 |
Bill Claybrook
Research director
Linux and Open Source Software
Aberdeen |
 |
|
Linux may leapfrog Windows
The long-term viability of Linux is excellent even though there will be some distributors that do not generate large revenues and may even go by the wayside. The success of Linux is measured by the interest of (and revenue generated by) the system vendors--Compaq, IBM, Dell, HP, VA Linux, SGI, and a few others. Sun is absent, of course, since they believe that Solaris offers anything that users want, even on the low end. Sun is apparently willing to continue pouring money into developing Solaris over the long haul. While Sun has a Linux appliance strategy via Cobalt Networks, it does not have a serious Linux or IA-64 strategy.
My feeling is that system vendors such as Compaq, HP, and IBM, which develop and market RISC Unix platforms, will eventually tire of development costs for their proprietary systems and that in seven to eight years, Linux on IA-64 will replace their RISC Unix platforms. I say this because these RISC Unix vendors have developed good Linux/Unix affinity strategies that allow Unix applications to run on Linux (and vice versa), thus paving a migration path from Unix to Linux.
It is conceivable that Linux, coming from the low end, will leapfrog Windows and become the high-end replacement for Unix over the long term. This seems natural because of the aforementioned Linux/Unix affinity and because Unix-based installations are generally very comfortable with Linux. Windows, however, will maintain a powerful position in the midrange. One of the keys to Linux moving toward the high end is the Open Source Development Lab in Portland, Oregon, which provides Linux developers with access to hardware systems for scaling Linux.
The development of Linux over the next two to four years will change. Not to fear, Linux will continue to be an open source project. There is demand, especially from ISVs, for a standard Linux distribution. ISVs have long tired of porting their applications to multiple RISC Unix systems. Even though distributions use the same kernel (possibly different versions of the kernel at any given time), the distributions are not the same. The real question is, how will the Linux community deliver a standard distribution? Red Hat believes that a standard distribution of Linux already exists: Red Hat Linux. Another possibility is to develop a standard Linux distribution through the Linux Standard Base (LSB) process.
One development that may change the Linux landscape a bit is a potential face-off between Red Hat and Caldera. Caldera, which recently completed the acquisition of SCO, is now the largest Linux distributor in terms of revenue (this may change, however, when Red Hat reaches the end of its fiscal year). Caldera appears to be the only roadblock preventing Red Hat Linux from becoming the de facto standard Linux distribution--in case the LSB approach fails to produce a standard distribution. Caldera inherits SCO's large, worldwide reseller channel (more than 15,000 resellers) and professional services/support organizations. Why is Caldera a significant competitor? In 1999, SCO sold almost 40 percent of the Unix licenses worldwide, via its OpenServer and UnixWare Unix operating systems. In 2000, SCO's percentage of Unix licenses sold dropped just below 30 percent, but SCO, and now Caldera, has a huge installed base of SCO Unix users. If Caldera is successful in migrating many of the OpenServer customers to its Linux distribution, OpenLinux, then it will be very competitive with Red Hat.
|
