September 15, 2005 11:19 AM PDT

Dell shuttering Itanium server business

Dell is phasing out Intel Itanium-based servers, a move by one of the chipmaker's closest allies that reflects Dell's emphasis on low-cost servers and the processor's failure to spread into the mainstream.

A representative with Round Rock, Texas-based Dell said Thursday that an Itanium transition is under way now. Itanium servers are no longer promoted on Dell's Web site, which now directs attention solely to servers using Intel's Xeon. But Itanium-based models such as the PowerEdge 7250 can be found by searching.

Though Dell remains the largest seller of Intel processors, the decision to drop Itanium has not caused widespread panic at Intel.

"Losing Dell as an Itanium customer is unfortunate, but frankly, we see their impact as negligible," Intel representative Erica Fields said.

Hewlett-Packard is by far the biggest in the Itanium server market, but Fujitsu, NEC, Hitachi, Unisys and SGI also sell their own high-end Itanium servers.

The writing was on the wall for Dell and Itanium once Intel started aiming the processor at large multiprocessor systems instead of the broader server market. For one thing, Dell has publicly spurned massive "big iron" computers as "less and less interesting." Indeed, in 2002, the company scrapped a deal to sell a 32-processor Xeon server built by Unisys.

Microsoft moves sway Dell
More recent decisions at Microsoft also reinforce Dell's direction. A coming version of Windows Server 2003 called R2 is geared for smaller servers and won't support Itanium, and the next Windows server operating system, code-named Longhorn server, will only be geared for high-end server tasks, Microsoft said this month.

Dell's interest in Itanium has ranged all the way from lukewarm to complete indifference. In November 2001, Joe Marengi, the co-manager for Dell Americas, explained that interest among customers for Itanium servers was "effectively zero."

"The investment involved in the transition in huge," he said at the time. "I don't see the speed and benefit to what the processor brings to the equation," he added. Itanium had also been subject to several delays.

In the summer of 2002, Dell was still in wait-and-see mode. But, by November 2002, Marengi said, Dell had a change of heart and would begin to produce Itanium 2 servers. However, Dell was also phasing out a line of Itanium workstations. Intel also designed and partly manufactured the Itanium 2 servers Dell sold, which reduced the time, money and effort Dell was required to spend.

Dell isn't the only one to turn its back on Itanium. IBM this year decided against Itanium support for a high-end server chipset that as a result only works with Intel's Xeon chips. IBM has its own Power processor to promote for high-end servers.

Itanium allies have also withdrawn from other markets. In 2004, Microsoft canceled a workstation version of Windows for Itanium after HP canceled its workstation.

Itanium shipments fell short of Intel goals in 2004. The company had hoped to double 2003 shipments.

Still-active developments
Despite the problems, allies are still plugging away at Itanium. A number of companies are expected to announce an Itanium market development alliance later this month to try to help with software support and other activities.

The alliance made up of Intel, HP, Microsoft, Oracle and others is expected to sponsor porting events to help programmers bring their software to the processor and find combinations of Itanium hardware and software for various tasks.

And there are years of new Itanium chips planned from Intel, starting with the first dual-core model later this year. That chip, code-named Montecito, will more than double performance over the current "Madison" generation of Itanium processors, Intel has said.

After Montecito comes a derivative called Montvale, as well as a new four-core design called Tukwila and another four-core design called Poulson.

7 comments

Join the conversation!
Add your comment
The chip just wasn't meant to be
Can't blame Dell. Intel and HP had long delays with the chip from the start. Too much competition has kept the x86 alive and thriving.
Posted by bobby_brady (765 comments )
Reply Link Flag
Why Itainium?
Why bother with the expense and hassle of a upgrade to Itainium when the x86 hardware works just fine? Problem here folks, is that the Itainium is a new hardware platform with not much software out there. And in the time of shrinking IT budgets, poor x86 emulation performance, and high cost, the Itainium never was a hit in the market place.

AMD on the otherhand made an excellent business decision with their Opteron chip by keeping with the x86 instruction set. So, not only can the Opteron run the new 64-bit software, it is also fully backwards compatible with all previous x86 software at full speed. Something that Intel seems to have forgot.

Sorry Intel, but if I have to spend many thousands of dollars to buy all new hardware and software just to have a 64-bit platform, I'm going to AMD because their solution is more cost effective.
Posted by Maelstorm (130 comments )
Reply Link Flag
Itainium? Why bother?
Why bother with the expense and hassle of a upgrade to Itainium when the x86 hardware works just fine? Problem here folks, is that the Itainium is a new hardware platform with not much software out there. And in the time of shrinking IT budgets, poor x86 emulation performance, and high cost, the Itainium never was a hit in the market place.

AMD on the otherhand made an excellent business decision with their Opteron chip by keeping with the x86 instruction set. So, not only can the Opteron run the new 64-bit software, it is also fully backwards compatible with all previous x86 software at full speed. Something that Intel seems to have forgot.

Sorry Intel, but if I have to spend many thousands of dollars to buy all new hardware and software just to have a 64-bit platform, I'm going to AMD because their solution is more cost effective.
Posted by Maelstorm (130 comments )
Reply Link Flag
True, but...
Intels solution is a big transition with a big payoff. AMDs solution is a small transition with a medium payoff and a permanent speed penalty. (All 64 bit instructions need a prefix instruction.)

Even so, I am surprised Dell held out this long.
Posted by Andrew J Glina (1673 comments )
Link Flag
Dell is not bowing to Intel pressure...Amazing!
Itanic is in the same catagory as the specialized CISC and RISC processors and as such only addresses a limited audience. So their competitors are SUN, IBM, and these companies have embedded audiences and are still struggling. What in the heck were Intel and HP thinking? Intel was thinking it as a next generation processor to kill off the x86 line and even killed their 32/64Bit processor project "since it wouldn't be needed after the launch of Itanic...WRONG. AMD put them back on course but behind the game.
But this revelation blows me away as the norm for Dell is "what ever you say Mr. Otellini", it's as if Dell is just a subsidary of Intel.

Bottom Line: Intel when are you going to cut your losses and give it up? Take the resources your are devoting to the Itanic and place them in developing better solutions to your current bottlenecks in the x86 area.

Fred Dunn
Posted by fred dunn (793 comments )
Reply Link Flag
Intel should examine their development process
The statement that Dell's decision is "negligible" is probably going to haunt Intel for some time to come. The marketplace abandonment of the Itanium processor and Intel's apparent disregard for customer needs are indicators that something is very wrong with their product development process. The old adage "if we build it, they will come" isn't holding up anymore. Intel, and other companies who rely on past market dominance should not lose sight of "how" they develop their products.
Posted by (1 comment )
Reply Link Flag
Very Well Said...
Intel still thinks that what ever they have planned is what the consumer wants/needs. The falicy of this approach is that they do not bother to do any market analysis, rather they depend on "yes" men within.
But they have started to feel the pain as others prove their approach wrong and millions of research and development costs are wasted.

Fred Dunn
Posted by fred dunn (793 comments )
Link Flag
 

Join the conversation

Add your comment

The posting of advertisements, profanity, or personal attacks is prohibited. Click here to review our Terms of Use.

What's Hot

Discussions

Shared

RSS Feeds

Add headlines from CNET News to your homepage or feedreader.