which protocols does MTUD help?

Craig Partridge <craig@NNSC.NSF.NET> Thu, 30 November 1989 12:33 UTC

Received: from decwrl.dec.com by acetes.pa.dec.com (5.54.5/4.7.34) id AA06882; Thu, 30 Nov 89 04:33:26 PST
Received: by decwrl.dec.com; id AA08729; Thu, 30 Nov 89 04:33:23 -0800
Message-Id: <8911301233.AA08729@decwrl.dec.com>
To: sytek!rfox@sun.com
Cc: mtudwg
Subject: which protocols does MTUD help?
Date: Thu, 30 Nov 1989 07:31:07 -0500
From: Craig Partridge <craig@NNSC.NSF.NET>

>  You do bring up a good point: Which packetization protocols are we really
>  trying to help/solve with the mtu discovery? 

TCP -- for obvious reasons

Some UDP-based applications -- for example, NFS *can* adapt to changing
    MTU sizes, if it has some way to learn them (see Nowicki's paper
    in the April '89 issue of CCR).

We may also use MTU discovery to help solve FDDI-Ethernet bridging
problems (mismatched MTUs) -- note this requires the use of an option.
It also isn't high on my list of things we *must* fix.  (But I don't
hold with folks who say let FDDI-Ethernet bridge builders stew).

Craig