It's tempting to say that Larry Ellison has finally got his boot on BEA's neck, but in fact he might have BEA's boot caught in his throat. This one is going to be much harder for Oracle to swallow than PeopleSoft and Siebel were, not to mention the countless smaller technology vendors that Oracle has acquired in recent years. Although some of BEA's technology complements Oracle's portfolio (as these prior acquisitions mostly did), a whole lot of it -- from the Java stack and enterprise service bus to portal and collaboration software -- competes directly with Oracle offerings.
Certainly BEA has technology that Oracle wants (see Eric Knorr's take, "BEA's sweet portfolio"). Most recently, BEA has even brought terrific innovation to the J2EE server, launching a "virtual" edition of WebLogic that runs in a VMware virtual machine without a resource-hogging operating system (see Andrew Binstock's review). Nevertheless, the monumental headaches that Oracle product strategists will face in reconciling BEA's offerings with Oracle Fusion Middleware supports the view that the acquisition is aimed mainly at bolstering Oracle's middleware market position, and not about nabbing important technology.
Oracle has been pushing to become a heavyweight in middleware, where the movement toward services-based integration and process orchestration opens up new opportunities. Acquiring BEA cinches that. According to Gartner's figures, the worldwide market for "portals, process, and middleware" software stood at nearly $12 billion in 2006, and IBM had the lion's share, with almost 32 percent of that revenue. Adding BEA's 10.5 percent to Oracle's 8.6 percent makes the combo a healthy #2, and way ahead of #3 Microsoft, which Gartner attributes a 4.2 percent market share.
Oracle had already been making steady gains in middleware at BEA's and IBM's expense. If Oracle succeeds in smoothly integrating BEA (a formidable task and wild assumption, but Oracle has proven to be good at this), IBM's lead will not be safe. But Oracle has more than Big Blue to worry about here, and pulling in BEA may also help the company keep the threat of open source alternatives at bay. Oracle has always had an ambiguous relationship to open source. It was quick to adopt Linux as a preferred platform, a good hedge against both Microsoft and Sun. But it has otherwise strived to keep the open source middleware and database genies in the bottle, going to a lot of trouble to run interference against up-and-coming rivals Red Hat and MySQL.
We all remember the rumors that Oracle would acquire JBoss (now part of Red Hat), then MySQL, and finally Red Hat itself, and in the case of MySQL the rumors turned out to be true. After that offer was spurned, Oracle dealt blows to MySQL by acquiring Sleepycat Software and then Innobase, both sources of transactional storage engines at the heart of MySQL's goal to deliver a truly enterprise-class database. Now that MySQL will be a part of Sun Microsystems, it will be a bigger threat than ever.
On the middleware front, the battle is not only with commercial solutions from IBM and Sun and Tibco, but with JBoss and open source brethren including MuleSource, WSO2, and Iona, which are casting a shadow over future sales prospects. Both IBM and Sun have even embraced open source middleware, most notably Sun with its Sun Open ESB.
Oracle is clearly not content to beat open source rivals by joining them. Caught between strong commercial rivals on one side and open source challengers on the other, Oracle had two options: acquire open source challengers to get them out of the way, or shore up its market position. By buying BEA, Oracle eliminates a strong commercial rival and expands its reach into enterprise customers for a new generation of middleware. Was buying a top commercial rival instead of the leading open source challenger the right strategy? Only time will tell.
Meanwhile, we need a fresh set of rumors around Red Hat. When will Sun make the move?