive@ietf.org>; Thu, 28 Aug 2003 17:24:32 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sUFw-0007PG-00
	for pmtud-web-archive@ietf.org; Thu, 28 Aug 2003 17:24:36 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19sUFv-0007P9-00
	for pmtud-web-archive@ietf.org; Thu, 28 Aug 2003 17:24:35 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sR5x-0004nW-Ed; Thu, 28 Aug 2003 14:02:05 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sNND-0002Xs-0G
	for pmtud@optimus.ietf.org; Thu, 28 Aug 2003 10:03:39 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA16489
	for <pmtud@ietf.org>; Thu, 28 Aug 2003 10:03:32 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sNNA-0007do-00
	for pmtud@ietf.org; Thu, 28 Aug 2003 10:03:36 -0400
Received: from brmea-mail-2.sun.com ([192.18.98.43])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sNN9-0007dk-00
	for pmtud@ietf.org; Thu, 28 Aug 2003 10:03:35 -0400
Received: from bu-ewat02-01.uk.sun.com ([129.156.199.2])
	by brmea-mail-2.sun.com (8.12.9/8.12.9) with ESMTP id h7SDxWkH019672
	for <pmtud@ietf.org>; Thu, 28 Aug 2003 07:59:33 -0600 (MDT)
Received: from sun.com (watford-109.UK.Sun.COM [129.156.199.109])
	by bu-ewat02-01.uk.sun.com (8.12.9+Sun/8.12.2/ENSMAIL,v2.2) with ESMTP id h7SDxWvm012657
	for <pmtud@ietf.org>; Thu, 28 Aug 2003 14:59:32 +0100 (BST)
Message-ID: <3F4E0AC4.6070100@sun.com>
From: Jeremy Harris <jeremy.harris@sun.com>
User-Agent: Mozilla/5.0 (X11; U; SunOS sun4u; en-US; rv:1.3) Gecko/20030314
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: pmtud@ietf.org
Subject: Re: [pmtud] the DF-bit honouring problem
References: <Pine.LNX.4.44.0308221347120.1275-100000@zippy.psc.edu>
In-Reply-To: <Pine.LNX.4.44.0308221347120.1275-100000@zippy.psc.edu>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Thu, 28 Aug 2003 14:59:32 +0100
Content-Transfer-Encoding: 7bit
Content-Transfer-Encoding: 7bit

Matt Mathis wrote:
> On Thu, 21 Aug 2003, Michael Richardson wrote:
> 
>>As we discussed at the meeting last month, the plpmtud proposal breaks
>>when there are tunnelling systems that do not obey the DF bit. A second
>>order effect is when tunnelling systems do not copy the DF bit to the
>>outer header when they encapsulate.
> 
> 
> ... snip ...
> 
> 
>>The risk to pmtumtud of DF-ignoring systems is that they cause the 
>>algorithm to fail, since all probes would get through. The result is more
>>fragments, and possible much higher changes of failure to communicate.

Why not have the frag-reassembling IP fire back an icmp saying "needed reassembly"?

Cheers,
   Jeremy



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



From exim@www1.ietf.org  Fri Aug 29 19:54:50 2003
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged))
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA19772
	for <pmtud-archive@odin.ietf.org>; Fri, 29 Aug 2003 19:54:50 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sshj-0001Xu-Cw
	for pmtud-archive@odin.ietf.org; Fri, 29 Aug 2003 19:30:55 -0400
Received: (from exim@localhost)
	by www1.ietf.org (8.12.8/8.12.8/Submit) id h7TNUsJi005920
	for pmtud-archive@odin.ietf.org; Fri, 29 Aug 2003 19:30:54 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19sqGe-0003ic-5j
	for pmtud-web-archive@optimus.ietf.org; Fri, 29 Aug 2003 16:54:48 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA04187
	for <pmtud-web-archive@ietf.org>; Fri, 29 Aug 2003 16:54:40 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19sqGc-0006yA-00
	for pmtud-web-archive@ietf.org; Fri, 29 Aug 2003 16:54:46 -0400
Received: from ietf.org ([132.151.1.19] helo=optimus.ietf.org)
	by ietf-mx with esmtp (Exim 4.12)
	id 19sqGb-0006y7-00
	for pmtud-web-archive@ietf.org; Fri, 29 Aug 2003 16:54:45 -0400
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19snvY-0005O8-BO; Fri, 29 Aug 2003 14:24:52 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org)
	by optimus.ietf.org with esmtp (Exim 4.20)
	id 19slc1-0007Ts-IW
	for pmtud@optimus.ietf.org; Fri, 29 Aug 2003 11:56:33 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1])
	by ietf.org (8.9.1a/8.9.1a) with ESMTP id LAA06427
	for <pmtud@ietf.org>; Fri, 29 Aug 2003 11:56:26 -0400 (EDT)
Received: from ietf-mx ([132.151.6.1])
	by ietf-mx with esmtp (Exim 4.12)
	id 19slc0-0006GD-00
	for pmtud@ietf.org; Fri, 29 Aug 2003 11:56:32 -0400
Received: from [205.150.200.166] (helo=noxmail.sandelman.ottawa.on.ca)
	by ietf-mx with esmtp (Exim 4.12)
	id 19slby-0006G3-00
	for pmtud@ietf.org; Fri, 29 Aug 2003 11:56:30 -0400
Received: from sandelman.ottawa.on.ca (marajade.sandelman.ottawa.on.ca [205.150.200.247])
	by noxmail.sandelman.ottawa.on.ca (8.11.6p2/8.11.6) with ESMTP id h7TFuHd04506
	(using TLSv1/SSLv3 with cipher EDH-RSA-DES-CBC3-SHA (168 bits) verified OK);
	Fri, 29 Aug 2003 11:56:23 -0400 (EDT)
Received: from marajade.sandelman.ottawa.on.ca (mcr@localhost)
	by sandelman.ottawa.on.ca (8.12.3/8.12.3/Debian -4) with ESMTP id h7TFuGDQ028427;
	Fri, 29 Aug 2003 11:56:16 -0400
To: Jeremy Harris <jeremy.harris@sun.com>
cc: pmtud@ietf.org
Subject: Re: [pmtud] the DF-bit honouring problem 
In-reply-to: Your message of "Thu, 28 Aug 2003 14:59:32 BST."
             <3F4E0AC4.6070100@sun.com> 
Mime-Version: 1.0 (generated by tm-edit 1.8)
Content-Type: text/plain; charset=US-ASCII
Message-ID: <28426.1062172576@marajade.sandelman.ottawa.on.ca>
From: Michael Richardson <mcr@sandelman.ottawa.on.ca>
Sender: pmtud-admin@ietf.org
Errors-To: pmtud-admin@ietf.org
X-BeenThere: pmtud@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=unsubscribe>
List-Id: Path Maximum Transmission Unit Discovery <pmtud.ietf.org>
List-Post: <mailto:pmtud@ietf.org>
List-Help: <mailto:pmtud-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/pmtud>,
	<mailto:pmtud-request@ietf.org?subject=subscribe>
List-Archive: <https://www1.ietf.org/mail-archive/working-groups/pmtud/>
Date: Fri, 29 Aug 2003 11:56:16 -0400

-----BEGIN PGP SIGNED MESSAGE-----


>>>>> "Jeremy" == Jeremy Harris <jeremy.harris@sun.com> writes:
    Jeremy> Why not have the frag-reassembling IP fire back an icmp saying
    Jeremy> "needed reassembly"?

  What's what I have proposed.

  If the receiver observed that they had to reassemble fragments, they
could figure out an MTU that did work.  (A lot depends upon how the
fragmentation was done. Chop in half, or big/little bit)

  If we were permitted to use the MSS option again, we could use that.

  This does not directly solve the problem of fragmentation *outside*
of a tunnel - it depends on where the ESP (for instance) packet is
reassembled.  For the *road-warrior* situation, the ESP and TCP termination
are co-located, so this can work.

]      Out and about in Ottawa.    hmmm... beer.                |  firewalls  [
]   Michael Richardson, Sandelman Software Works, Ottawa, ON    |net architect[
] mcr@sandelman.ottawa.on.ca http://www.sandelman.ottawa.on.ca/ |device driver[
] panic("Just another Debian/notebook using, kernel hacking, security guy");  [
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)
Comment: Finger me for keys - custom hacks make this fully PGP2 compat

iQCVAwUBP093n4qHRg3pndX9AQFz7gQAhHks9l6mqqnUnqyqvCRDolaPAq5OALri
9TW43qcPRO/cz8tEpIij/i5pV1IpaDriRfjIyL1L6shbXNJwRJRX6nNoyuq7z9/B
gPZzqOWJd3ftEms1g9mxGV1CaXB5FXB8rBElOXKrh6E3ItY11XM/jxefsZ+qYrpd
8LgQEBC6Fmk=
=bW60
-----END PGP SIGNATURE-----

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



