Date: Wed, 03 Sep 97 18:02:08 -0700 From: "Studded" <Studded@dal.net> To: "chat@FreeBSD.ORG" <chat@FreeBSD.ORG>, "Sean Eric Fagan" <sef@Kithrup.COM> Subject: Webserver/OS review Message-ID: <199709040102.SAA15402@mail.san.rr.com>
next in thread | raw e-mail | index | archive | help
On Wed, 3 Sep 1997 16:40:09 -0700 (PDT), Sean Eric Fagan wrote: >> http://techweb.cmp.com/internetwk/reviews/reviews.htm >> >> Comparsion of FreeBSD, SCO's Enterprise Server, Microsoft's Windows NT, >>Red Hat's version of Linux, Berkeley Software Design Inc.'s BSD/OS 3.0. > >An interesting article. FreeBSD does not come out as well as I would have >liked, though -- better than NT, but not as good as BSD/OS or Linux. (They >don't give any numbers for SCO.) > >The numbers they do give for Linux are surprising, in fact -- it would seem >to indicate that Linux is considerably ahead of FreeBSD 2.2.2 in terms of >performance as a Web server, even with several hundred "simultaneous" >connections. I also appreciated this article, thanks to the original poster. The article made a very important point. They did not make changes to the settings (kernel compile or otherwise) for anything that didn't generate an error message. One of the things I learned when setting up the FreeBSD systems that my two servers (and in fact most of DALnet now :) run on was that FreeBSD will chug happily along with suboptimal configurations right up to the point where it dies a quick, painful death. You can look at this as a bug, or a feature.. it depends on your perspective. I will say though that when I asked for help, it was forthcoming from the FreeBSD community (and one member of the core team in particular) in spades. They pointed out that FreeBSD was way ahead of the pack right up to the 100 user mark, then it started to fall off. I would have predicted something similar based on our experiece. If they had taken the test up an order of magnitude and simulated several *thousand* users instead of several hundred, they would have seen linux crash and burn in a fairly spectacular manner. :) We used to have several systems on our network using various flavours of linux. All but one of them have switched to FreeBSD as a result of our success. Our ircd software doesn't scale well on solaris, so I can't really comment on that platform, but I know that our FreeBSD servers beat our BSDi servers hands down, even when the BSDi platforms have superior hardware. The big lesson I got from this is a reinforcement of something that I already knew was in the works, namely helping FreeBSD scale a little more gracefully from a light -> heavy user load (with the steps in between obviously). It *is* possible to get a dynamite heavy load system out of FreeBSD by tweaking maxusers, nmbclusters, and a few other little knobs and buttons, but it's still too arcane a process for the non-programmer who can't dig into the source code and "just see" what's wrong and how to fix it. I know that -current has a lot of motion in this direction, but if we really want a product that is "consumable" for a more intermediate target market, more needs to be done. Doug PS, I finally got around to making a little more "exciting" .sig, guess this was a good day for it. *Big Grin* *** Proud operator, designer and maintainer of the world's largest *** Internet Relay Chat server. 4,168 clients and still growing. :-) *** Try spider.dal.net on ports 6662-4 (Powered by FreeBSD) *** Do thou amend thy face, and I'll amend my life. - Henry V
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199709040102.SAA15402>