From owner-freebsd-net Wed Mar 8 7: 1:31 2000 Delivered-To: freebsd-net@freebsd.org Received: from quercus.ifn.fr (mailhost.ifn.fr [62.161.215.67]) by hub.freebsd.org (Postfix) with ESMTP id CDA0837B55E for ; Wed, 8 Mar 2000 07:01:26 -0800 (PST) (envelope-from rguyom@mail.dotcom.fr) Received: from pingoo.ifn.fr ([192.9.200.33]) by quercus.ifn.fr with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2232.9) id FLZQQKVR; Wed, 8 Mar 2000 16:01:23 +0100 Received: (from rguyom@localhost) by pingoo.ifn.fr (8.9.3/8.9.3) id QAA10258; Wed, 8 Mar 2000 16:00:05 +0100 Date: Wed, 8 Mar 2000 16:00:03 +0100 From: =?iso-8859-1?Q?R=E9mi_Guyomarch?= To: freebsd-net@FreeBSD.ORG Cc: Jonathan Lemon Subject: Re: TCP performance problems, Linux faster than FreeBSD ? Message-ID: <20000308160003.A10234@pingoo.ifn.fr> References: <20000303165052.C23732@pingoo.ifn.fr> <20000305193317.B22005@pingoo.ifn.fr> <20000305161638.A85421@prism.flugsvamp.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mailer: Mutt 1.0.1i In-Reply-To: <20000305161638.A85421@prism.flugsvamp.com>; from jlemon@flugsvamp.com on Sun, Mar 05, 2000 at 04:16:38PM -0600 Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Sun, Mar 05, 2000 at 04:16:38PM -0600, Jonathan Lemon wrote: > On Sun, Mar 05, 2000 at 07:33:17PM +0100, Rémi Guyomarch wrote: > > So, Linux use the time of the last non-acked packet as a starting > > point for the 200ms timer, but *BSD stacks use a fixed 'clock'. > > This is completely different ! > > > > I put an ugly hack in sys/netinet/tcp_timer.c to try to fix the > > problem. And now I'm getting **way** less duplicate packets. > > Hm, yes. That's the way it works -- BSD uses a fixed clock > with a maximum delay of 200ms and a statistical average of 100ms. > > You may want to try upgrading to 4.0 (about to be released shortly). > Not only does it change the timer behavior to a per-packet (instead > of global basis) but it allows you to tweak the delayed-ACK time > via sysctl. E.g.: > > net.inet.tcp.delacktime: 100 I (mostly) finished the upgrade to -current. And I'm happy to thank you for this idea, it works much better now ! :-) Here's the results I'm getting with various delacktime settings : 100ms -> 48% dups, 4.39 KB/s 200ms -> 6% dups, 6.18 KB/s 250ms -> 3% dups, 6.88 KB/s <-- :-) 300ms -> 13% dups, 4.80 KB/s With 250ms, I'm beating Linux :-) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message