From owner-freebsd-net@FreeBSD.ORG Fri Mar 26 14:58:58 2004 Return-Path: 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 2E40A16A4CE for ; Fri, 26 Mar 2004 14:58:58 -0800 (PST) Received: from xorpc.icir.org (xorpc.icir.org [192.150.187.68]) by mx1.FreeBSD.org (Postfix) with ESMTP id 1F91443D31 for ; Fri, 26 Mar 2004 14:58:58 -0800 (PST) (envelope-from rizzo@icir.org) Received: from xorpc.icir.org (localhost [127.0.0.1]) by xorpc.icir.org (8.12.9p1/8.12.8) with ESMTP id i2QMwvgd052706; Fri, 26 Mar 2004 14:58:57 -0800 (PST) (envelope-from rizzo@xorpc.icir.org) Received: (from rizzo@localhost) by xorpc.icir.org (8.12.9p1/8.12.3/Submit) id i2QMwv0I052705; Fri, 26 Mar 2004 14:58:57 -0800 (PST) (envelope-from rizzo) Date: Fri, 26 Mar 2004 14:58:57 -0800 From: Luigi Rizzo To: Wes Peters Message-ID: <20040326145857.A52363@xorpc.icir.org> References: <20040326141509.G90406@ganymede.hub.org> <34426.62.242.151.142.1080329152.squirrel@mailbox.wingercom.dk> <200403261425.34253.wes@softweyr.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5.1i In-Reply-To: <200403261425.34253.wes@softweyr.com>; from wes@softweyr.com on Fri, Mar 26, 2004 at 02:25:34PM -0800 cc: Per Engelbrecht cc: freebsd-net@freebsd.org Subject: Re: Looking for switch recommendations ... X-BeenThere: freebsd-net@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Networking and TCP/IP with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 26 Mar 2004 22:58:58 -0000 On Fri, Mar 26, 2004 at 02:25:34PM -0800, Wes Peters wrote: ... > In the Xylan (now Alcatel) second-generation switches (The "X-Frame" > backplane) the switching hardward was capable of switching on the MAC > header *or* other predefined parts of the packet if no MAC header matches > were found. This feature was used to implement hardware routing (the HRE-X > module), allowing us to route packets between IP networks at a million > packets per second. i think you need to tell the full story, such as what was the limit on the routing table, and whether switching packets for which there wasn't a host-specific entry was slower. Finally, cost is not an inessential detail here... I pointed to an L2 switch which can switch around 2.5Mpps and costs Eur 60, retail... An L2 switch has two big advantages over an L3 switch: + only an exact match on the MAC address is necessary, as opposed to the longest prefix match which is required for a router. Surely you need more/different hw to do longest prefix match than the one needed for L2 exact match. Sure, you can install host-specific entries and then use an exact match on those, but the 'miss' case is more expensive, and if you want to do a worst-case rating, then you need to use that number; + in case of a miss, an L2 can flood all ports, a router can't (well, in principle even a router could do that, but i think the reviews wouldn't be so nice if a product did this). So an L2 thing is inherently cheaper as it can play tricks to cut costs down and still behave within the specs. cheers luigi