Contact: John Abbott
With Oracle likely just two months or so away from closing its $7.4bn acquisition of Sun Microsystems, speculation is now picking up about what parts of Sun’s technology portfolio will be dropped. (And make no mistake, cost-cutting is a major driver of this deal. Oracle has pledged to wring at least $1.5bn of operating profit from Sun in the first year that it owns the company.) But Oracle is currently working hard to counter suggestions that it won’t take on Sun’s core hardware business, and in particular, that it will give up on Sparc processor development. That’s not the case, CEO Larry Ellison insists. In fact, Oracle will increase investment in Sparc, Ellison says.
His argument is that, by designing hardware and software together to work as a system, it’s possible to avoid the low-margin trap of the commodity server business. Sparc is a key part of that, says Ellison. He adds that, as IBM has found, some system features are best done in silicon. That said, Oracle doesn’t plan to work on a Sparc-Solaris version of its Exadata database machine. Instead, it will keep the arrangements it has with Hewlett-Packard in place over its current systems activities for the Exadata database machine, which Ellison claims has been the most successful product introduction in Oracle’s 30-year history.
However, it’s still hard to believe that Oracle will make a long-term commitment to the continuing development of a proprietary RISC chip architecture. IBM’s Power and Intel’s Itanium are now the only other significant architectures: Power has been bolstered by some lucrative and high-volume gaming console contracts, while Itanium sales, driven almost exclusively by HP, have done little more than replace shipments of older HP architectures (such as Alpha, PA-RISC and NonStop) without any significant market growth. So how does Ellison see his way out of this? He plans to work in partnership with Fujitsu to add features to Sparc aimed at improving Oracle’s database performance. But reading between the lines, it’s possible that this could lead to handing over most or all of the ongoing development work for Sparc chips to Fujitsu. Provided, of course, the Japanese tech giant wants to take that on.