Linux, an NT Killer in the Making

• 09/23/1998

There is a transition happening in the industry where so-called "open source" operating systems and applications are gaining some credibility for corporate use. The past couple of months we’ve seen several major software vendors announcing that they will port their applications to Linux, including the likes of Informix Corp. and Oracle Corp.

The question I keep asking myself is, Why? What’s the real motivation here? Is it because these vendors will support anything that’s not Microsoft-proprietary and has a slight chance to derail Windows NT, or is it because Linux is really a viable alternative to NT and an enterprise-class product on its own merits?

My take is that these vendors are anxious to help Linux along just in case it should prove to have the potential to compete with Windows NT. Yet at the same time they’re afraid to make a big deal out of it lest they incur Microsoft’s wrath -- and be perceived as going into competition with Redmond. Oracle alone could provide the funding and support to make Linux a household word if it wanted to.

Linux is seeing deployment in corporate America for tasks that include Web serving and print and file services, and in some situations, it has been deployed as a workstation operating system for custom applications. The cost is right, yet the deployments have not shown up in a huge variety of situations. This probably has something to do with the fact that many common applications are simply unavailable for Linux. Informix says it will port all of its products to Linux. But the company has not yet made a commitment to exactly when that will happen.

I watched a panel discussion on the subject of Linux in the enterprise, held at the Comdex/Enterprise show earlier this month in San Francisco. The panelists included representatives from several of the larger Linux vendors such as Red Hat Software Inc. (Research Triangle Park, N.C., www.redhat.com) and S.u.S.E. Inc. (Oakland, Calif., www.suse.com), as well as newbies such as Intel Corp., Informix and Oracle. There also were a couple of end users that have used Linux in their corporate projects.

I was disappointed to see that, despite literally sending show personnel out into the halls to recruit attendees to view the session, there were only about 60 people scattered about in the vast room, which was capable of seating at least 1,000 people. Of those, a show of hands requested by one of the speakers found that about half of those people already are Linux users. This tells me that people who don’t use Linux are not terribly interested in learning more about it, while those who do are loyal to it.

There were some interesting points raised in that session, especially the bias that exists against Linux. One of the users who deployed Linux as the client operating system in a data warehousing project admitted that the only way he successfully brought Linux into the project was to do so without telling anybody until it was done. The results were good, but he believed his IT manager would have shot down the use of Linux had he been given the choice. Another user conceded, "If I had to convince you to use Linux, it would be pretty hard. That’s the scary truth."

Representatives from Red Hat Software and S.u.S.E. believe that Linux is about to "cross the chasm" from early adopter phase to mainstream. If it makes it through this transition, it will begin to be viewed as a credible operating system. If not, it could become yet another blip steamrolled into the pavement by Microsoft.

However, it doesn’t have to go that way. Consider the Apache HTTP server. Like Linux, Apache also sports a collaborative development history and is available at a hard-to-beat price. Apache already has been ported to Windows NT, is heavily used by ISPs and is the most widely used Web server on the Internet.

Yet for corporate users it still carries some of the same credibility problems that threaten to strangle Linux, particularly the lack of a large vendor to hold responsible when things go wrong.

Enter IBM Corp. The company will start shipping Apache Web server next month with its WebSphere application server product. When a customer buys a copy of WebSphere, Apache is included at no extra cost, but the customer gets IBM support for both products. If something breaks, a customer calls IBM and IBM fixes the problem.

This model has to be more attractive to IT managers than the prospect of posting a problem on a Listserv and hoping for a solution to appear via e-mail. The WebSphere/Apache combination will be available for AIX, OS/2, Sun Solaris and Windows NT.

What if a major player such as IBM or Sun Microsystems Inc. was to sell support services for Linux? Could Linux really grow into a viable alternative to NT? Are IT managers really interested in having yet another emerging operating system within their corporate environments? Hard questions to answer. If you have an opinion on the subject, let me know.