The news that Covalent Technologies is offering commercial support for Apache Geronimo 1.0, an open source J2EE application server developed and recently made available by the Apache Software Foundation and distributed under the Apache license, sort of acts like a young Dutch finger being pulled from a dike.
It signals that IBM's intentions and actions since acquiring Gluecode last year are not to let Geronimo wither on the vine (or worse). Indeed, for Covalent to dive deep on Geronimo shows that it's willing to big bet on the success of Geronimo. And that should mean that others will bet big on it too.
What's more, the backers of the Services Component Architecture (SCA), notably IBM and BEA, are perhaps setting their sights higher up the software infrastructure foodchain for their future profits. They may have reckoned that an open source core for their infrastructures makes the most sense, first as an alternative to their own J2EE implementations, and perhaps even eventually as a replacement. Take the service and maintenance contracts and save on the R&D.
Why not focus on SOA orchestration and process management as the new areas of differentiation, and let the core Java plumbing be best fulfilled by a strong, lightweight, large community-backed development process offering, and with the right license? Same could be asked about Apache Axis and Synapse. If the goal of Enterprise Java is to make the core middleware standardized, then make it standardized in the way the largest community wants along with the blessing of the biggest commercial player: IBM.
It now seems that the Apache community trusts IBM to do the right thing with Geronimo, and that should open the floodgates of those enterprises and ISVs that will increasingly gravitate to it. That, in turn, will allow Geronimo to mature into a more strategic role than the J2EE applications server of choice for SMBs. It could well lead to the de facto standard as core for SOA-tier runtimes: A lean, mean runtime machine with few entanglements, soon to be aided and abetted by an equally accepted ESB and Web services platform components.
All it takes is the chicken-and-egg synergy of wide adoption that begets wider community development and involvement to push projects like Geronimo into massive adoption. Once that seems like a safe bet then the best-of-breed higher-abstraction services layer and solutions sell and full support race among vendors continues. But commercial development to the Java layer does not.
This, of course, begs the question of what Sun Microsystems will do with its CDDL Java stack and Glassfish runtime. Geronimo should look as attractive to Sun shops as any other organization seeking the best runtime for the job at hand. Right?