Speculation on the splintering of Sun's hardware business out of Oracle

What did Oracle plan to do with Sun’s hardware? There has been some conversation about this here and there already, with some people arguing the purchase was brilliant and would let Oracle control the entire user experience, from hardware to application. These people think (or thought) Oracle would keep the hardware alive.

Then there are those arguing that the HP’s involvement in the start of the Sun purchase as the entity that would buy the hardware (leaving Oracle with the software) is a pretty striking testament that Oracle never really wanted the hardware and was only making nice in public to ensure the sale when through at all and it got what it wanted. These people are working on the observation that, hey, Oracle isn’t a hardware company.

I suppose we won’t know for sure until two weeks after it happens, but the pitch of the speculation is heating up. Termination of Sun’s delayed but much-ballyhooed next generation Rock chip could be seen as cleaning up the hardware house before the sale. The Register has been stirring the pot from the beginning, and published a few new tidbits late last week

Sun logoA source close to Oracle has told The Reg that Oracle has continued to shop Sun’s hardware business around to potential buyers after the official announcement of its intention to buy the whole of Sun – and after it moved to re-assure Sun employees of its love for their hardware.

The Register’s source qualified the price Oracle was asking for Sun’s hardware business as “unrealistic.” Oracle declined to comment for this article.

…The disclosure of Oracle’s attempt to sell Sun’s hardware business would seem entirely logical given Oracle was originally only interested in Sun’s Java, Solaris, and MySQL assets.

It also makes sense, as Joe Landman and The Register article itself point out, given that Oracle’s hardware partners stand to lose a lot from an Oracle making its own hardware, and would likely not be at all happy about such a move.

My comment in this post to the effect that HPC wasn’t likely to survive in any case at Oracle drew some pointed emails my way. The counter argument to mine was that HPC is part of the new world order and it makes perfect sense for every IT company to want in. That argument still strikes me as incorrect — yes, computation is increasingly important, but desktops are increasingly powerful and while there is an unquestioned need for parallel software, I don’t believe that adoption of 1-2 socket servers with 8-16 cores qualifies as HPC of the kind that would support a separate business unit. Those are (or will be) just workstations from a hardware perspective. And that market is well served by existing companies today, none of which are Sun. The larger systems are just too big for Oracle customers to need, and the middle is so crowded it would seem to make little business sense for Oracle to risk trying to shoulder its way in over top of its long-standing hardware partners…after all, it didn’t work that well for Sun in the first place.