By Stuart J.
Johnston
In fact, most Linux true believers fully expect that Linux will overthrow Windows NT, and
possibly even Unix, in the next few years. Linux has shown some impressive progress. Only
introduced in the early 90s, its adherents claim 7 million copies are already installed in
businesses. In June, a multiprocessor Digital Alpha machine running Linux was listed 315th on
the list of the 500 fastest supercomputers.
Linux, supporters claim, is already more scalable, more reliable, more open, runs on a broader
range of processors, and requires less memory and less processor power to do its own internal
housekeeping functions compared with NT. It has a growing list of major vendors releasing
important products to run on it -- including Oracle, Informix, Netscape, and Corel (see story) Two weeks ago, Caldera announced it is shipping a
version of NetWare for Linux.
It is also largely free, and that is part of the problem for CIOs.
Most CIOs, and most senior IT staffers for that matter, are conservative when it comes to
technology. They have jobs to do, and that is to serve business needs with the appropriate
technology. From my observations and experience, they are wary of promised quick fixes and
superior new technologies.
For most of them, the cliché that "you get what you pay for" rings true. One reason IT decision
makers are leery of Linux is its extremely low price -- about $50 should you choose to buy it
from a vendor like Caldera or Red Hat, or free if you download it from the Web.
A recent InformationWeek Research recent survey found that only 3% of 150 randomly
chosen IT decision-makers have any plans whatsoever to deploy any significant amount of Linux
in their companies in the next two years. IT staffers like to have a big company with a large
support organization that they can turn to when there is a problem. They want someone they can
point the finger at, someone with a large vested interest in fixing the problem immediately.
Microsoft fits that description.
The Linux model of "post a message on an Internet bulletin board and some other enthusiastic
user will respond with a helpful answer within an hour or two" doesn't suit the corporate model
of round-the-clock support. Sure, you can buy support from some of the vendors that sell Linux,
but that is still a far cry from the IBM or Microsoft support model.
Admittedly, Netscape got to where it is today by giving away its browser for free and charging
for support, which is a very similar model to Linux. Netscape even copied some of Linux's
success factors recently, such as releasing the source code for the browser to the public.
In contrast to InformationWeek Research's survey, a February survey by DataPro, a Gartner
Group subsidiary, found that 14% of business have deployed Linux. In a recent E-mail, a Linux
proponent asserted that the reason InformationWeek's findings didn't match DataPro's is
that surveys like ours do not track "paradigm shifts" well. That may be true.
Still, I suspect there is a fundamental problem with the argument that Linux will soon supplant
NT through rapid, grassroots adoption by the rank and file.
Certainly, PCs originally became popular at businesses because of a grassroots movement that
swept them into the office, often in the form of petty-cash purchases, despite strong resistance
from IT. PCs were not networked then, however, and as soon as the PC mavericks started
demanding to hook into the corporate network and link to the company mainframe and its
production data, IT reasserted its role as keeper of corporate access. Today, enterprise operating
systems, such as networking, are a top-down decision, not subject to most grassroots
pressures.
Microsoft learned that fact the hard way with Windows for Workgroups (WfW) in the early
1990s. Because WfW had peer-to-peer networking built in, Microsoft product managers believed
that users would create their own peer-to-peer workgroup networks without IT's oversight. Once
they reached critical mass, bingo, they'd all hook up together on an NT-based network and
Microsoft would depose NetWare, thus becoming the de facto networking standard.
Of course, Microsoft's naive vision never happened. WfW did eventually become popular as the
standard desktop, but that's primarily because Microsoft phased out other versions of Windows
3.X that it licensed to PC manufacturers. But Microsoft's grand vision of a grassroots
groundswell for Windows networking didn't happen that way -- despite the fact that WfW had
some significant improvements to Windows' networking technologies.
And when Microsoft did finally begin to make serious progress in business networking, its
adoption was both slow and top-down all the way.
About a year and a half after the WfW fiasco, Microsoft removed the peer-to-peer networking
client from MS-DOS 6.0 just prior to its release. Why? Microsoft's answer was that IT staffs
asked them to remove it because they didn't want maverick users setting up their own networks!
So the Linux scenario sounds all too familiar and the flaw is the same.
What will most likely happen is that Linux believers in lower-level IT positions will sneak Linux
into a few niche applications areas, but when it comes to running business-critical applications,
whether the apps are available on Linux or not, those platform choices will be high-level
decisions. And those kinds of decisions are made in a very slow, deliberate manner.
Unix took almost 20 years before businesses began to take it seriously. NT has taken five years
so far and, by most analysts' estimates, it will be at least four or five years more before NT
becomes ubiquitous. The technological advances of the 1990s apparently have shortened the
adoption cycle to perhaps 10 years from 20 or 25.
Linux has a long way to go before it begins to rival NT, if it ever does. NT has made only a small
penetration in that space -- despite the fact that Microsoft systems ship on more than 90% of
all desktop computers.
Note that I am not saying that Linux will never beat out NT, only that if it does, it will take much
longer than those who want an instant solution to "the Microsoft problem" can ever imagine.
As the world moves more towards a "Web lifestyle" -- to steal a phrase from Microsoft -- there
are significant opportunities for Linux in the Web server arena. But as far as Linux becoming the
primary system in use at corporate sites on the desktop and server? Anything can happen, but I
suspect that five years from now, NT will dominate the desktop and will share the enterprise
with Unix, not Linux.
Linux: Enterprise
Contender?
uddenly, after being a cult favorite for several years, Linux has apparently broken out and
achieved status as a genuine contender in the enterprise operating systems wars.
Send Us Your Feedback
Top of the Page
Enterprise View
Eye On IT
Internet Zone
The Observer
Upcoming Events
Live Events
This Week's Issue
Free Print Subscription
Subscribe
Special Issue
Current Government Issue
Featured Whitepapers
Featured Reports
Video
