*BSD News Article 78591


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!nntp.coast.net!news.sprintlink.net!news-stk-200.sprintlink.net!news.sgi.com!www.nntp.primenet.com!nntp.primenet.com!cam-news-hub1.bbnplanet.com!cpk-news-feed2.bbnplanet.com!cpk-news-hub1.bbnplanet.com!newsfeed.internetmci.com!netnews.nwnet.net!symiserver2.symantec.com!usenet
From: tedm@agora.rdrop.com
Newsgroups: comp.os.linux.misc,comp.unix.bsd.freebsd.misc,comp.infosystems.www.misc
Subject: Re: Unix too slow for a Web server?
Date: 19 Sep 1996 08:27:57 GMT
Organization: Symantec Corporation
Lines: 58
Message-ID: <51r06d$gol@symiserver2.symantec.com>
References: <323ED0BD.222CA97F@pobox.com> <51oph4$4vj@due.unit.no> <51qa91$89o@manuel.anu.edu.au>
Reply-To: tedm@agora.rdrop.com
NNTP-Posting-Host: 198.6.34.1
X-Newsreader: IBM NewsReader/2 v1.2.5
Xref: euryale.cc.adfa.oz.au comp.os.linux.misc:130026 comp.unix.bsd.freebsd.misc:27506 comp.infosystems.www.misc:43774

In <51qa91$89o@manuel.anu.edu.au>, rxt651@leonard.anu.edu.au (Rohan Tronson) writes:
>Tor Iver Wilhelmsen (toriver@pvv.ntnu.no) wrote:
>: Subhas Roy <subhas@pobox.com> writes:
>: >A ZDnet article says in the web page
>: >http://www.zdnet.com/pccomp/features/fea1096/sub4.html#jump2
>: >that Windows NT-based servers run much faster (as much as 13
>: >times) when client counts are low.
>: >
>: >Is that possible? Anybody wants to comment on the
>: >article's claim?
>
>Ok, these comments are all very well.  However in the field of Chemistry,
>which is my background, if you disagree with a published result, and can
>prove it, you publish your own, contraditory results.  Which it seems is
>what we need to do as a community (with more than just this issue)!
>

The problem with this is that this still won't prove anything to anybody.

I feel compelled to point out here that Microsoft is not making a dime from
Internet server, since a peer-to-peer version is included with NT workstation, and 
a so-called full version is included with NT server.  

The Apache group is also not making a dime off of Apache since it's free.

So, this benchmarking is really meaningless in my mind, since it's basically 2 groups
that are fighting over the market for free software.  I don't know how you all
make money from your businesses, but if I was making all the money off of
IIS and Apache that Microsoft and the Internet groups are, I'd be broke!!  :-)

At any rate, the problem with benchmarking something as complex as a computer
system is that unless _all_ the variables are printed that can affect the results
then they are meaningless.  This goes beyond the old "well what model of
computer did they set up?" nonsense.  For example, how many times on these
benchmarks have you read any statements that discussed network utilization?
Do these benchmarkers even bother to put a packet sniffer on the wire to
see if the network adapter card is bad or something?  Sheesh!

If you really want to settle this debate once and for all the only way to do it
is through a real live contest, not a so-called "impartial" benchmark study.  In
other words, you get a group of pro-NT people, a group of pro-Unix people,
a group of pro-Netware people and a group of pro-OS/2 people all together and 
tell them that they their goal is to set up the fastest server.  Then, you make
them all agree to a common hardware platform, (such as a commercial platform
such as a Gronkulator 500 Sextium Pro 200Mhz machine) and tell them that other
than a standard motherboard, SCSI disk and adapter card, and RAM amount, that
they can pull whatever customization tricks out of their sleeves that they want,
except that they cannot modify the source code of the operating system.

Then you sit back and clock each group and see who is the winner.

Otherwise, if you attempt to set up a so-called "impartial" test without having
the enemy participate, you are going to be wide open to charges of favoritism.


Of course, we all know perfectly well this will never happen, because the
commercial OS people know that their stuff is slower, and they will only
agree to benchmarks where they have control of the competitions setup.