The big boys are coming to Linux

A quick scan of the players at Japan’s largest Linux
event underscores the degree to which traditional IT developers have
come to embrace and dominate the Linux platform. At the recent LinuxWorld
Expo in Tokyo, the bulk of the North American and Euro-Linux shops
were no-shows. SuSE wasn’t there. Neither was Red Hat or Mandrake. Who
was? Japanese developers such as HDE,
Plat’Home and so was Turbolinux,
certainly. But big-time software and hardware vendors were there in force,
including a lot of names not traditionally connected to Linux or Open Source. The market is changing and this show illustrated the degree to which
the “big iron” manufacturers and enterprise software developers are
beginning to dominate a market and a platform that the hacker community thought
it owned.

For example, one of the biggest news stories in Japan
involved Oracle, and its majority-owned subsidiary Miracle Linux, which
were busy pushing a package consisting of Borland’s Kylix development,
together with Oracle 8i and the Miracle Linux distro. It’s an excellent
package and one that should find a solid market in Japan and elsewhere.
It is the type of enterprise-level platform that is making Linux a
major player in back-end systems environments.

However, the biggest announcements surrounding the
show concerned hardware developers, especially IBM. Hitachi, NEC,
Fujitsu and IBM have agreed to cooperatively develop Linux-based backbone
systems. The technical scope of the announcement and the financial resources
member companies are committing to it are an excellent illustration of why
control of Linux and Open Source development is slowly but inevitably moving
out of the hands of the volunteer hacker community and into the
well-compensated engineering departments of traditional IT market leaders, BMD (before Microsoft dominated).

The announcement involves a solid re-work of many
critical Linux elements. Corporate research teams will focus on
scalability and non-uniform memory access (NUMA) capabilities, all the better to
support the multiprocessor
architecture found in a lot of “big iron.” The group has set a target
date for rolling out code and schematics by mid 2003. IBM is to head up core
code development, while its partners are to be more focused on specific
scalability issues.

Whether the community accepts the “suggestions”
of IBM and its partners, those suggestions will be backed up by an
enormous development budget. That budget works out to “only” $250 million
dollars, rather than the $400 million reported elsewhere, but that investment
represents more than double the total gross receipts of Red Hat over the last 12
months, and represents the kind of resources that can overwhelm independent
development or effective control of the Linux platform.

The marketing types at IBM have been at pains to assure
Linux zealots that this did not represent any sort of cabal to usurp
the hacker community. They were also discounting rumors that Big Blue’s
engineers or sales staff were making negative comments about Linux’s “amateurish”
development process or concerns about the 2.5 kernel and the degree to which it
may or may not meet enterprise system requirements. Still, Dan Powers,
IBM’s Internet Technologies chief told Computer Resellers News that he thinks
Linux is “still an immature operating system” that needs an
enterprise-class upgrade. It’s likely to get it.

It is almost unimaginable that the specifications
coming out of their efforts won’t rapidly become a de facto standard,
even if there is resistance among the Open Source community. It is nearly as
unimaginable that IBM, its partners and a handful of other corporate
developers will fail to get their way as long as they agree on standards and
support each other.

IBM along with a half-dozen other vendors, are just too powerful to have their wishes ignored.

Go ahead — try and ignore IBM. As I noted a couple of
weeks ago, IBM, Dell and Sun hold 86% of the Linux/Unix server market.
One reason: IBM’s powerful position as an integrated supplier. For
example, according to the Gartner Group’s DataQuest survey released this week,
Oracle and IBM jointly control 63% of the database marketplace, even as
Microsoft has leveraged its stranglehold on its own platform to push competitors
out of the NT market. Oracle and IBM now account for well over 80% of
the Unix/Linux database market. Another reason that IBM is essential to the
Open Source movement is the company’s astonishing Services division.
Last year the division did $33 billion in revenue. In an average week, it
will produce more sales than Red Hat, SuSE and VA Linux will generate
together in a year.

The bulk of that revenue is based on enterprise systems
work. One of the happiest discoveries in the Linux community has been
the degree to which enterprise environments are uniquely suited for Open
Source development. Clients and vendors have agreed that they both benefit
from open architecture, even if the resulting code isn’t GPLed. Clients can
tweak software and participate more effectively in development. Vendors
drastically reduce their support costs. Enterprise software is also
maddeningly complex. In most settings that is a negative; code can’t easily be
re-used. But that negative is a benefit for enterprise developers. It means they
won’t compete with their own code somewhere else, and that their clients
will still need their assistance with support and maintenance. This is
one of the most important reasons IBM has supported Open Source. For
IBM, there is virtually no downside.

Other large vendors are using their economic power to drive Linux
development. Many are working with Linux developers to create unique Linux and
Java-based applications on their platforms. Hitachi has set up an alliance with
TurboLinux to develop e-commerce and sales support systems. Sharp is working with
Lineo and a handful of other Linux specialists to jump-start its next
generation of PDAs and related devices, one of which is on display at
the
JavaOne
Conference in San Francisco this week. Sony is working with a handful of Linux and Java developers to morph PlayStation2 into a viable home computer. The
needed hard drive and software should be shipping before fall. This
isn’t really a new idea. Blokman, a Czech firm, has created a Linux distro it
claims runs on the older PlayStation platform. However, the fact that
Sony is committing resources to the effort is what gives the project
credibility.

Each of these vendors has a vested interest in driving the development of
the Linux platform. Some are doing it directly. Others are doing it by
becoming an important customer — sometimes the primary source of revenue for
companies focused on Linux, and that inevitably means that their requirements
become Linux requirements, and their priorities become Open Source priorities.