From owner-freebsd-net@FreeBSD.ORG Wed Feb 1 05:15:22 2006 Return-Path: X-Original-To: freebsd-net@freebsd.org Delivered-To: freebsd-net@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 064D316A423 for ; Wed, 1 Feb 2006 05:15:22 +0000 (GMT) (envelope-from silby@silby.com) Received: from relay03.pair.com (relay03.pair.com [209.68.5.17]) by mx1.FreeBSD.org (Postfix) with SMTP id C170D43D5D for ; Wed, 1 Feb 2006 05:15:15 +0000 (GMT) (envelope-from silby@silby.com) Received: (qmail 37622 invoked from network); 1 Feb 2006 05:15:12 -0000 Received: from unknown (HELO localhost) (unknown) by unknown with SMTP; 1 Feb 2006 05:15:12 -0000 X-pair-Authenticated: 209.68.2.70 Date: Tue, 31 Jan 2006 23:15:10 -0600 (CST) From: Mike Silbersack To: Greg 'groggy' Lehey In-Reply-To: <20060201012011.GP97116@wantadilla.lemis.com> Message-ID: <20060131230540.R47296@odysseus.silby.com> References: <20060201012011.GP97116@wantadilla.lemis.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed Cc: freebsd-net@freebsd.org Subject: Re: Van Jacobson's network stack restructure X-BeenThere: freebsd-net@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Networking and TCP/IP with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 01 Feb 2006 05:15:22 -0000 On Wed, 1 Feb 2006, Greg 'groggy' Lehey wrote: > Last week, at the Linux.conf.au in Dunedin, Van Jacobson presented > some slides about work he has been doing rearchitecting the Linux > network stack. He claims to have reduced the CPU usage by 80% and > doubled network throughput (he expects more, but it was limited by > memory bandwidth). The approach looks like it would work on FreeBSD > as well. I spoke to him and he confirmed. > > He's currently trying to get the code released as open source, but in > the meantime his slides are up on > http://www.lemis.com/grog/Documentation/vj/. Yes, this is my web > site. The conference organizers are going to put it up on their web > site soon, but in the meantime he's asked me to put it were I can. > > Comments? > > Greg The slides alone don't tell much. There seem to be two possibilities - either "channelizing" everything is responsible for the improvements, or the fact that it waits until the socket is woken up to process the packets is responsible for the improvements. I can't understand why the final step involves a userland TCP stack. The rest of the presentation doesn't explain why that is necessary. I'm sure we'll learn more once we see the source. Mike "Silby" Silbersack