By Rob Landley (TMF Oak)

In the beginning, Sun Microsystems (Nasdaq: SUNW) was in an enviable position. It recognized the importance of the Internet a decade before the rest of the world, as evidenced by its motto "the network is the computer." It built up an enviable installed base and a strong brand name. It recognized the importance of open standards, and built everything on the rock-solid technical foundation of Unix. Microsoft (Nasdaq: MSFT) huffed and it puffed, but it just couldn't touch them. That's how Sun got where it is today.

But as we move into a post-Microsoft world, things are changing, and Sun hasn't been moving with the times very well. Its competitors aren't offering proprietary solutions anymore; they're offering commodity open-source software like Linux and Apache, on clusters of commodity PC hardware.

If you need an OS that can stay up for months at a time, you don't have to pay for Solaris. Download Linux. It's free, and it outperforms Solaris on Sun's own hardware. (Solaris on Intel (Nasdaq: INTC) is dead and buried. Because of Linux, Sun can't even give it away for free in noticeable volume.) You can even get the complete source code to Linux (also for free) so you don't have to rely on the original vendor to fix any problems you might have. You can hire anybody from anywhere to fix your bugs, support personnel have to fight for your business and don't just answer questions, they provide code patches pronto. Or you could just provide your own support in-house: Having the source code to your software gives you complete control of your business infrastructure in a way proprietary solutions (from Sun or elsewhere) just can't touch.

As for hardware, commodity PCs are cheaper than anything else. For the price of each Sun box you can get dozens of PCs, and run them in geographically
distributed clusters providing not just protection from hardware failures but floods and earthquakes as well. If a box burns out, you pull it and drop in a new one. With a decent load balancing front-end (from any of a dozen vendors) your customers will never notice. Scalability isn't a worry either; to get more speed, buy more boxes. They're so cheap you can have extra PCs plugged in and sitting there idle, just in case.

Or if you like having one box do it all, Linux can run on multiprocessor PC boxes with terrabytes of RAID drive space. You can get monster servers preconfigured, with a service and support contract, from big players like IBM (NYSE: IBM), Dell (Nasdaq: DELL), or Compaq (NYSE: CPQ), or smaller fish like VA Linux (Nasdaq: LNUX) or Penguin Computing.

As for Web serving, Sun's strategy there leaves me mystified. When America Online (NYSE: AOL) bought Netscape, Sun acquired Netscape's Web server software. They immediately renamed it "iPlanet." WHY? Netscape's entire strategy was to give away their Web browser as a loss leader to build their brand name, both to drive traffic to their website and to sell a Web server from the same company that made the Web browser. A Web server is an almost trivial type of program to
write, and the performance is primarily determined by the hardware it runs on.

The market share of Netscape's Web server fell to single digits a while ago. The undisputed leader (with 60% of the market) is the open source "Apache" Web
server. Add in the fact that a Web server is not a very difficult type of program to write (a college student could do a simple one as an in-class exercise, and years ago Sun started running parts of its own website on a Web server it developed in-house, in Java), and just about the ONLY value in Netscape's server software was the brand name. Since Sun threw the brand name out the window first thing, why did they want the software at all?

Another problem is that Sun's management is wasting opportunities left and right, and as competitive as their market has become they really can't afford to miss a trick. Sun's profound inability to understand how to play in a commodity space is incredibly frustrating to watch, with missteps like their laughable "Community" license, and the snubbing of the Blackdown developers when they finally decided to release Java for Linux. On the one hand, only techies really care about this sort of thing for now. On the other hand, who provides the all-important "third-party support"?

Java itself is an excellent example of wasted potential. As potent a force as it is in the computer world, it could have been so much more if Sun hadn't kept it on a short leash. Java could have become much bigger than Sun, but Sun didn't want that. As a result, instead of becoming a ubiquitous advertisement for Sun's technical expertise, Java has retreated into niche markets in the server space, even though its primary strength is portability and security across client platforms. As for adjacent technologies like the HotJava Web browser or Java Web server, they have been eclipsed by newer developments from outside Sun.

Sun's best strategy at this point would probably be to focus on becoming a "solutions" provider, like IBM. Unfortunately for Sun, IBM is already doing it, embracing the commodity hardware and software Sun refuses to touch, and focusing on service revenues, volume, and ongoing relationships with customers. IBM has explicitly targeted Sun's market with these commodity weapons, and Sun hasn't much chance of making a dent in IBM's market with Solaris on Sparc.

The Bull Rebuttal »

 This Week's Duel