From owner-freebsd-current@FreeBSD.ORG Tue Mar 2 07:35:42 2004 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 5FE0916A4CE; Tue, 2 Mar 2004 07:35:42 -0800 (PST) Received: from guns.icir.org (adsl-68-76-113-50.dsl.bcvloh.ameritech.net [68.76.113.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id DAAF443D31; Tue, 2 Mar 2004 07:35:41 -0800 (PST) (envelope-from mallman@guns.icir.org) Received: from guns.icir.org (localhost [127.0.0.1]) by guns.icir.org (Postfix) with ESMTP id A84E977A6FA; Tue, 2 Mar 2004 10:35:40 -0500 (EST) To: Andre Oppermann From: Mark Allman In-Reply-To: <4044928C.AF49FD38@freebsd.org> Organization: ICSI Center for Internet Research (ICIR) Song-of-the-Day: Blow Up the Outside World MIME-Version: 1.0 Content-Type: multipart/signed; boundary="=-=-="; micalg=pgp-sha1; protocol="application/pgp-signature" Date: Tue, 02 Mar 2004 10:35:40 -0500 Sender: mallman@guns.icir.org Message-Id: <20040302153540.A84E977A6FA@guns.icir.org> X-Mailman-Approved-At: Wed, 03 Mar 2004 04:46:37 -0800 cc: Wes Peters cc: freebsd-current@freebsd.org cc: freebsd-net@freebsd.org Subject: Re: My planned work on networking stack X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: mallman@icir.org List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 02 Mar 2004 15:35:42 -0000 --=-=-= > TCP buffer sizing involves mainly two areas. One is good RTT > measurements to be able to estimate the bw*delay product well and the > other is information about memory (mbuf) usage in the networking > system to do the right thing if memory gets low. Why try to measure the bw*delay? Why not use the trick from PSC's autotuning paper whereby you just try to ensure that the socket buffer size is always some multiple (2-4, I think) of the congestion window? I.e., so the congestion window dictates the performance and the socket buffer is not a factor. Of course, you have to figure out what to do to all the connections when there is not enough memory for such socket buffer sizes. But, fundementally, that seems like a much better approach to me. And, thanks for taking this all on! It sounds wonderful! allman -- Mark Allman -- ICIR -- http://www.icir.org/mallman/ --=-=-= Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.3 (FreeBSD) iD8DBQFARKnMWyrrWs4yIs4RAuq3AJ9Tu73LDANRPd/Lgz2lbBf0Umm11wCghIwK 4cj9M/HPJMuhPbHtAZw2t7I= =avhH -----END PGP SIGNATURE----- --=-=-=--