January 24, 2000

Printer ready

TechView:
Linux: A Strategic Choice?

By Gregory Smith

he flurry of mail generated by my last foray into Linux has finally died down, so I'm ready to brave another maelstrom of criticism on my search for the meaning of Linux in large companies.

In November, my quest for the real meaning of Linux lead me to conclude that the open-source operating system is a great tactical product, but for most businesses it simply isn't a strategic choice. None of the responses to that column have really changed my mind.

A couple of readers pointed out one interesting effect of the open-source nature of Linux--one that is indeed a very powerful advantage. By publishing the source code to the entire operating system, the Linux community creates a uniquely strong aftermarket for support services. As with other platforms, Linux vendors sell support contracts for their particular implementations. While reselling value-added features and support to a freely available operating system is somewhat different than what Windows NT resellers offer, for example, it's the potential for fundamentally equal and thoroughly informed support vendors that may finally break the stranglehold that single-source vendors have on businesses.

There's something very disconcerting about what developers and larger companies are saying about Linux. Most of these reports start with a general enthusiasm for the operating system, but they all point to scalability issues and the increasingly fractured Linux landscape as hindrances to large-scale deployment. The scalability issues are manifest. Most developers of enterprise products who are exploring Linux ports complain about process and thread limitations, file descriptor limitations, the lack of large file support, and the scalability of the shared-memory architecture. That deficiencies exist in some critical aspects of Linux isn't necessarily indicative of anything other than an operating system in its adolescence. With care and feeding, this teen-ager could grow into a very reliable adult.

What concerns me is who's going to nurture this thing. Certainly, the community development efforts that have brought Linux to the forefront have done an admirable job, and given the status quo, I'd expect Linux to continue down this promising path. But as more developers look to Linux as a host for their enterprise applications, more pressure will be brought to bear on these problems. That may be a good thing, but the corporatization of Linux might well be its undoing. Linux's core strengths flow largely from the interest of far-flung programmers. As heavyweight developers begin to steer development of Linux to address their business needs, I fear they will alienate too many of the contributors to Linux that made it so successful. Should that happen, Linux may be lost to the misfortunes of OS/2 or an even more fractured code base than Unix spawned during the 1990s.

None of this naysaying is a fait accompli. But looking at the near-term benefits of deploying Linux as a strategic initiative as opposed to the relative dearth of important management applications and the potential for a stunting of Linux's growth, I wouldn't bet the farm on Linux quite yet. Of course, if I'm wrong, then Linux will soon coalesce as a mature platform for big companies--and I'd love to start 2001 by eating these very words.

Back to the Columnist page
Send Us Your Feedback
Top of the Page

---