A brighter future for Sun One?

Mark Bauhaus, vice president of Java Web Services at Sun, is just the kind of smart, results-oriented executive the company needs. He’s spent the past nine years in Sun’s services division, so he brings a customer and solutions-focused perspective to the company’s rudderless Java efforts. Given enough time, freedom, and exposure, Bauhaus could help turn Sun’s largely uncelebrated customer successes into respect for Sun One as a foundation for enterprise solutions.

Bauhaus genuinely believes that Sun’s server-side software is the best available. He admits that Sun took too long to embrace Web services, despite the company’s crucial involvement in its development. And he’s right when he portrays Java as the biggest gift any vendor ever handed to IT. Bauhaus has the right values and perspective for his job. Unfortunately, we think there is little chance that his hopes for enterprise Java will play out, at least at Sun.

Bauhaus speaks convincingly of Sun’s “teach the fishermen to fish” approach to consulting. He blasts IBM Global Services for engineering solutions that result in dependent customers. Like Sun’s CEO Scott McNealy, Bauhaus blasts the press for missing the point. His unusual perspective on Sun’s server software — an insider who plied Sun One as a solutions toolkit in much the same way IBM uses WebSphere — lends credibility to his core argument: Sun’s server software is already respected, widely deployed in large accounts, and on its way to racking up more big wins. Sun doesn’t need to fix or rescue Sun One, but only to make the product more competitive.

What does Sun need to do to accomplish this feat? The stones most often tossed at Sun One include its cost and locked-in design, and the comparatively poor quality of integration among its parts. Bauhaus wasn’t ready to go public on the pricing of Sun’s soon-to-be-announced Java server bundles, but he repeated other executives’ claim that at least a basic edition of the Sun One application server will be included with new Sun server purchases. We expect Sun’s current approach, which limits the scalability of the bundled edition of its server software, will continue. Customers can develop and test their Java applications for free, but deployment requires a license.

Sun can afford to push the apparent cost of its licenses down. For one thing, much of Sun’s hopes are pinned to the Sun Fire Blade Platform, which stuffs 16 CPUs into a space that previously accommodated three or four. That’s a boon for any application that’s licensed on a per-CPU basis.

The other consideration is that Sun’s revitalization of its branded and partner-based consulting programs shifts revenues using an IBM-like model. IBM can claim that WebSphere’s base price is a mere $7,500. But IBM’s position, which Sun simultaneously criticizes and seeks to imitate, is that services are always part of the sale.

“No customer wants to do this for themselves,” Bauhaus says. The difference, as he describes it, is that Sun’s consultants eventually go home. IBM’s never do. In contrast, McNealy said at the Network Computing 03 Q1 launch that he wants to tell Sun’s server customers, “Keep your hands off.” In other words, Sun will take care of everything in perpetuity.

Sun will address the integration issue as well as it can with its upcoming server software configurations. In this regard, Sun faces the same challenges as competitors IBM, Oracle, and Microsoft. All players fall back to integration through documentation, code samples, and knowledge shared primarily with partners and consultants. However, Sun lacks the one element that unites competitors’ enterprise stacks: an enterprise-grade relational database.

DB2, Oracle, and SQL Server databases are often the wedges used to drive application servers into large accounts. The scalability of a J2EE application is tightly bound to the scalability of the data services on which it depends. Here, IBM, Oracle, and Microsoft can optimize their entire enterprise stacks for their branded relational databases; Sun cannot.

No analyst could write off Sun as an enterprise Java player, but the chances are slim that Sun will steer J2EE mind share (and revenues) away from its own licensees. By declaring its Java server software to be the best and emphasizing how different it is from others, Sun is deflating the primary selling point of J2EE: It’s essentially the same no matter what brand is on it. That’s always been a debatable claim. With Sun’s upcoming changes and the likely responses from licensees, the “write once, run anywhere” contract might finally be voided.