Sun and Netscape: Will 1 + 1 = ½?

Until recently, Sun Microsystems hasn’t had much of a strategy to make money directly off of Java. True, being the progenitor of this smoking hot technology has pumped up its stock price, and keeping Microsoft at bay is essential to selling Solaris boxes, the heart of its revenue stream. But until the middle of last year, the company had few Java-based products to sell to end users.

Then Sun bought NetDynamics, a leading application server vendor. This made perfect sense. The latest version of this now Sun-owned product is a showcase for enterprise Java technologies. Release 5 of the NetDynamics Application Server supports Enterprise JavaBeans (EJB), the Java Transaction Service and more. It’s an impressive product.

More recently, Sun announced the details of its alliance with Netscape. Much of Netscape’s software, such as its Web server, LDAP server and certificate server, fits well into Sun’s world. The general idea of creating a complete portfolio of Java-oriented enterprise technologies makes good sense, and this alliance makes Sun a more effective competitor to Microsoft, which is also a good thing.

Yet the most important technology in the portfolio is the application server, and here’s where the Sun-Netscape alliance runs into trouble. The Netscape Application Server is a fine product, but it competes directly with Sun’s NetDynamics Application Server. I’d be willing to bet that Sun now regrets buying NetDynamics only a few months before forming its alliance with Netscape, but the reality remains: For today’s linchpin middleware technology, Sun has two products. What to do?

I’m afraid Sun has chosen the worst of all possible worlds. Rather than bite the bullet and retire one product or recoup some of its investment by selling one off, Sun and Netscape have announced plans to combine the two code bases. This worries me. First, it means that every current customer of both the Netscape Application Server and the NetDynamics Application Server has to be a little nervous. Although I’d expect the combined product to support all the current interfaces -- no serious vendor will intentionally break existing customer applications -- it will likely declare some of them "legacy." This means that building new applications on either of these platforms is dicey until the first combined release, scheduled for next year, actually appears.

The pending code base consolidation should also make new customers nervous. Suppose you’re considering buying either of these two products today. Buying into an infrastructure product that you know is about to undergo substantial change is scary. Who wants to bet a business-critical application, the kind that app servers are meant for, on a moving platform?

Rather than announcing plans to merge the two products, suppose Sun had instead committed to enhance one product with the features of the other. The company would still support and help migrate users of the product not chosen, but there would now be one stable code base for customers to bet on, and one set of application interfaces that were destined for dominance. This may have caused more pain for the people negotiating the Sun-Netscape alliance, especially given its unusual organizational structure. Still, it would likely have been better for both current and new customers.

Application servers are not simple products, and large code bases are difficult to make reliable. Combining two large code bases in what is bound to be a politically complex environment is not an enviable task. Sun’s ambition is admirable, and I think their goal is exactly correct. Offering a complete infrastructure with a Web server, directory and security services, application server and more, is the best way to compete with the similar set of services that Microsoft builds into Windows NT. But they’d get there faster, with less disruption to both current and future customers, by picking one of their two application servers and going with it. As it now stands, there’s a very real chance that, at least for the next year, the Sun-Netscape alliance will make life worse, not better, for customers of both application servers. --David Chappell is principal of Chappell & Associates (Minneapolis), an education and consulting firm. Contact him at