Feeds

Intel nuances Itanium; Microsoft ignores it

Some things never sell

  • alert
  • submit to reddit

Security for virtualized datacentres

It's no longer Sun Microsystems and IBM working to undermine the Itanium processor. Both the chip's maker - Intel - and close software partner Microsoft are doing the dirty work now.

As Intel this week eased a fatter third-generation Itanium 2 into the market, it also pushed back the generally accepted delivery date for the fourth generation "Montecito" processor. Meanwhile, Microsoft, one of Itanium's biggest backers, snubbed the chip, saying it won't support the chip with its upcoming version of Windows aimed at high performance compute clusters. The cluster segment of the server market happens to be one of the few places where Itanium has enjoyed success, but Microsoft hinted that lack of demand was behind its lack of support. Go figure.

There's a bit of a hair-splitting exercise with Montecito. Intel had once planned to bring out a single core fourth generation Itanium in 2004, but then saw rivals hyping dual core products and decided that was the best direction to take. So, Intel killed the 2004 chip known internally as Shavano and declared it would bring out Montecito with two processor cores in 2005.

Although the processor will arrive in 2005, it won't enjoy a "volume ramp" - by Itanic standards - until 2006. Intel insists that it never expected things to play out any differently, but close industry watchers know that isn't the case. Intel has had more than ample opportunity to fess up to the 2006 "ramp" over the past two years, as reporters grilled the company on Itanium's dismal prospects. Funny enough, however, 2006 never came up; it was 2005 and loads of servers all the way.

In the big scheme of things, a three or four month slip for a chip is hardly newsworthy. Some of Intel's competitors - Sun, we're looking at you - miss delivery dates by years not months. But for Intel, this situation is uniquely desperate.

IBM delivered a dual core processor - Power4 - way back in 2001. Sun and HP doubled up their RISC chips this year. So the world's greatest chipmaker - and more importantly its OEMs - will be at least two, solid years behind rivals with a dual core part. Lucky for Intel, HP accounts for almost the entire Itanium ecosystem and already has a clunky module for placing two Itanium 2s where there should only be one.

Overall, the last thing Intel needs is any kind of delay getting new, exciting Itaniums to market. Customers aren't buying what Intel is currently rolling. A dual core boost sooner rather than later would have helped.

And then there's Microsoft.

Redmond will overlook Itanium when Windows Server 2003 Compute Cluster Edition hits the streets in beta next year. The x86-64-bit chips from AMD and Intel will be supported, but not IA-64.

This move makes sense for a couple of reasons. For one, Opteron and Xeon servers provide the obvious volume market where Microsoft plays best. In addition, the operating system most commonly found on Itanium clusters is Linux, and that's not likely to change anytime soon. Microsoft has been giving Itanium a shot in conjunction with HP on high-end Superdome systems but with little return on its investment. ®

Related stories

Intel relaunches Itanic
Major server vendors in giant, supercomputing cluster cluck
IBM's Power5 beast takes on Sun, HP - and IBM

New hybrid storage solutions

Whitepapers

Providing a secure and efficient Helpdesk
A single remote control platform for user support is be key to providing an efficient helpdesk. Retain full control over the way in which screen and keystroke data is transmitted.
Top 5 reasons to deploy VMware with Tegile
Data demand and the rise of virtualization is challenging IT teams to deliver storage performance, scalability and capacity that can keep up, while maximizing efficiency.
Reg Reader Research: SaaS based Email and Office Productivity Tools
Read this Reg reader report which provides advice and guidance for SMBs towards the use of SaaS based email and Office productivity tools.
Security for virtualized datacentres
Legacy security solutions are inefficient due to the architectural differences between physical and virtual environments.
Secure remote control for conventional and virtual desktops
Balancing user privacy and privileged access, in accordance with compliance frameworks and legislation. Evaluating any potential remote control choice.