Monday, April 20, 2009

Sun and Oracle: What End To Vertical Integration?

There will be many pundits discussing this issue over the next few days, so I felt motivated to get something down in print before I read it from someone else!

I was surprised to wake up this morning to hear that an old employer, Sun Microsystems, would be purchased by Oracle. After licking my wounds from dumping my options last year, I tried to imagine what a Sun/Oracle business model would look like. The question that came to mind was "what does massive vertical integration in the computing systems market look like in 2015?"

There are two options. First, an IBM approach where you pick a company of trusted advisers, architects, integrators, service professionals, etc. to take all of the pieces a company has to offer and build out whatever business solution you need. The second, that you build the extreme corporate IT cloud (public/shared or private) and operate it all as service(s).

The first model is well understood. years from now, we may look back at the whole piece part interchangeable OS (MS/Unix/Linux)/Processor(Intel)/application stack(Java/.NET) Server business (specialists) as a blip in history where the pace of innovation outstripped the utility and efficiency of vertical mainframe models, but eventually maybe it was destined to all come from the same company. For as many advancements are made from standards and interoperability, one could argue progress is also slowed by the endless infighting over those standards and the added software, management, integration headaches, and low service levels that come with this approach. Now Oracle can have the ability to deliver a complete optimized solution for many business applications. They can concentrate resources on optimizing around a common stack (down to the kernel), keep a little internal competition around (mySQL) to keep their engineers honest, and invest in service levels, functionality, transaction speed, etc. etc. all of the critical to customer values that they ultimately get paid for. On the surface this sounds like a good logical pendulum swing for the industry.

The second option is far more interesting. If you subscribe to cloud computing as a business model (as opposed to almost any kind of network based computing architecture that also fits under the umbrella of cloud) does this massive reverticalization of the computing industry really just a way to control the costs of the internal bill of material for building a cloud offering? Google has already taken that leap. They knew they couldn't afford to pay hundreds of vendors 50% margins on layers of systems and software just to offer their services. As computing power itelf continues to concentrate into larger more efficient data centers and the process for building and operating these services becomes more "productized" and less "custom" there will be more and more margin pressure on business models across the value chain. In this case, Oracle's move to control systems assets (and an OS) could be seen as a deft defensive move.

In either case early reports seem to indicate that Oracle made a sound financial move regardless of strategic motives. Putting together an accretive deal that better positions you against your strongest competitor is a no brainer. One could argue that Oracle is well positioned to play out either strategy. The only caution is that if you put too much energy into being a products company, can you really make the shift to services? Both Sun and Oracle had/are making attempts, but the acquisition, size, revenue targets and shareholder expectations could make a shift to cloud services that much harder.