Just in case someone didn't see this

mogul (Jeffrey Mogul) Mon, 10 September 1990 18:04 UTC

Received: by acetes.pa.dec.com (5.54.5/4.7.34) id AA00936; Mon, 10 Sep 90 11:04:11 PDT
From: mogul (Jeffrey Mogul)
Message-Id: <9009101804.AA00936@acetes.pa.dec.com>
Date: 10 Sep 1990 1104-PDT (Monday)
To: mtudwg
Cc:
Subject: Just in case someone didn't see this


------- Forwarded Message

From: Greg Vaudreuil <gvaudre@NRI.Reston.VA.US>
Date: Mon, 10 Sep 90 10:10:01 EDT
To: Jon Postel <postel@venera.isi.edu>, IAB@venera.isi.edu
Subject: MTU Discovery to Proposed Standard
Cc: IETF@venera.isi.edu, IESG@NRI.Reston.VA.US

To:   Jon Postel, IAB
From: IESG
Cc:   IETF
Re:   MTU Discovery

The IESG recommends to the IAB that the internet-draft
<draft-ietf-mtudisc-pathmtu-02.txt> be published as a proposed
standard.  This is the work of the MTU Discovery Working group of
the IETF.

This draft specifies a mechanism for discovering the path MTU.  The
mechanism consists of occasionally sending a long datagram with the
don't fragment bit set (DF).  If the datagram is too long, an ICMP
error message "datagram undeliverable due to DF bit set and
fragmentation needed" is returned.

This protocol requires a small change to routers.  This protocol uses
16 bits of a previously undefined header field of the ICMP message to
return the size of the slower link causing the error.  This protocol
will also work with non-upgraded routers with performance at least as
good as current practice.  This draft presents many implementation
techniques to improve performance with old routers.

Status of this Memo

This document defines a proposed standard mechanism for discovering
path MTU. It is the product of the IETF MTU Discovery Working group.
Send comments to mtudwg@decwrl.dec.com.  Distribution of this memo is
unlimited.


Greg Vaudreuil
IESG-Secretary

------- End of Forwarded Message