RE: [TCLUG:7739] NT vs Linux (was Re: [TCLUG:7737] )

["Schoenhofen, Kelly" <KSchoenhofen@Carlson.com>]

There's alot of other baggage to mention as well. 
First off, the benchmarks really aren't technically "wrong". I do believe
they are quoting the second Mindcraft benchmark that MS and Linux people
both supervised. While some of the hardware was stacked against Linux, and
the primary test itself - X-sized peak connections/hour, in the end it's a
moot point. We can rip on the BS of the testing method endlessly, but nobody
cares. 

In those static test conditions, a high end hardware box running NT
Server/IIS was able to consistently outperform a Linux/Apache box. With the
benchmark they used, it indicated that the linux box could run 1/4 of the
entire web, and the NT box could run 1/2th the web, given enough bandwidth.
This is the equivalent of a Ziff-Davis 3d benchmark that shows Brand-X 3D
video card capable of  "800fps in ANY 3D game!" That's how relevant the
benchmark was. 

IMO, the real deciding factor in the testing that MS is extremely quick to
gloss over is the stability factor - I do believe MS had them reboot the NT
server between each and every benchmark. You ask anyone who works in a
lights-out environment, touch-free maintenance (of lack of) on a server is
the #1 cause for a successful technology [in the server room].



-----Original Message-----
From: ^chewie [mailto:chewie@guinness.urw.org]
Sent: Friday, August 20, 1999 8:32 AM
To: tclug-list@mn-linux.org
Subject: [TCLUG:7739] NT vs Linux (was Re: [TCLUG:7737] ) 


On Fri, 20 Aug 1999, Clayton T. Fandre wrote:
Re: http://www.microsoft.com/ntserver/nts/exec/compares/ntlinux.asp

> I never know Microsoft was so good at fiction.

The benchmarking isn't actually fiction.  What they quoted were
documentable statistics.  What they fail to mention is that the
computer configurations used in these tests were set up in NT's favor.