Re: MTUs on different paths

Philippe Prindeville <philipp@gipsi.gipsi.fr> Tue, 12 December 1989 13:52 UTC

Received: from decwrl.dec.com by acetes.pa.dec.com (5.54.5/4.7.34) id AA13434; Tue, 12 Dec 89 05:52:33 PST
Received: by decwrl.dec.com; id AA01718; Tue, 12 Dec 89 05:52:26 -0800
Received: from gipsi.gipsi.fr by inria.inria.fr (5.61+/89.0.8) via Fnet-EUnet id AA08185; Tue, 12 Dec 89 14:51:32 +0100 (MET)
Received: by gipsi.gipsi.fr; Tue, 12 Dec 89 14:52:05 -0100 (MET)
Date: Tue, 12 Dec 89 14:52:05 -0100
From: Philippe Prindeville <philipp@gipsi.gipsi.fr>
Message-Id: <8912121352.AA02310@gipsi.gipsi.fr>
Phone: +33 1 30 60 75 25 / +33 1 47 34 42 74
To: art@sage.acc.com
Subject: Re: MTUs on different paths
Cc: MTU Discovery <mtudwg>

When TCP receives a connection request, is it compelled to
complete the connect (SYN-ACK, and all that) at the same TOS?
If so, then the piggy-backing problem isn't (a problem), since
the information will be symmetrically conveyed...

-Philip