IT Stuff

Oracle's Quest for IT Dominance

In response to my post on The BEA / Oracle Saga, blogger John Harvey (of Memphis Crime)queries:

Do you really think Oracle (Larry Ellison) wouldn't love to be the newest Microsoft?

Spot on, John. This sentiment is echoed by the 451 Group as they try putting Unbreakable Linux into perspective:

The other significant factor here is Oracle’s desire to increase its influence over its customers’ IT estates. At OpenWorld it became increasingly clear that Oracle’s strategy is to own as many customer ‘touch points” as it can. The company will tell you this is so it can provide customers with more value and increased services, but locking out the competition is a nice little side win.

As I mentioned earlier this week, that’s what Oracle VM appears to be all about. With Oracle VM and Unbreakable Linux Oracle can now claim to provide customers with everything that sits on top of its server hardware. It’s no coincidence that the company will not be certifying its database and applications on other virtualization software.

Oracle wants to own the entire IT stack, from the database drivers on the PC (just try running Microsoft ODBC drivers against an Oracle database) to the software that manages resources on the server (virtualization). The problem is that while Oracle is very, very good at writing database code, it really kinda sucks at everything else. Hence Larry's 4-year, 45-company if-you-can't-build-it-then-buy-it acquisition tear.

But Oracle's virtualization initiative is further hobbled by Oracle's outdated licensing strategy, as outlined by VMWare's Brian Byun:

For instance, a company that runs an Oracle database in a VMware virtual machine and wishes to port that single instance from one physical server to another would technically be forced to buy an additional Oracle license for each physical server used, Byun said. Or if an application in a VM is apportioned to use only 1 out of 4 CPUs in a server, the company would still have to buy a license for the entire server.

In my experience, Oracle is slow to react to market pressure for little things like licensing. So no matter how good Oracle's virtualization software is (and I doubt it is anywhere near as good as they say it is), adoption is going to be limited. Which, as far as I am concerned, is a good thing.