April 5, 2001

Linux catching Windows on all fronts

Author: JT Smith

- by Jack Bryar -
Open Source Business -

Linux stocks are down 90 to 95%. Most of the major Linux developers have yet to
a distribution that makes use of the 2.4 kernel. The one company that
SuSE, got slapped around in the tech press. Nothing, not even Comdex
re-focus a business press fixated on the collapse of the dot-com and
financial bubble.And yet--- something is happening out there.

Linux has won the Web server marketplace. It's making serious
inroads into LAN environments. Independent industry analysts are putting out
some surprising projections about the Linux installed base. The new version of Gnome looks good. Could Linux really take on Microsoft?

Eric S. Raymond thinks so. He recently told ComputerWorld
that he thought that Linux would emerge as a viable Windows alternative
before the end of next year. In fact, he worried that Microsoft's "desktop
monopoly will break due to operating system pricing issues before Linux
is really ready" to replace it. Could Linux actually take on Microsoft to
become the dominant platform over the next three to five years?

Admittedly Raymond is not the most objective evaluator of the computing
But consider these facts:

One battle is already won.

Microsoft has lost the ISP market. Research conducted by Idaya
indicated that the Linux growth rate among ISPs and corporate
webmasters would easily match the pace set last year. The research indicated that
ISPs expected that the Linux market would grow by 150% this year,
slowdown or not. If anything the study suggested that growth rate would
accelerate as distributions based on the 2.4 kernel came on the market.

If those numbers are even close to being accurate, it means that
Linux will have become the de facto webserver standard by mid 2002. Forty-five percent of ISP managers told Idaya's researchers that they expected Linux to be the
underlying technology for Web hosting services by 2003.

Another battle is being joined.

Two thirds of Idaya's correspondents also thought that Open Source
software was scalable for enterprise applications. Given the rising profile of
administrators who cut their teeth on web applications, that's a significant

Surprisingly, Linux already controls better than a quarter of server
market, according to IDC. And that number is likely to accelerate. The
support of Linux by major hardware vendors like Oracle, Dell, Compaq,
HP and IBM may not have helped their individual bottom lines, but it has
given Linux credibility in the corporate market as a platform that's here to

The local network environment is critical to both Microsoft and
Linux developers. Yet Microsoft could lose its lead here. Although Microsoft
has spent a lot of cash promoting itself as an enterprise solution, and
has improved the quality of its offerings, the "improvements" in
Windows 2000 have been far less significant than the changes to Linux over the
last year. The 2.4 kernel's journaling file system, volume management
and large memory handling abilities finally make the OS a viable, even
compelling alternative to 2000 or XP. Although the first 2.4 version of SuSE's
distribution was considered a bit disappointing, it's likely that at least some of
the 2.4 based offerings from Red Hat, TurboLinux and Caldera will win
significant chunks of the LAN and enterprise market.

Microsoft is already feeling the pinch. The company had counted on
revenue from servers to offset flattening desktop sales. According to
Information Week, as many as two-thirds of IT executives said last fall they had planned a Windows 2000 server upgrade this year. So far those numbers aren't
materializing. The company has encountered stiff resistance in the marketplace. Sales
of Windows 2000 Advanced Server and DataCenter Server, and NT
Enterprise Edition accounted for only about 8% of the estimated 4 million servers
shipped last year. Sales of Windows 2000 DataCenter Server, in
particular, were "insignificant" according to Microsoft officials. Steve Ballmer
told Information Week he's "not panicking," but admitted, "We've got work to

The response, so far has been typical Microsoft. The company wants
to reduce the functionality of its successor platform, the new Whistler
Server, which means that customers will have to "upgrade" to the Advanced
Server to get the same functionality they already have ... more than tripling
the per unit cost. This strategy may have worked when Microsoft was cheaper
than comparable offerings from IBM and Sun, but will it work if Linux
emerges as a significant alternative?

And elsewhere ...

Engineering and other high-end systems environments are leading a
to Linux across the enterprise. John Cooley followed up a recent EE
Times column complaining about Sun by expressing surprise about how many
system administrators wrote to him stating they had shifted big chunks of
their enterprise environments over to Linux. He mentioned the example of
graphics developer Nvidia. Over the course of the last 20 months, the company had
grown. So had their IT infrastructure. Nvidia had nearly doubled its
Unix (Sun) boxes, cut the number of NT machines by a third and increased the
number of Linux machines 60-fold, and it has become the
dominant platform. The changeover to Linux may be even more complete once Red
Hat ships a 2.4-based distribution, Cooley was told. Nvidia may be
exceptional, but it isn't isolated. Academic and engineering environments, in
particular, are becoming Linux strongholds.

The desktop battle to come

As interesting as these market wins may be, the desktop is still
king. And core applications such as Microsoft's Office suite are critical to
success there. This is the reason that the tech press continues to
feature columnists like Sm@rt Partner's Steven J.
Vaughan-Nichols, who advise people not to bother with Linux on the desktop, (although
he's using KDE 2.0 on Caldera). Why? In a word, MS Word, and the other
extensions of the Windows Office suite are too ubiquitous. You can't fight a

That could change as well. Because Win XP simply can't run on most
home machines, or many corporate boxes, and it can't run current software
all that well. The RAM requirements (256 MB) and clock speed requirements
are just too great. Run XP and forget about workers who telecommute.
Further, many programs written to Win9X, 2000 and NT run in "compatibility mode," i.e., in a relatively bogus emulation of the older operating systems. I can get
that running WINE, thank you very much.

Corporate customers ought to find this lack of effective
functionality, paired with the company's shift to an ASP model, to be a real break
point. It's one thing to have one's data effectively locked up in Microsoft
applications, it's another thing entirely to admit that and cede control permanently
to Redmond. There should be a lot of soul searching in the corner
offices of corporate America (or Canada, or Australia, or the UK) before making
that kind of commitment in the midst of an economic downturn.

All the parts come together (or not)

What happens on the desktop will affect the future of Linux. Despite
the marketing message being sent out by systems developers about handheld
and embedded devices, the desktop is critical to the future.

The analysts sense this, too. It is one of the reasons for negative
sentiment about Palm. Sure, the Palm OS has attracted a reported 100,000-plus
developers. It's also extending its reach into next-generation wireless telephone
devices. Despite this growth, analysts worry that the company could get
whipsawed. The company is being challenged by two platforms that can extend
applications from the PC to the handheld and other, middle range and specialty
devices. The core of Microsoft's strategy over the next few years is to position
the handheld device as a detached peripheral -- a traveling extension of
the desktop -- a device that works fine by itself, but works better if
it can talk to the home and office PC. It's a good strategy. Palm is
vulnerable to it.

Linux is not. It is the only other platform that extends from the
handheld to the desktop to the server. It has an advantage among ISPs,
a growing advantage in the enterprise server space and a small leg up
over Microsoft in the handhelds market. If it can get a toehold on the
desktop, it could emerge as a credible alternative to the Windows marketplace.

As Eric Raymond suggested, there's a window of opportunity here,
courtesy of Linux momentum and Microsoft pricing and design decisions. But that
window is likely to be short-lived.

NewsForge editors read and respond to comments posted on our discussion page.


  • Open Source
Click Here!