report fragmentation

Craig Partridge <craig@NNSC.NSF.NET> Tue, 28 November 1989 13:24 UTC

Received: from decwrl.dec.com by acetes.pa.dec.com (5.54.5/4.7.34) id AA11965; Tue, 28 Nov 89 05:24:21 PST
Received: by decwrl.dec.com; id AA05045; Tue, 28 Nov 89 05:24:17 -0800
Message-Id: <8911281324.AA05045@decwrl.dec.com>
To: deering@pescadero.stanford.edu
Cc: mogul, mtudwg
Subject: report fragmentation
Date: Tue, 28 Nov 1989 08:15:09 -0500
From: Craig Partridge <craig@NNSC.NSF.NET>

I still have to read Steve's RFC (sorry Steve -- I'm currently limited to
one page memos...) but one other point I believe to be true about RF.

You can't learn the MTU except by trying to exceed it -- in other words,
you gotta cause fragmentation to learn the MTU.  The IP option doesn't do
that.  I can start with 576 IP datagrams and only after the option comes back
do I expand to the MTU size I've learned.  Fragmentation only occurs when
the path changes....  (Note in this case, the RF bit would help).

Craig

report fragmentation Craig Partridge