IBM wants to finish JBoss

After losing its founder and leader to music and other personal interests shortly after being acquired by Red Hat, JBoss has a new set of issues to deal with. Marc Fleury left after sharing with his colleagues how he felt about working with Red Hat: "I am increasingly experiencing diminishing returns on my emotional and professional investments at Red Hat."

The new danger comes from outside. IBM and Covalent announced today that they had each contributed a significant amount of IP to Apache's Geronimo to help users migrate from JBoss to Geronimo. Although I personally consider this a serious attack on JBoss, Shaun Connolly (VP of product management in the JBoss division of Red Hat) begs to differ calling the IBM-Covalent initiative "uninteresting". I still think that JBoss today is far more superior than Geronimo if they play a feature war. But we all know the better products don't always win. I also believe that existing JBoss customers may not want to switch to a less performant application server if they are already in production. If it works why fix it? But I still think that when giants like IBM or Microsoft go after a much smaller company, we know how that movie ends.

Speaking of movies let's rewind the JBoss/IBM movie. Let me refresh your memory on a few events that put together are confusing to say the least:

- In September of 2003, JBoss and IBM team up to cool off the growing popularity of Microsoft's C#
- Not too long after this Marc Fleury started bashing BEA and IBM on his blog
- In May of 2005 IBM purchases Gluecode a company that employed most Apache Geronimo contributors and positions this acquisition as their entry level, lightweight application server. They later called the Geronimo-based product WAS Community Edition.
- Exactly two years after deciding IBM was nice and C#/Microsoft evil, Marc Fleury partners with Microsoft (Sept 2005). Their partnership shocked me (and I was not the only one) but I thought it was pretty clever after all. Marc described that day as his best day ever.
The idea there was that half JBoss servers were running on Windows so let's work together on making JBoss work even better on Windows and SQL Server, Active Directory and single sign-on, etc.
- Then Red Hat buys JBoss, Microsoft becomes great friends with Novell and Fleury doesn't like working for Red Hat, fakes a paternity leave and never comes back to work. [Sorry I had to compress the story]
- Next IBM feels JBoss is kind of vulnerable and decides to partner with Covalent to hurt them even more, hence the announcement.

Concretely IBM (which roughly employs half of Geronimo's committers) and Covalent (which already provides support for Apache's Tomcat, HTTP Server and Axis) are getting together to provide quality support for Geronimo and lure people away from JBoss. Paul Buck, director of IBM WebSphere open source said that they were going to provide a migration tool that would go through the J2EE application itself and look for any required changes at the source that we know are different between JBoss and Geronimo.

I am interested in your thoughts, do you think IBM with this move is going hurt JBoss' business in a significant way? Can somebody tell me why we never see Glassfish in these battles? It's also a Java EE open source application server but no one seems to take it seriously. I'd be interested to hear from anybody who reads this blog who uses Glassfish in the enterprise.