Re: How to use an IP-header bit for Path MTU discovery.

Philippe Prindeville <philipp@Gipsi.Gipsi.Fr> Mon, 26 February 1990 23:44 UTC

Received: from decwrl.dec.com by acetes.pa.dec.com (5.54.5/4.7.34) id AA02900; Mon, 26 Feb 90 15:44:47 PST
Received: by decwrl.dec.com; id AA23457; Mon, 26 Feb 90 15:44:40 -0800
Received: from [192.33.166.11] by inria.inria.fr (5.61+/89.0.8) via Fnet-EUnet id AA08522; Tue, 27 Feb 90 00:44:03 +0100 (MET)
Received: by Gipsi.Gipsi.Fr (4.12/4.8) id AA20321; Tue, 27 Feb 90 00:45:01 -0100 (MET)
Date: Tue, 27 Feb 1990 00:45:01 -0100
From: Philippe Prindeville <philipp@Gipsi.Gipsi.Fr>
Message-Id: <9002262345.AA20321@Gipsi.Gipsi.Fr>
X-Phone: +33 1 30 60 75 25 / +33 1 47 34 42 74
To: mogul
Subject: Re: How to use an IP-header bit for Path MTU discovery.
Cc: MTU Discovery <mtudwg>

Well, for X.25 (to pick an example, not necessarily my favourite)
one can provide an IP mtu equal to the maximum packet size or
one can use the M-bit to fragment an IP datagram transparently.
For raw HDLC, one doesn't have acknowledgement or sequencing (at
LLC1), so would have to be layered on top.  Of course, with raw
HDLC, one can usually set the MTU arbitrarily large (since this
is typically over private point-to-point lines) so this isn't a
problem...  At least until Frame Relay happens.  Unless you are
referring to LAN technologies (which I feel offer a 'decent' size
of MTU).

-Philip