From owner-freebsd-performance@FreeBSD.ORG Thu Jul 12 16:04:59 2012 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 9278C106566C; Thu, 12 Jul 2012 16:04:59 +0000 (UTC) (envelope-from kob6558@gmail.com) Received: from mail-we0-f182.google.com (mail-we0-f182.google.com [74.125.82.182]) by mx1.freebsd.org (Postfix) with ESMTP id 5AD198FC0C; Thu, 12 Jul 2012 16:04:58 +0000 (UTC) Received: by weyx56 with SMTP id x56so2388641wey.13 for ; Thu, 12 Jul 2012 09:04:57 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; bh=OasibeoZE5t3aAlLHBP002THGK6vqv47092vHuH/Wp0=; b=Utc7PhHe1kcp+i7I9ZBNwtS7Ruku8IZlDIzZ/ygjkXrTEyjiAx59Z77pxDotGb0JYX XDNCGsoQoSzTYhjvh0MdCstA9PGpkw+AmuSjvSfo7lls7FyDdzl22aZQmUS92WsavqSY Da/j6ChYmc28JN+j9J9Akq9UlBQA97xYsZXxiq9gBlDq4RjtXQQoDYMtYW6RvRU6onIz hOMyPil9851+V6F2IFceQ8KHfGVSxue9QhZfXrHm5a2Og4Nc3mQ9MZJ7MIUdtM1oOo/t uy4Z1XxHC/IHIgeI0pwVAM0ki35Yvs7bt+n5whmUMzpC86jcazArbC6vyKVU6oVb7LJp CRrA== MIME-Version: 1.0 Received: by 10.216.4.146 with SMTP id 18mr15911156wej.83.1342109097156; Thu, 12 Jul 2012 09:04:57 -0700 (PDT) Received: by 10.223.88.217 with HTTP; Thu, 12 Jul 2012 09:04:57 -0700 (PDT) In-Reply-To: References: Date: Thu, 12 Jul 2012 09:04:57 -0700 Message-ID: From: Kevin Oberman To: Sami Halabi Content-Type: text/plain; charset=UTF-8 X-Mailman-Approved-At: Thu, 12 Jul 2012 16:33:07 +0000 Cc: Luigi Rizzo , freebsd-net@freebsd.org, Alexander Motin , freebsd-isp@freebsd.org, Jack Vogel , "Bjoern A. Zeeb" , freebsd-performance@freebsd.org Subject: Re: ating 100Gbit transfer rate X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 12 Jul 2012 16:04:59 -0000 On Wed, Jul 11, 2012 at 9:45 PM, Sami Halabi wrote: > Hi, > Thank your for your response. > > i have 2 questions: > 1. can you explain the looping method that allowed you to reach 100GB ? > 2. Alcatel-Lucent is routers are given for research internationally ? or > its locally? what routers we are talking about here and what link do they > have? i appreciatre if you explain more how do these routers saturate 100GB. Sure. You create an LSP with a vrf (routing-instance in Juniper-ese) to place the traffic onto it. The LSP is manually configured at each hop to traverse to the far end router and that one then points back at the input router where it is again reversed back towards the destination. Loop as often as required to saturate the link. (I was tempted to just say "rinse and repeat, but that might not be clear to those not in the U.S.) For information on ANI (and I am not sure if new proposals are being accepted), see: http://www.es.net/RandD/advanced-networking-initiative/ > On Thu, Jul 12, 2012 at 6:43 AM, Kevin Oberman wrote: >> >> On Wed, Jul 11, 2012 at 1:31 PM, Sami Halabi wrote: >> > Hi, >> > >> > We have several boxes using 10G cards and using most of the bandwidth. >> > as a future vision i would like to ask if someone ever combined hardware >> > with freebsd/linux to saturate 100Gbit of traffic. >> > what hardware (server, NICs, platform) and software do you recommend >> > that >> > would allow me to acheive my goal? >> > >> > Is it possible with servers ? or i need dedicated hardware >> > (cisco/juniper/other?) if dedicated hardware needed, i would be glad to >> > hear from your experience what do you recommend in terms of performance >> > and >> > price. >> >> I don't know of any 100GE hardware for any PC, but I may be a bit >> behind on the times. >> >> The way we saturate a 100GE with a FreeBSD (or Linux) system is using >> a 10G transmission stream and loop the data stream over the net using >> MPLS. Works quite well, though no end system ever sees more than about >> 9.9G, the routers do. >> >> We are using Alcatel-Lucent routers at this time for our national test >> network. It is available for research by educational, commercial and >> research organizations for a little longer as a federally funded >> testbed for 100G research. When the funding for that project runs out, >> most of the hardware will be re-purposed and will no longer be >> available for research. gnn@ mentioned it about a year ago and >> suggested that some FreeBSD people might want to submit proposals, but >> I sw no responses. We have tested with Juniper and they will work, >> too. All 100G hardware is just a mite pricey, though it has dropped >> tremendously over the past year and a half and I expect it will >> continue to do so. >> -- >> R. Kevin Oberman, Network Engineer >> E-mail: kob6558@gmail.com > > > > > -- > Sami Halabi > Information Systems Engineer > NMS Projects Expert > FreeBSD SysAdmin Expert > -- R. Kevin Oberman, Network Engineer E-mail: kob6558@gmail.com