[Tsvwg] Proposed changes to draft-allman-tcp-sack-11

Ethan Blanton <eblanton@cs.ohiou.edu> Thu, 18 July 2002 15:07 UTC

Date: Thu, 18 Jul 2002 09:35:30 -0400
From: Ethan Blanton <eblanton@cs.ohiou.edu>
To: tsvwg@ietf.org
Cc: kfall@intel-research.net, Mark Allman <mallman@guns.grc.nasa.gov>, lwang@netlab.uky.edu, thomas.r.henderson@boeing.com
Message-ID: <20020718133530.GA16116@paco.paco.myip.org>
Mail-Followup-To: tsvwg@ietf.org, kfall@intel-research.net, Mark Allman <mallman@guns.grc.nasa.gov>, lwang@netlab.uky.edu, thomas.r.henderson@boeing.com
Mime-Version: 1.0
Content-Type: multipart/signed; micalg="pgp-sha1"; protocol="application/pgp-signature"; boundary="PNTmBPCT7hxwcZjr"
Content-Disposition: inline
User-Agent: Mutt/1.4i
Subject: [Tsvwg] Proposed changes to draft-allman-tcp-sack-11
Sender: tsvwg-admin@ietf.org
Errors-To: tsvwg-admin@ietf.org
Precedence: bulk

As I think we have all come to agreement on the major points, here is
a summary of the changes I plan to make to this I-D.  Hopefully I'm
covering the salient points here...

1.)  An RTO terminates the recovery algorithm specified in the draft,
     but RecoveryPoint is preserved as-is.

2.)  The recovery algorithm in the draft MUST NOT be triggered after
     an RTO until the cumulative ACK has covered RecoveryPoint.

3.)  Update () MAY be used after an RTO; the reciever SHOULD record
     the new sack information provided by incoming ACKs after a
     timeout in some fashion.  That information SHOULD be used when
     determining what segments to send following a timeout, but the
     selection algorithm for those segments is explicitly unspecified
     by the draft.  (we think a straightforward
     pick-the-first-unSACKed-octets algorithm is acceptable.)

Ethan

-- 
Now if I wasn't such a weenie do ya think you'd still love me,
Pretendin' I'm an airplane on the living room floor?
                -- The Offspring, "I Choose"

[Tsvwg] Proposed changes to draft-allman-tcp-sack… Ethan Blanton