Event
Summary
On February
17th, Microsoft delivered Windows 2000, the long-delayed, de-featured operating
system successor to Windows NT 4.0. Initial acceptance is expected to be lower
than for earlier releases of consumer operating systems, such as Windows 95.
However, Microsoft officials said nothing had changed in their forecasts for
the product and, indeed, the lowered expectations make it easier for the company
to exceed them. "We don't expect the 'midnight madness' thing of Windows 95
to happen," said Keith White, director of marketing for Microsoft's Windows
division. Still, he added, "large corporations and dot-coms are clamoring for
new technology."
In
addition to the problems generated by the delays, Microsoft finds itself having
to fight off a strong challenge from Linux, now holding the #2 spot (with 25%)
for server OS unit sales. [Source: IDC] Microsoft still retains the lead with
38% of unit sales.
Recently,
Novell Inc. has reported an apparent security hole in the Active Directory feature,
one of the features Microsoft has been promoting heavily. Although Microsoft
originally disputed the report, someone in its organization has now said, in
effect "The problem exists, but we don't believe it is a realistic scenario".
To
compound Redmond's problems, Michael Dell, CEO of Dell Computer Corp., said
he saw increased demand for Linux and predicted that corporate adoption of Windows
2000 will likely be slow. He did some damage control earlier this week, but
partial damage had already been done.
Finally,
reports of 63,000 bugs/defects have appeared. In response, Keith White said:
"We didn't ship this with any show-stopper bugs, which means bugs that cause
lost data or crashes. This product is rock solid."
Market Impact
Anytime Microsoft releases a new operating system, it affects the market in
a big way. Unfortunately for Microsoft, the impact is probably not as great
as it would have been, had it been released when originally promised (1998),
or even last year. Since that time, Linux has gone from being an interesting
"nuisance" for Redmond, to a big player in the Intel server OS market. This
has led to various MS attacks, frontal and otherwise, on Linux. Sensing weakness,
Sun and Novell have launched their own slings and arrows, with MS fighting back.
Although Microsoft is in little danger of losing the Intel market lead, they
are not quite as feared as they have been in the past.
We
do not expect Windows 2000 to increase market growth beyond the current pace.
Because of the expected slower-than-normal adoption rate, we believe it will
not increase consolidation. We believe first-year adopters will be in the 20+%
range, somewhat lower than originally anticipated. In addition, we expect a
large portion of the early adopters to be upgrades from existing Windows installations,
rather than switchovers from Solaris/UNIX/Linux. This will provide Linux (and
Solaris to a lesser extent) a greater opportunity to grab more market share.
Even
with the problems described above, by virtue of Microsoft's market position
and strength, Windows is not yet prey. Helping this along will be Microsoft's
well-oiled marketing/PR machine at full tilt.
User
Recommendations
Two classes of customers will be unaffected by comments here:
1)
Loyal Microsoft customers who "need" to have the latest version no matter what.
2)
Linux/Solaris/Unix adherents and anti-Microsoft customers.
For customers who are planning to change/upgrade to Windows 2000, but not required
to do so immediately, we recommend waiting until at least Service Pack 1 (SP1),
or perhaps SP2. Microsoft OSes historically have not been considered sufficiently
robust until at least SP1, and usually SP2 or later. (Substitute "Service Release
#X" for "Service Pack #X" where appropriate.) Although Microsoft has put significant
effort into testing Win2K, there are still reports of some 63,000 bugs. A high
percentage (30+%) are not true bugs, but the implication is that Win2K is still
not "complete".
In
addition to the issue of bugs, potential purchasers must consider the following:
1)
Desktop systems must meet certain minimum requirements, in excess of NT 4.0
requirements for things like RAM, CPU speed, etc. This may mean clients will
need to upgrade, which can be expensive.
2)
Although Active Directory has (in theory) numerous benefits, it has two big
"gotchas":
A) A potential security hole, as reported by Novell -- This may be fixable
through a patch or hotfix.
B) The probable
need to revamp your internal network -- The second requires time, $$$, and
effort.
3)
License fees can be expensive for the higher-end flavors such as Data Center.
4) Compatibility with existing hardware/applications is not guaranteed, and
needs to be checked thoroughly.
Those
users considering, but not committed, to Windows 2000 for their Intel systems
should also assess Linux. Although we believe Linux still lacks the feature
set of Win2K, and Linux-based applications are far fewer than those for Windows
(other than Win2K), Linux users have generally been happy. On an ironic note,
there may presently be more Linux applications available than Windows 2000-certified
applications.
The
above notwithstanding, beta testers/customers have generally been pretty happy
with Win2K. It is clearly an improvement on NT 4.0, which is itself a pretty
decent OS. We are not recommending customers stay away from it, only that they
delay implementation or proceed with the understanding that it will require
potentially major changes to their infrastructure.
Finally,
although it is probably obvious to most casual observers: Windows 2000 is geared
more toward enterprise/business computing. It should not be thought of as a
consumer OS.