From owner-freebsd-net@FreeBSD.ORG Thu Oct 4 01:09:46 2007 Return-Path: Delivered-To: freebsd-net@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id C3E1B16A47B for ; Thu, 4 Oct 2007 01:09:46 +0000 (UTC) (envelope-from pyunyh@gmail.com) Received: from nz-out-0506.google.com (nz-out-0506.google.com [64.233.162.224]) by mx1.freebsd.org (Postfix) with ESMTP id 7D55313C4BA for ; Thu, 4 Oct 2007 01:09:41 +0000 (UTC) (envelope-from pyunyh@gmail.com) Received: by nz-out-0506.google.com with SMTP id l8so4295nzf for ; Wed, 03 Oct 2007 18:09:41 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:received:received:date:from:to:cc:subject:message-id:reply-to:references:mime-version:content-type:content-disposition:in-reply-to:user-agent; bh=pR59+52cPXAR9m6CjoBw9vCbfEgVUTfZgIGvZXy63ro=; b=f9xuoXvXGXOfNd4F+sEsGKit9xdWaZe8edreB+HNMHLesSCpz8Gk1prf+Cg0Ss2HrF0H4gSuD8Wtbj0cv/Bnf2tp7NamiJ/1GmPt2/m1kv4g3Neo3tXMKG40bjvg0YSNPIlN7MWHXsq6fXr0QJXrsGSQ1OdndxUfWy+qTQb5wyE= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:date:from:to:cc:subject:message-id:reply-to:references:mime-version:content-type:content-disposition:in-reply-to:user-agent; b=PR3tNW5BP2o0n/pZm5GKrCFFS8SdHxxCnkBPcN1YDYxaaOVhkSvkPxAsv8ihwCqGdr+lEv1lztVH5mrqtzxGIpimkFjgxsyjdhblWpfn3Eo9rhKozI3c3lff1Xlj4eFuvrkpKn+KRqwgFtwNp5aO6L2UoJtjeFlP23mh5WhG5p8= Received: by 10.114.178.1 with SMTP id a1mr4914081waf.1191460180107; Wed, 03 Oct 2007 18:09:40 -0700 (PDT) Received: from michelle.cdnetworks.co.kr ( [211.53.35.84]) by mx.google.com with ESMTPS id m40sm2317690waf.2007.10.03.18.09.37 (version=TLSv1/SSLv3 cipher=OTHER); Wed, 03 Oct 2007 18:09:38 -0700 (PDT) Received: from michelle.cdnetworks.co.kr (localhost.cdnetworks.co.kr [127.0.0.1]) by michelle.cdnetworks.co.kr (8.13.5/8.13.5) with ESMTP id l9415vfM031078 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Thu, 4 Oct 2007 10:05:57 +0900 (KST) (envelope-from pyunyh@gmail.com) Received: (from yongari@localhost) by michelle.cdnetworks.co.kr (8.13.5/8.13.5/Submit) id l9415rru031077; Thu, 4 Oct 2007 10:05:53 +0900 (KST) (envelope-from pyunyh@gmail.com) Date: Thu, 4 Oct 2007 10:05:53 +0900 From: Pyun YongHyeon To: Cristian KLEIN Message-ID: <20071004010553.GA30781@cdnetworks.co.kr> References: <4703F9C3.2060601@net.utcluj.ro> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <4703F9C3.2060601@net.utcluj.ro> User-Agent: Mutt/1.4.2.1i Cc: freebsd-net@freebsd.org Subject: Re: FreeBSD as a gigabit router X-BeenThere: freebsd-net@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list Reply-To: pyunyh@gmail.com List-Id: Networking and TCP/IP with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 04 Oct 2007 01:09:46 -0000 On Wed, Oct 03, 2007 at 11:21:23PM +0300, Cristian KLEIN wrote: > Hi list, > > A few days ago I tested whether a FreeBSD 7 box is able to handle Gigabit > traffic. So I used a Cisco 7600 and added static routes from the router to the > box and from the box to the router, so that some packets would loop between the > two. Then I externally injected 30Mbps of "ping -f -t 255 -s ", which > should have generated a "maximum" of 3,6Gbps. I then used nload on the box to > graph the bandwidth. > > The box is a Intel Core 2 Duo, with a PCIe re NIC. I used FreeBSD i386 with > polling and fastforwarding. No WITNESS, INVARIANTS or firewalls. > Though RealTek GigE is not best suitable hardware for gigabit traffic handling how about overhauled re(4)? I've tried hard to fine tune the driver and it may have performance improvements over stock re(4). http://people.freebsd.org/~yongari/re/re.HEAD.patch > I was amased to see that injecting 1000 bytes packets gave a maximum throughput > of 650Mbps, while 1400 bytes gave 750Mbps. During both tests one core was 98% > idle, while the other one was more than 80% idle. > > Can anybody point me what the bottleneck of this configuration is? CPU was > mostly idle and PCIe 1x should carry way more. Or is the experiment perhaps > fundamentally flawed? > > Thanks. > -- Regards, Pyun YongHyeon