Thursday, April 23, 2009

Oracle Buys Sun. A Natural Progression, or Unnatural Mistake?

Can the Software Giant make sense of Hardware?

Java and Solaris are the prize.

The announcement that Oracle will be taking over Sun Microsystems has generated a huge amount of reaction in the blogs and within the IT industry generally. There have been questions about what Oracle's strategy is, what the future will be for Solaris on Sparc, where the free MySQL database lives etc. etc. There have also been been some questions in some minds about Larry Ellison's sanity. It has certainly been a bold move. Some claim to have seen it coming - I certainly did not.

My own interpretation is that Oracle are being opportunist. There was an ailing company - Sun Microsystems - whose heydays in the .COM boom were long gone. They had a huge commitment to R&D without much to show for it. They have some distinctive products (the Sparc chips), some OEM market (Storage from Hitachi), and some very interesting free, or nearly free software and commitment to the Open Source world (Solaris 10, Star Office which forks development in Open Office, and MySQL). They also own the JAVA stack. Maybe they are worth a few billion, even if they don't currently make a profit.

The Profit Motive

And that is the key point. If Oracle is about anything, it is about a business that exploits their assets to make a profit. I suspect that there was no "grand strategy". Charles Wang, the former head of Computer Associates, once said that at the level he worked, people "make it up as we go along". Oracle is driven by a Profit Motive. That, and a hatred of Microsoft. Add to that the fact that IBM and Cisco are circling around Oracle's historical profit levels, and the deal makes sense.

More than just a database company

Oracle has been more than a database product company for many years. They started with the database, but over the last years have positioned the company as an Application Platform. Oracle Financials was one attempt. Then add PeopleSoft, and myriads of other acquisitions. So they have diversified away from the database. If you look at their latest figures, you see that the Oracle database itself is less that 50% of the revenue of the company as a whole. So this is an exercise is diversification. Move up and down the software stack to ensure that you can offer everything the customer could possibly want. All at a profit.

Hardware is not Software, Larry

Oracle has tried to move into the Hardware space before. They created a product called "Raw Iron" which was an embedded hardware product for running the Oracle database. Coincidentally (maybe?) this was based on Sun hardware. There is a very interesting FAQ released by Oracle yesterday which says "Oracle's ownership of two key Sun software assets, Java and Solaris, is expected to provide our customers with significant benefit.". This suggests strongly that Oracle are still seeing Sun as a software vendor. Whilst Oracle have lots of experience in integrating companies, they have always been other software companies. Running a Hardware company is a different thing. The sales model is different, the lead times are different. And you have to ship physical equipment all round the world. It will not be an easy integration.

Predictions

Everyone else is making predictions. Usually, these are based on what the author would do. However, these are my predictions on what Oracle themselves will do. Whether they are accurate predictions, I will leave history to determine. Whether they are good business, that will be the realm of Economics.
ProductPrediction
Java I doubt if Oracle want to upset the Java community. In fact, I suspect that Java will become more open. Oracle's view will be - why do the work ourselves when there are so many willing volunteers to do it for us? Oracle want Java so that they can ensure that all their applications have a good strong application server stack. But watch out for "Oracle Extensions" to the main product.
Solaris Oracle claims that it can now optimize the Oracle database for some of the unique high-end features of Solaris. It has always had this option, but was afraid of "lock-in" to another vendor's product. I predict that there will be some new features of Solaris that Oracle can exploit. But there won't be much. They don't want to alienate the Linux users.
Sparc Chips This is the nub of the question. Oracle have said that they will grow the business. Oracle salesmen may clinch deals by selling integrated hardware alongside the application. They will be able to point to Oracle-specific APIs in Solaris to show performance gains. However, if Fujitsu decide to come calling, I would not put it past them to hand over Sparc development, and OEM the solutions.
Sun Storage If Oracle can sell storage at a marginal profit, they will do so. Particularly if it means software license sales.
Star Office Maybe this is a product that can be stacked against Microsoft. But you have to sell a lot of Star Office licenses to equate to a single Oracle DBMS license. Is it worth it? I suspect that, as with Sun, Star Office will be a sideshow.
MySQL Who cares? The fact is that MySQL generates relatively little profit. So R&D will be cut back. The product will still be there, but will rely on the OpenSource community to develop it. Oracle knows that MySQL is not much of a threat. It will be allowed to stand, or fall, on it's own.
It will be worth watching this one...