Great Circle Associates Firewalls
(April 1997)
 

Indexed By Date: [Previous] [Next] Indexed By Thread: [Previous] [Next]

Subject: Re: Opinion: Unix vs NT
From: Todd Graham Lewis <lists @ reflections . eng . mindspring . net>
Date: Wed, 23 Apr 1997 09:20:28 -0400 (EDT)
To: Sean McPherson <sean @ ntr . net>
Cc: firewalls @ greatcircle . com
In-reply-to: <199704231222 . IAA19208 @ ha1 . ntr . net>

On Wed, 23 Apr 1997, Sean McPherson wrote (I snipped some of his words):

> Let me toss in a real life example here. I'll note and admit that by no
> means is this site saturating anything, and by no means is this a catch
> all , end all example, but it's real.
> 
> We have 2 exactly identical machines, as far as hardware is concerned. 
> Had a web site running on the slackware box for 5 months.
> Machine was up for months at a time (sounds like a unix box, eh?).
>
> Moved the website to the NT box (moved the IP, DNS name, transparent to
> users).  Machine groans under the load (hey 32 MB is not a lot for NT),
> swaps frequently, and has had to be restarted multiple time this month
> Was told "Buy more RAM". Traffic dropped to 66% of what it was, both in
> number of hits and transfers. 
> 
> NT Just won't cut it for a lot of these people when they have
> to buy more ram, faster processors, and faster drives
> 
> How does NT perform under some conditions? Great! But under others, it
> just flat doesn't perform at all. What Todd said about the curve on
> equivalent PC hardware is true, from everything I have seen so far.

I've heard this too many times for me not to think that it's true.  I've
_seen_ it myself four or five times.  (I'm a full-timer, not a contractor,
so I don't see as much of the world in person as you do.) 

NT's network performance, both in raw networking performance (pps,
goodput, etc.) and in overall terms (resource consumption on the server,
stability) are inferior to most unices, including among others Linux,
FreeBSD, Digital Unix, Solaris, and AIX.

The Unix people have been doing this for a long time, and Microsoft will
not be able to replicate their results in a few years time.  (Notice I
haven't even mentioned security issues.)  I am still perplexed when
confronted by people who seem to be normal, well-adjusted individuals with
a firm grasp of reality who refuse to admit this. 

Is NT's user interface better than Linux or any other Unix?  Yes, by a
long shot.  Are manpower costs lower?  In many cases, yes they are.  I am
perfectly willing to admit these shortcomings of Unix. 

Why won't the NT advocates admit their shortcomings, manifest as they are?
They remind me of _Pravda_ back in the good old days, when every year the
Soviet Union had more grain than the year before, where all the workers
in Russia are happy socialists, and everything is all roses.  Why is it
that the only time Microsoft will admit its OS'es shortcomings is when
they want their customers to upgrade?  Why is it that NT advocates will
only badmouth old versions of NT, and then only when Microsoft gives the
OK?  I must admit that it puzzles me.

Apologies for carrying on a very tired thread.

__
Todd Graham Lewis          MindSpring Enterprises      tlewis @
 mindspring .
 com



References:
Indexed By Date Previous: HTTP put/get restrictions
From: "Robertson, Dan" <robertsond @ espn . com>
Next: Re: scan detectors
From: Adam Shostack <adam @ homeport . org>
Indexed By Thread Previous: Re: Opinion: Unix vs NT
From: Sean McPherson <sean @ ntr . net>
Next: Re: Opinion: Unix vs NT
From: Ken Wilcox <wilcox @ poss . com>

Google
 
Search Internet Search www.greatcircle.com