On Tue, 22 Jun 1999, Tracy Tilton wrote:
> Charles Camp wrote:
> >
> > Just read over article at
> >
> > http://www.microsoft.com/ntserver/nts/exec/compares/ntlinux.asp
> >
> > and thought I would throw it out for whoever wanted to read and
> > comment.
> >
> > Being familiar with the "spin doctoring" of companies and advertisers
> > in general was wondering if anybody with some technical expertise
> > would make comment on how germain all that verbage is to a little user
> > like me with a couple of old boxes tied together with ethernet.
>
> For those of you who are not already familar with the FUD from MicroSoft
> and MindCraft you should visit this page to get the rest of the story;
>
> http://slashdot.org/features/99/04/23/1316228.shtml
>
>
> As for it being germain to what you are doing; that would all depend
> on which story you decide to believe once you have read the article
> at slashdot.org.
Hi
I think it is very simple really, you prove what you set out to prove. The
main problem and the one which microsoft are taking advantage of is the
different ways these to products(since they are comparing IIS with Apache
and not NT vs Linux!) are set up. By default apache is set up very
conservatively, it doesn't cache, it uses very little memory and puts very
little strain on your machine. IIS on the other-hand is a monster of a
program which brings my little 166 to its knees. IIS does a LOT of
caching (just look at the amount of memory it consumes). If you read the
specs of the test they have a "warm-up" period (essentially priming the
cache) and then the test begins.
Under these circumstances IIS can perform 680% better, especially when you
consider that the test is done "under dynamic loads". So at the base of
all this is the problem of how well you've setup your software. Apache can
be told to cache, which under the conditions of the test improves its
performance substantially. A even more impressive approach is to use
Apache and Squid together as an accelerated httpd. In this scenario apache
is run on a port other than 80, apache(proxy server) is run on port 80 and
set as and "accelerated httpd", this way apache does all the caching and
apache just serves it pages, IIS can never dream of being able to out
perform these sort of systems.
The CGI tests are meaningless, the two operating systems are not binary
compatible. (Based on using CGI on Linux and ISAPI on Windows NT Server
4.0)
Another problem is how your Linux box is set up. I won't comment on the
rest of the article because I think it all follows from here.
I think the simplest answer is, you've used the two OSes what do you
think!
--
Clint
