From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Fri, 24 Aug 2001 09:44:48 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Fri, 24 Aug 2001 09:44:26 -0400 Received: from ns.suse.de ([213.95.15.193]:9222 "HELO Cantor.suse.de") by vger.kernel.org with SMTP id ; Fri, 24 Aug 2001 09:44:14 -0400 To: Bernhard Busch Cc: linux-kernel@vger.kernel.org Subject: Re: Poor Performance for ethernet bonding In-Reply-To: <3B865882.24D57941@biochem.mpg.de.suse.lists.linux.kernel> From: Andi Kleen Date: 24 Aug 2001 15:44:29 +0200 In-Reply-To: Bernhard Busch's message of "24 Aug 2001 15:39:15 +0200" Message-ID: User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Bernhard Busch writes: > Hi > > > I have tried to use ethernet network interfaces bonding to increase > peformance. > > Bonding is working fine, but the performance is rather poor. > FTP between 2 machines ( kernel 2.4.4 and 4 port DLink 100Mbit ethernet > card) > results in a transfer rate of 3MB/s). > > Any Hints? Bonding reorders packets, which causes frequent retransmits and stalls in TCP. One setup that doesn't is multipath routing (ip route .. with multiple nexthops over different interfaces). It'll only load balance (srcip, dstip,tos) tuples though, not individual flows, but then it has the advantage that it actually works. -Andi