Re: TCP Checksum Interoperability

Alan Cox <alan@lxorguk.ukuu.org.uk> Fri, 05 April 2002 23:44 UTC

Subject: Re: TCP Checksum Interoperability
To: crawdad@fnal.gov (Matt Crawford)
Date: Sat, 6 Apr 2002 00:44:06 +0100 (BST)
Cc: L.Wood@eim.surrey.ac.uk (Lloyd Wood), tcp-impl@grc.nasa.gov
In-Reply-To: <200204052313.g35NDvZ10445@gungnir.fnal.gov> from "Matt Crawford" at Apr 05, 2002 05:13:57 PM
X-Mailer: ELM [version 2.5 PL6]
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-Id: <E16tdNG-0000uh-00@the-village.bc.nu>
From: Alan Cox <alan@lxorguk.ukuu.org.uk>
Sender: owner-tcp-impl@grc.nasa.gov
Precedence: bulk
Status: RO
Content-Length: 588
Lines: 11

> hop limit and have no connection to time.  A TCP riding on IPv6 may
> receive old segments an unbounded time later without any other
> network element breaking a spec.

This was true with IPv4. A satellite link or a store and forward physical
layer can have multi-second timings anyway. Take RFC1149 for an obvious extreme
example. In the experimental study the packet dispatch time was a relatively low
proportion and the flight time is not visible to the router nor was the
large amount of flapping.

But we know TCP is broken. The late Ian Heavans wrote a nice draft on the 
subject.