The Geronimo Effect

Opinion: When the Apache Software Foundation's Geronimo Java application server arrives, it could set off a chain of events that change the application server market forever.

You can look at the Apache Software Foundations Geronimo project from two perspectives. For some developers, services firms and customers, Geronimo is the fulfillment of the dream: a commoditized, open J2EE (Java 2 Platform, Enterprise Edition) server market. For some major application server players, Geronimo is the realization of a nightmare: a commoditized, open J2EE server market.

Sure, theres already an open-source (in the Lesser GNU General Public License sense of the term) Java Community Process-certified J2EE application server: JBoss. JBoss effect on the market is already being felt. But Geronimo, which aims to be as good as any commercial or open-source implementation of J2EE, will be licensed under the ASFs terms—which amount to, essentially, "Give us credit, and do whatever you want with it".

Do you hear that giant sucking sound? Thats the air supply for proprietary J2EE implementations being removed from the market.

Sure, Sun has been giving away (as in free beer) the current J2EE reference implementation, aka Java System Application Server Platform Edition (given Suns market share, giving it away was probably not a difficult decision). And Sun is still flirting with open source for Java. But whether Sun ever decides to open source its own J2EE reference implementation or not will become moot—its likely that Geronimo will become the de facto reference implementation.

When Geronimo is fully cooked and certified, any developer will be able to build and distribute a certified J2EE server just by incorporating Geronimo. While the end products might not carry Suns seal of approval, that may not matter a whole lot—all that will matter is that it says "Geronimo Inside." And since Apaches license is a lot less restrictive than the JBoss LGPL, that could make supporting Geronimo a lot more attractive.

It also makes commercially built J2EE application servers look a little less attractive—at least from the standpoint of continuing to develop proprietary J2EE "containers," the guts of Java app servers. They can try to compete based on performance, but over time the Geronimo platform will evolve and erode away that technical advantage—just as Linux and Apaches Web server have done.

But that doesnt mean that BEA, IBM, and Sun are all going to suddenly dry up and blow away. In fact, all three companies have been sliding toward the new center of gravity for the application server market—services—for some time now. Once Geronimo drops, the only value proposition that commercial application server vendors will have left is their reputation for service and support.

But theyll be battling people with the same credentials whove chosen to go to open source. HP recognized this when it shut down its own application server efforts, and started focusing on open source (as it has in its deal with JBoss).

Sun (and perhaps the Java community at large) faces an even bigger problem post-Geronimo. If Geronimo becomes the de facto reference implementation, theoretically companies could use the ASF to do an end run around the JCP with new features—donating significant pieces of code that could break compatibility with the JCP-blessed version. How will Sun, or anyone else for that matter, be able to preserve compatibility once the J2EE genie is out of the bottle?


Check out eWEEK.coms for the latest news, reviews and analysis in Web services.