I've been out of town for most of the debate about NT and Linux, but I'm
glad to add a few points to follow up...
As a Systems Analyst, one of my tasks is to identify platforms for
networking. I go about this asking two primary questions: which package
contains the features that I need to do the job that I need done today; and
which package stands the best chance of being popular and well supported
five to ten years down the road.
Both NT 4.0 and Linux now contain the bulk of the features that most
Internet mavens require. Both are being supported reasonably well -- albeit
by entirely different means -- by their respective constituencies. And both
carry a significant price tag: either in cash up front or in a steep
learning curve and additional support time. But I expect to see both take a
different path in the coming years and, as much as I like Linux, I suspect
that NT will become the dominant OS for small and medium LANs and the Internet.
There's a phenomena in the industry known as the "First Time Buyer
Syndrome." First time buyers pay more attention to the price tag and aren't
particularly discouraged when they spend hundreds of hours trying to make
their equipment hop, skip, and jump. This phenomena was especially
prevalent in universities and medium-sized companies, where yet today you
can still find support people spending inordinate amounts of time trying to
make disparate systems, most of which have fallen on hard times or have not
been upgraded, function appropriately.
A great deal of market research has shown that second time buyers have, for
the most part, learned their lesson and are willing to pay the added premium
to have off-the-shelf solutions that function out of the box and require as
little management expertise as possible. ("Been there. Done that. Gimme
the big Kahuna please...")
It is usually the case that new networks are being put together by those who
have to convince their bosses and cohorts that the technology is useful in
the first place and who must pin together what they can with far too little
money. These are ideal candidates for my beloved Linux and and clone computers.
But in the not-too-distant future we're going to see corporations, NGOs,
universities, and governments in developing countries latch onto the idea of
networking and the Internet in a big way (just as they have in every other
sector in the so-called developed world). When they finally commit the
resources to connectivity, they won't want to play with marginal products --
they'll go mainstream (most of us won't even be consulted, so there's no use
practicing your arguments now) and buy the popular OS that the VARs and
corporate shops support.
That will be, warts and all, NT. (Even Novell has essentially conceeded
this point and has moved on to enterprise networking and value-added
applications like Groupwise...)
So those with NT skills will be more marketable and those with NT systems
will be able to make better use of the "critical mass" of NT services and
Finally, look who Microsoft has hanging out in their stables: most of the
originators of UNIX and IP are now working in "the belly of the beast"
(making much more money than they ever did in universities...) Microsoft
has made a firm commitment to the Internet and it will not -- cannot -- back
So what would be my advice to someone setting up a new network today?
Put together what you can afford today, be only moderately worried about OS
performance benchmarks (they can shift dramatically in a short time frame),
keep your eye -- and your boss' wallet -- aimed at the easiest-to-manage
product (remembering that the less time you spend pinning together the basic
infrastructure of your network means you have more time to develop added
functionality for the human beings who use it and really prove your
network's worth), and be willing to make the jump early to the popular OS --
late adoptors almost always rue the delay.
In my eleven years of networking PCs I have had to learn a half dozen OSs
and adopt one new outlandish technology every year. (This year its the NC:
network computer -- no brains but lots of well connected friends...) It
doesn't promise to get any easier.
So if Linux is what you can afford, great! But keep your eye on NT (or
maybe even Novell/Oracle <grin>)..
To clairfy the following, NT server and workstation do actually support
scads of TCP connections. Microsoft tried, for a very short time, to impose
a license agreement limiting NT workstation to ten simultaneous connections
but was quickly disabused of this notion by an outraged Internet community.
The folks at O'Reilly and Associates have shown that the kernal for NT
Workstation and Server are essentially the same, only a registry entry
distinguishes one from the other. (See http://www.ora.com)
At 07:56 PM 10/26/96 -0400, you wrote:
>Another thing is that as Dr. Lisse mentionned, Microsoft does not
>follow open standards, so there are other limitations. Recently,
>Netscape complained to the government that Windows limited
>applications to 10 TCP connections max, so that if you choose to run
>server application made by Netscape or someone else, you can have only
>ten clients at a time (more if you use Microsoft's own server
>application because it uses interfaces to the OS that are not given to
>third party developers).
Iowa City, IA
[log in to unmask]
Internetworks in International Development
Cliff Missen's Home Page