problems with RFC 1063

Craig Partridge <craig@NNSC.NSF.NET> Mon, 27 November 1989 13:25 UTC

Received: from decwrl.dec.com by acetes.pa.dec.com (5.54.5/4.7.34) id AA05481; Mon, 27 Nov 89 05:25:08 PST
Received: by decwrl.dec.com; id AA07673; Mon, 27 Nov 89 05:25:04 -0800
Message-Id: <8911271325.AA07673@decwrl.dec.com>
To: MTU Discovery <mtudwg>
Subject: problems with RFC 1063
Date: Mon, 27 Nov 89 08:21:09 -0500
From: Craig Partridge <craig@NNSC.NSF.NET>

Hi folks:

    I'd like a fuller discussion of what's wrong with RFC 1063.  Not because
I desperately want to see RFC 1063 endorsed, but because I'd like to understand
what is wrong with it.

    The only problem I'm aware of is that, in principle, gateways don't
like options.

    I have two problems with that objection:

	(1) the option doesn't appear very often (I think we suggested
	once every 100 segments -- I've been convinced that we could
	make that number much higher like in the 1,000s).  So you don't
	take an option processing hit too much...

	(2) by doing the occasional option processing, you should save on
	fragmentation processing.....  I don't think we're going to
	continue to see people willing to live with 576 forever...

Craig