A little 'off-topic' from BSM per se, put we think it's worth a look at what's happening as Larry Ellison and his Oracle team try to move from software giant to full system solution provider.
The recently announced 5% drop in Oracle's HW revenues (continuing a drop that began with the Sun acquisition) once again casts doubts on its future as a hardware vendor. Before acquiring Sun, Oracle claimed the hardware division would be profitable in a very short time. We're still waiting.
Oracle executives (Safra Catz and Mark Hurd) insist the drop is a result of efforts to increase the profitability and margin in hardware sales. To be honest, the margins did improve. And, Ellison continues to insist they will make no attempt to hang onto unprofitable business segments - hence, his comments about exiting the high-volume x86 server market. (Ellison's stated goal is the get operating margins to around 46% where it was before the Sun acquisition. This is up from the last reported 42 %.)
These comments are consistent with what you would expect from executives in the midst of trying to execute a turn-around in a very tough market. None of the statements are a surprise. None appear outrageous. Vendors unwilling to surrender a costly, barely profitable attachment to a hardware platform or even a software product - tend to end up at the side of the road. Turning away customers, deteriorating sales numbers are not obviously great strategies for the long term. Let's see how this might play out.
In our view, once customers begin to hold back on their purchases of hardware (as they tend to do when the platform future appears shaky); the downward trend is likely to continue and expand. At the very least, until the customer perception of the prospects for the hardware division turns positive. This holds doubly true when a company, like Oracle, is moving into a totally new market space. There's uncertainty in spades when a software company attempts to become a full service solution vendor, no matter what the quality of new executives or past performance. The problem is compounded when the acquired company seemed to be having problems prior to its purchase. No one wants to invest in dead-end hardware, which is why it is most companies avoid talking of 'end-of-life' until the replacement product is available and accepted.
In the case of Oracle/Sun such a change in perspective will take time. It's not good when customers start publicly raising questions about the future of SPARC/Solaris. The risk increases when the platform was positioned as THE critical technology for the division. Restoring confidence will take time and money.
Oracle will have to demonstrate its commitment and willingness to make the necessary investment to restore confidence in SPARC with a detailed roadmap of its future. In concert, Oracle has to execute a credible and successful campaign to increase SPARC penetration and position not simply as a place holder for Exadata-like appliances.
Yes, Exadata is x86-based running Linux. But, it is a clear winner for Oracle today, with a promising future. That makes Exadata a strategic platform. The question that follows is: can Oracle be a hardware powerhouse without SPARC? We believe they need SPARC, IF they are to deliver to Larry's vision of offering a system with a fully integrated stack. The question then becomes can Oracle afford a dual-platform strategy? And, will Oracle be able to convince customers of the validity of their commitment to a multi-platform solution as other vendors have done? Both questions are quite reasonable and must be convincingly answered by Oracle.
When customers are convinced of the commitment and the success at execution, then they will reevaluate SPARC as a solution. We expect this process will take some time and see no obvious way to short circuit or jump start the process. Oracle needs to make a major SPARC announcement this year. (Larry Ellison announced the intention to do that at Oracle Open World this fall.) It remains to be seen how its performance improvements compare to competitor products from IBM and others. Assuming ORACLE can pull all these things together, it is possible that some time in 2012, the prospects for Sun's success will begin a recognizable turn around.