Gavin King of Hibernate/JBoss fame dropped us a comment regarding the impact of IBM's WebSphere announcement on JBoss.
While Derby is an interesting product, and while I very much hope that the pure-Java database space matures soon, I don't think it's at all reasonable to believe that people will be deploying production web or enterprise applications on Derby in the near future. (This is certainly a niche for pure Java databases, but this is probably not it.) Rather, people in pursuit of a complete-opensource stack will continue to use MySQL or PostgreSQL as they do today. We have a great relationship with MySQL, and I don't frankly see Derby as any kind of a threat to their market penetration. The JBoss / MySQL stack is a many times more credible combination than WebSphere Community / Derby! Meanwhile, our customers have huge existing legacy databases and most aren't interested in moving off of Oracle or DB2 or Sybase or SQL Server or Teradata or whatever at the same time they migrate away from WebSphere or WebLogic. I mean, why should they? In most cases, switching databases is going to be a much, much more expensive proposition than switching appservers.So, in conclusion, I don't see there being any real synergy in having a database bundled with the application server. Frankly, if I were IBM, I would not be trumpeting my story around data without having some (any) kind of credible story for /persistence/ first. Hehe, I should say that the proposed stacks are:
WebSphere Community / EJB CMP / Derby
JBoss AS / Hibernate / MySQL
Gavin makes a good point. I don't think that Derby is really all that relevant. The issue that remains important is that IBM is capable of creating a good stack, and with it's sales force could definitely sway customers to purchase the whole package from them. The best product doesn't always win-if it did we'd all be using Macs ;>
BTW-Thanks to Gavin for shutting me up.
Posted by Dave Rosenberg on October 28, 2005 12:13 PM
