When a company the size of Oracle buys a company the size of BEA, it's natural to wonder what will happen to the product line of the company being acquired. In this case, however, there's less reason for speculation, because Oracle's recent history tells us everything an IT manager needs to know about how — assuming its US$6.7 billion bid goes through — this will all playout.
It would be suicide, if not utter stupidity, for Oracle to discontinue BEA's middleware and application server product lines. Even if this is just the cynical play for a good customer list that some observers think it is, Oracle demonstrated through the PeopleSoft and J.D. Edwards acquisitions that it isn't about to pull the rug out from anyone. A few years ago it pledged lifetime support for those products, and I suspect it would to the same thing for BEA.
Given that many, if not most BEA customers run WebLogic on Oracle databases, why on Earth wouldn't it? Of course over time it might push BEA users towards its own Fusion middleware. But it could be that some potential Oracle customers weren't prepared to adopt Fusion until it was a bit more mature. BEA might have represented a safer bet for production environments, and if Oracle buys BEA those customers might be more serious about testing Fusion than they were before.
The main difference between this Oracle acquisition and the PeopleSoft acquisition is how inevitable analysts and other experts seem to think it is. Despite the machinations of activist investor Carl Icahn, no one is expecting BEA to successfully resist Oracle's siren call for long. It could be that unlike PeopleSoft, which had a reasonably broad line of applications, BEA is too narrowly focused on middleware to remain independent. It may make more sense that middleware be part of a comprehensive software stack, and the list of possible takers with holes in their existing stacks is pretty slim. IBM has Websphere, Microsoft has Windows Server. It feels like there's no one left. But of course there is.
HP is the obvious candidate. It has the operating system (HP-UX or Linux), the management software (OpenView/Mercury) and has even moved into the applications space with NeoView and other business intelligence add-ons. It's a long-standing partner of Oracle's, so the database portion is already there. There would be no overlap. It would mean closer competition with SAP, perhaps, but given its forays into business applications that may be happening more anyway. With the right services, it would allow HP to go head-to-head against IBM with much greater confidence, providing it kept the right BEA executives around, particularly with enterprise customers struggling or beginning their first service-oriented architecture project.
Perhaps it will never happen, but before everyone starts accepting Oracle's acquisition strategies as if they were their own, we should at least consider what might have been. BEA may have been destined for a buyout, but not necessarily this one.
Shane Schick is the editor of Computerworld Canada.