Thursday, June 14, 2007

Itanic: not quite dead yet?

Intel announced 3 new Itanium-based chips today.

Itanium, is that thing still around? Well, according to this article, the Itanic server market is a $3.4b market. You've got to be kidding me. Hold on, it gets better. The market is up 40% year over year.

I can only imagine one thing is keeping Itanic alive, and it goes something like this:

  • Customer currently on expensive Sun or IBM POWER hardware decides to try to save money by moving to Linux or Windows on clusters of x64 boxes.
  • x64 boxes turn out to be woefully unprepared to deal with things like 1 TB databases and the build quality of the hardware is shit.
  • Customer can't deal with ditching expensive migration to Linux/Windows, so instead opt for Itanium boxes over going back to Solaris or AIX.
Does that sound feasible? Because I can't really think of any other reason someone would wake up one day and decide to buy Itanium unless his name is Andy Grove or Paul Otellini. Truth be told, people actually say some good things about the chip -- after all it was based on the PA-RISC, which I used in the mid-90s and really kicked ass. Plus, it's just not as hard as it used to be to support multiple architectures. It's relatively trivial to compile your code for Itanium, x64 and x86.

But let's remember we're talking about a chip that barely anyone has bought in the 7 years that it's been in release. Why is it gaining momentum now?

Then I read this part of the article on the Wall Street Journal: "Because Itanium sells in small volumes compared to other chips, the future of the product line is often questioned. Ms. Bryant said the company wanted to reinforce its commitment to the technology by showing that design teams are working on models for years into the future."

That seals the deal, Itanium is SCREWED! Anytime a company "reinforces its commitment", you know it's the end of the road for that product. As an Intel shareholder with too many shares, I wish they'd run away from this product as fast as possible. Focus on crushing AMD with the x64 architecture. The Woodcrest and Clovertown chips are insane and blow away the competition. Everyone's talking about them. Bet the company on that great stuff instead of screwing around with Itanic, which no one cares about anyway.

No comments: