From owner-freebsd-questions Sat Mar 7 00:51:38 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id AAA12267 for freebsd-questions-outgoing; Sat, 7 Mar 1998 00:51:38 -0800 (PST) (envelope-from owner-freebsd-questions@FreeBSD.ORG) Received: from scanner.worldgate.com (scanner.worldgate.com [198.161.84.3]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id AAA12259 for ; Sat, 7 Mar 1998 00:51:33 -0800 (PST) (envelope-from marcs@znep.com) Received: from znep.com (uucp@localhost) by scanner.worldgate.com (8.8.7/8.8.7) with UUCP id BAA02723; Sat, 7 Mar 1998 01:51:29 -0700 (MST) Received: from localhost (marcs@localhost) by alive.znep.com (8.7.5/8.7.3) with SMTP id BAA17775; Sat, 7 Mar 1998 01:45:24 -0700 (MST) Date: Sat, 7 Mar 1998 01:45:24 -0700 (MST) From: Marc Slemko To: Webmaster cc: freebsd-questions@FreeBSD.ORG Subject: Re: Webserver benchmarks? Ultra 2 vs. FreeBSD In-Reply-To: <19980306164853.AAA27430@stimpy> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Fri, 6 Mar 1998, Webmaster wrote: > Hello all, > > I am running a Sun Ultra2 with Netscape Enterprise server 3.5.1, also have > P166 running 2.2.5-R and Apache 1.2.4. Both are on 100/Ethernet. > > Is there a good reason why the P166 would bench at about 60 requests/second > and the Ultra2 only about 15 requests/second? These tests were run using 3 > NT workstations as "web" clients. > > My employer insists that this is wrong. I used Web Bench 1.1 from ZDNet. It could be anything. Could be RAM, could be disk, could be server configuration, could be your benchmarking setup. Benchmarking is hard to do right and you can end up generating a lot of data with hard numbers that are meaningless. Ziff Davis has done it enough times, and they write the software. (although you have people in one place writing it, people in another place doing the testing, and people in a third place doing the article normally...) Check the latency for the Solaris box. If you are seeing consistent ~200ms latencies for all requests under all loads, you are probably running into an interaction between NT's and Solaris's interpretations of slow start. If this is the problem, you can "fix" (ie. make it agree with what NT wants) it on your Solaris box somehow, but this is a FreeBSD list and I don't remember what to do. Let me just say that as much as I like FreeBSD and Apache and think they can perform very well when configured correctly, the difference in numbers that you are seeing is not due to any inherent performance difference unless your Ultra is massively underequipped in some way (eg. 8 megs of RAM). -- Marc Slemko | Apache Group member marcs@znep.com | marc@apache.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message