European Union Approves Oracle Gobbling Up Sun Microsystems
January 25, 2010 Timothy Prickett Morgan
The many-year drama and 25-year history of Sun Microsystems is nearly over, and a new page in the history of software giant Oracle is about to be turned as Oracle buys a hardware company that controls the Java programming language and attempts to transform itself into a provider of integrated IT stacks.
While Sun’s own shareholders (including a number of venture capitalists who had kicked in funds to the struggling server maker over the years) and the U.S. Department of Justice’s Antitrust Division approved the $7.4 billion acquisition of Sun last summer, the European Union’s antitrust regulators have been stalling the deal over concerns about Java and the open source MySQL database, which has its origins in Sweden and represents one of the clear winners in the IT racket that any European company can boast of. (Sun bought MySQL for $1 billion in January 2008, radically overpaying for the company because of the popularity of its eponymous database. It was one of many moves that Sun has made since the dot-com bust that made sense, but not at the numbers Sun had to pay.) As 2009 was winding down, Oracle made a commitment to continue to support MySQL Enterprise Edition for at least the next five years and to continue to develop the product despite the fact that Oracle no doubt wishes MySQL would be erased from the fabric of the universe so it could make gillions more in sales.
On Thursday morning in Brussels, Belgium, the heart of the European Beast (well, some organ, anyway), Neelie Kroes, the European commissioner for competition, released a terse statement: “I am now satisfied that competition and innovation will be preserved on all the markets concerned. Oracle’s acquisition of Sun has the potential to revitalize important assets and create new and innovative products.”
We’ll see. That certainly is the plan according to Larry Ellison, Oracle’s co-founder and chief executive officer, who has talked up the future integrated stack of servers, storage, networking, systems software, and application software that will be enabled once Sun is part of Oracle and Oracle knocks its product lines into shape.
That last bit could be tough, of course. Both the Sun and Fujitsu Sparc processor and related server roadmaps are a mess. Sun killed off its “Rock” UltraSparc-RK 16-core, high-end processors, the ones that were supposed to crush the Power6+ chips last year, and its entry “Niagara” Sparc T series of multicore chips, while interesting, are not a price/performance leader and cannot scale enough to make up for the missing Rocks. Fujitsu makes decent quad-core Sparc64-VII processors, but is moving at much too slow of a pace to keep up with IBM in the Unix racket. Fujitsu is not expected to get eight-core Sparc64-VIII chips to market until 2011 or 2012; IBM will do so in a few weeks.
The easiest–and perhaps smartest–thing for Oracle to do is spin out its Sparc development operations in a partnership with Fujitsu, tap GlobalFoundries (the former AMD chip manufacturing operation) to be its foundry, and hook up with Fujitsu as a contract manufacturer for Sparc and X64 rack, tower, and blade servers. Yes, I know Sun doesn’t sell tower servers. That’s one of the things Sun should have done more than a decade ago and didn’t. Anyway, by spinning the Sparc out with a partnership with Fujitsu, Oracle can maintain its commitment to Sparcs while not fully bearing the cost of development. Oracle needs to figure out what the long-term plan is for Sparc-based machines in relation to the X64 boxes peddled by Sun, too. And finally, Oracle has to figure out a coherent storage and networking strategy as well, since Sun’s product lines in these areas have some holes.
In many ways, the fact that IBM got OS/400 and AIX running on what was essentially the same hardware platform back in 1997 made coping with its legacy platform (OS/400) a lot easier as its Unix platform was elevated and pushed very aggressively. Sun and now Oracle won’t have it as easy because the legacy platform is Solaris running on Sparc compared to Solaris, Linux, and Windows running on X64 iron. The OS/400 platform made the jump to Power iron and lots of smarts and trickery to get legacy RPG and COBOL code to run on the new machines. Sun’s Sparc/Solaris architecture was not designed with seamless and (largely) invisible application portability despite lots of serious hardware changes, which, quite frankly, is why Sun has gotten its ass kicked by IBM since the dot-com bust. And with IBM owning the QuickTransit emulation environment (the one that lets Sparc/Solaris apps run in emulation mode on X64 iron running Solaris), Oracle is going to have to be clever to get out of this bind while not wasting a lot of money and ticking off the remaining Sparc shops.
Good luck, Larry.
The new Oracle has a convoluted set of middleware, server virtualization, systems management, and application development tools it needs to sort through, too. In just about every instance, you can expect a Sun product to lose out to an Oracle one over the long haul.
Personally, I think that for political and marketing purposes, Oracle is going to have to make Sparc and X64 variants of its integrated Solaris and Linux stacks for years to come, and then let the market decide. Once the market has decided, and there seems little doubt that the market will decide on a mix of Oracle Enterprise Linux and Solaris running on X64 machinery, Sparc will go the way of all flesh. Despite however much Oracle has said it wants to invest in Sparc. But, there are always possibilities. It could turn out that Oracle really believes in homegrown hardware (like Apple clearly does for its consumer devices) and that over the long haul, it creates such a good Sparc server lineup that Fujitsu ditches its own and buys stuff from Oracle.
Stranger things have happened. Like, for instance, Oracle buying Sun in the first place.