To:
Edward Lewis <lewis@tislabs.com>, "'ietf-provreg@cafax.se'" <ietf-provreg@cafax.se>
cc:
Allison Mankin <mankin@isi.edu>, Scott Bradner <sob@harvard.edu>, Ned Freed <ned.freed@mrochek.com>
From:
Patrik Fältström <paf@cisco.com>
Date:
Thu, 10 Jan 2002 22:02:05 -0800
Content-Disposition:
inline
In-Reply-To:
<v03130302b85f6839e103@[199.171.39.21]>
Sender:
owner-ietf-provreg@cafax.se
Subject:
RE: Requirements Document Update
--On 2002-01-07 10.02 -0500 Edward Lewis <lewis@tislabs.com> wrote: > At 11:48 AM -0500 1/4/02, Hollenbeck, Scott wrote: >> More to the point, I think the intention is to ensure that IETF-specified >> protocol documents don't build on "problematic" transports. Maybe Ed can >> provide more detail based on the exchange he had with Patrik. > > Sorry, I can't. When I asked about the motivation of the new text, I was > told it was to ensure (okay, require) that congestion control was part of > the transport. My guess is that this is some concern that blew through > the IESG in response to some other networking technology, perhaps > telephony related. But I am just guessing. The view in the IESG on new protocols is that _somewhere_ in the protocol stack there has to be congestion control. Either one have to include congestion control in the protocol itself, or one rely on underlying technologies for it. What IESG suggested was to have a requirement that the protocol you are developing must run on a transport that itself have congestion control (like TCP or SCTP), and that way, you don't have to include such things in your protocol. IF you want to also run this protocol on for example UDP, your protocol would end up being much more complicated. I have cc:ed the transport area directors which can correct me if/where I have been wrong in the explanation above. Now, Scott seems to have created a new version, even though I still only see version -03 in the archive. Scott, Ed, can you let me know if you see the new version before me? paf