Is It Time for Linux
May 31, 1999
By Greg Shipley

Original link

 You've heard the stories about Linux at major companies, medium-sized cities and even at NASA. You've read about its use in bizarre clusters that outperform Cray supercomputers. You've seen the manifestation of its computational power in movies like Jurassic Park and Titanic. You have friends who swear by it, and maybe some fellow managers who curse at the very mention of its name. And over the past 12 months, you've witnessed Apple, Computer Associates, Compaq, Corel, Dell, Hewlett-Packard, IBM, Informix, Intel, Lotus, NAI, Netscape, Novell, Oracle, SAP and Sybase launch support for what appears to be the world's fastest-growing OS. But there's still a question with no obvious answer: Is Linux right for your enterprise?

Over the past year we've been working to address this issue. We've conducted in-house lab tests, hounded vendors, interviewed CIOs and kept a hand on the pulse of the Linux community. We've been watching, testing, logging and inquiring about the use of Linux in corporate environments. Our findings? Not only is Linux ready for the enterprise, it currently occupies the enterprise--but not in the manner you might expect. Linux is not powering Oracle databases yet. It doesn't drive the financial services, and it usually doesn't sit at the heart of all system deployments.

Instead Linux currently serves as the Swiss Army knife of networking. In many ways, it's a stable version of what NT aspires to be. Linux has been creeping in from the bottom up--introduced by engineers to solve esoteric tasks in a hurry. From there, it has grown to house printing systems, support custom manufacturing systems, perform traffic graphing and analysis, support custom intranet applications, and function as a platform for Web and e-mail gateways.

Recently, Linux has begun making power moves into areas traditionally reserved for other OSes. Organizations have begun considering Linux for their large-scale Web, database, and file- and print-sharing needs. Science and engineering-related industries have begun replacing high-end Unix clusters with inexpensive but computationally superior Linux clusters.

Why the sudden surge of interest? What has finally made Linux a credible corporate citizen? While we've had our eye on the Linux phenomenon, we've suffered through the same problems corporate America has faced: no 24x7 support, no drivers for storage and network devices, no support by major server vendors, and lack of vendor participation in the large-scale DBMS (database management system) arena. But real solutions to many of these hurdles now exist.

During the past six months, the Linux playing field has been redrawn: Caldera Systems, Compaq Computer Corp., Hewlett-Packard Co., Linuxcare, and Red Hat Software now offer call centers for 24x7 support. The distribution issue is also disappearing. Hardware vendors including Compaq, Dell Computer Corp., HP and IBM Corp. offer Linux as a preloaded option. Informix Corp., Oracle Corp. and Sybase have all ported major components of their product lines to Linux. With the backing of these major players, it's reasonable to wonder if Linux can compete with Microsoft Corp., Sun Microsystems and other enterprise leaders. On some fronts, the answer is yes, on others, no--it depends entirely on how and where Linux is deployed. Moving a company's financial system onto an early beta of Oracle 8 for Linux is a bad idea; using Linux for FTP and Web server farms will not only save you money, but may improve overall performance and reliability.

As Compaq's John "Mad Dog" Hall pointed out to us in a recent interview, there is a distinction between "Linux in the enterprise" and "enterprise Linux." Enterprise Linux is still a little ways off. Hall, senior leader of the Unix Software Group at Compaq, is executive director of Linux International, one of the largest Linux user groups. But even he concedes that lack of robust SMP (symmetric multiprocessing) support, an unpolished clustering technology, lack of a robust 64-bit journalized file system and advanced options for high availability leave it trailing market leaders, such as Compaq's Tru64 and Sun's Solaris, on the high end. However, the current shortcomings won't hold it back indefinitely. Linux's rapid evolution continues, and projects addressing these issues are already under way.

Vendors who ignore the rising tide of Linux eventually will lose out. The real challenge for Linux lies not in its technology, but rather in the evolution of its community of supporters. Linux's success in the enterprise will hinge on how Linux vendors handle standardization, the quality of the support infrastructure, the evolution of education and certification systems, and the methods by which new technology will be assimilated into the various distributions of the OS.

It appears that world-class support is close: The training and certification routes are materializing. But Linux's nature as an open-source offering will make training an ongoing challenge. While some organizations may shy away from Linux until these issues are resolved, the smart ones won't wait to start exploring how Linux can help them today. Deployed for the right reasons for the right tasks, it can be incredibly powerful.