Linux has come a long way: The once-insurgent operating system is now a solid corporate citizen. And the belt-tightening climate of this year should play to Linuxs strength as a low-cost alternative.
Those who havent considered Linux, or who have given it a chance and found it wanting, should take a look at current distributions; with gains in key areas during the past year, Linux is ready to take on more important and varied roles.
The Linux 2.4 kernel, which shipped just more than a year ago, gives the operating system muscle that was lacking. In eWeek Labs tests, Linux now matches, and in several respects outperforms, Windows 2000 in Web and file serving.
The kernel has had its ups and downs: Memory management problems forced a complete replacement of that subsystem. Such jarring, rip-and-replace episodes might be a thing of the past, however, since additional testing resources are now available to Linux kernel developers.
The business proposition of Linux vendors was in ferment last year, but the tidal wave of cash that flowed into Linux companies in 1999 and 2000 pumped a huge, one-time subsidy into Linux software and application development. The Linux 2.4 kernel, Nautilus file manager and Evolution e-mail client all were major beneficiaries.
This fluctuating financial backdrop has made the support of IBM crucial to Linuxs success. That backing has paid dividends in Armonk as well, as IBMs zSeries system sales grew at double-digit rates last year, thanks in part to the combination of Linux and virtual machine technology.
Forecasts show Linux rising rapidly as the choice of embedded-system developers, whose operating systems are behind Web-based services and in server room appliances.
Linux still isnt strong on the desktop. Productivity applications such as Suns StarOffice 6 and Ximians Evolution are a big step forward, but both need time to mature and to gain credibility with those who deploy and support end-user tools.
Facing the relentless enmity of Microsoft and the apathy if not antipathy of Sun cant be an asset, but Linux is too diffuse to be destroyed or acquired.
Hewlett-Packard, IBM, Compaq and others should be relieved at the prospect of getting out of the business of maintaining a Unix derivative. In five years, we believe there will be only three server operating systems that matter: Linux, Solaris and Windows. And no one will be fired for recommending Linux.