From owner-freebsd-hackers Sat Jun 24 17:15:44 1995 Return-Path: hackers-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id RAA01522 for hackers-outgoing; Sat, 24 Jun 1995 17:15:44 -0700 Received: from gndrsh.aac.dev.com (gndrsh.aac.dev.com [198.145.92.241]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id RAA01516 for ; Sat, 24 Jun 1995 17:15:41 -0700 Received: (from rgrimes@localhost) by gndrsh.aac.dev.com (8.6.11/8.6.9) id RAA14146; Sat, 24 Jun 1995 17:15:29 -0700 From: "Rodney W. Grimes" Message-Id: <199506250015.RAA14146@gndrsh.aac.dev.com> Subject: Re: FreeBSD as a router To: davidg@Root.COM Date: Sat, 24 Jun 1995 17:15:28 -0700 (PDT) Cc: nc@ai.net, hackers@freebsd.org In-Reply-To: <199506242258.PAA00155@corbin.Root.COM> from "David Greenman" at Jun 24, 95 03:58:07 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 1673 Sender: hackers-owner@freebsd.org Precedence: bulk ... > I understand your point - that perhaps pc-route could be used as some sort > of best-case of which to compare to. My point is that I already know that PC > (PCI) hardware is capable of moving packets at full speed. Hopefully when some > of us gets some time, we'll look into improving the bootlenecks. I'm actually > surprised to hear Rod's 400KB/sec benchmark. I was getting better than that > a year ago with a pair of 486/33 boxes w/ISA ethernet cards. If that's all we > can do these days with Pentiums and PCI, then something is very wrong. > The 400KB/sec benchmark is the number from a year ago with the 486/66 boxes and ISA ethernet controllers. This is *not* what I am seeing today, and if I had my house all put back togeather they way it is suppose to be I would produce the new numbers using PCI and 10/100MB/sec cards. Give me another week and I should have the 4 card router back on line and in full swing use so I can produce the numbers using the current state of affairs. I do know that when I was running this at the other place I could route somewhere around 20MBit/sec on 100MBit/sec ether as a minimum for real world usage types of things (ie, routing NFS). I will be happy when we can route at NFS at the 50MBit/sec rate, as then NFS performance through my router will be as good as local disk performance. For that matter, once things are back togeather here, David, you are welcome to drop by and do some testing and looking with me so that the problem areas can be identified. -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation Company Reliable computers for FreeBSD