Intel's Itanium -- time to reconsider? 12 October, 2006 By Charles King

PromoPipeline Exclusive Channel Promotions Find Out How You Can Make Money Today! ENROLL FREE! >> Underpromising William Vanderbilt - Innovative Learning Channels Services: I Can See Clearly Now Beth Vanni - Amazon Consulting Cutting the margin so many times William Vanderbilt - Innovative Learning Channels The Sign of Success - High Maintenance William Vanderbilt - Innovative Learning Channels Question for the Times: Build or Resell? Gary Bixler - AMD

The Intel demo was accomplished with the help of Transitive, a virtualization vendor whose Quick- Transit technology allows applications that have been compiled for one operating system and processor to run on servers that use a different processor and operating system.

Vendors including Intel, Apple, SGI, and IBM work with Transitive, and at IDF the company announced plans to join the Itanium Solutions Alliance, a promotional group that aims to speed the IA-64 platform's position in the marketplace.

Gelsinger's dig at Sun was nothing new. In fact, considering the history of enmity between the two companies, it would have been a shock if IDF had come off without some similar stunt. Such high jinks also reflect Intel's original lofty plans for Itanium to become an industry-standard platform for all enterprise operating environments, but they also call to mind a simple question: Are they appropriate given Itanium's current state and potential?

We believe not.

Calling Itanium's prospects into question now may seem odd, at best. After all, the platform's recent dual-core upgrade (AKA Montecito) enabled Intel to boast its best Itanium performance metrics to date and also put the company and its IA-64 partners on a more even footing against other vendors' well-established multi-core solutions. Additionally, while total Itanium sales are still anemic when compared against IBM POWER and Sun UltraSPARC-based servers, they continue to gain incremental ground quarter-to-quarter.

In other words, the horizon never looked clearer for Itanium, and by extension for its parent, partners, and customers, so why the gloomy forecast? First of all, despite positive news and good vibes, Montecito's arrival obscured a singular issue: Itanium's failure to deliver on promises originally proposed by Intel and its primary IA-64 partner HP. Secondly, while Intel continues to press Itanium's potential as a replacement/migration solution for competing UNIX platforms, HP is the only vendor that is likely to gain significantly from that position.

A bit of ancient history might help shed light on Itanium's future. When Intel and HP began their development efforts in 1989, the companies envisioned a Golden Era when Itanium would stand alone as a 64-bit platform that would supplant, not just Intel's own industry-standard x/86 architecture, but other workstation and server architectures, as well. This was not pie-in-the-sky daydreaming so much as good economic sense. Since it costs billions to develop and maintain enterprise- class processor platforms, server vendors could save a pot of money and boost earnings by adopting an "industry standard" processor architecture.

Initially, both Sun and IBM seemed interested in the Itanium parade. Sun worked on a Solaris port to the new processor until abandoning the effort in mid-2000 (amidst much finger-wagging between the company and Intel). Meanwhile IBM collaborated with SCO and Sequent on "Project Monterey" to port AIX to Itanium until canceling the project in 2001 due to IA-64 production delays. Despite Intel's continuing protestations to the contrary, the Sun/IBM one-two punch delivered over half a decade ago effectively knocked out Itanium's shot at ever becoming an industry-standard platform.

Which brings us to the second point. Since its marketplace arrival, Itanium has been largely Hewlett- Packard's baby, with the company continually driving and reaping the vast majority of IA-64 sales. This is no surprise given that HP moved all its myriad 64-bit enterprise platforms and operating environments to Itanium, including HP-UX, Microsoft Windows, Linux, OpenVMS, and the Non-Stop OS (gained in the Compaq acquisition). However, appearances suggest that the lion's share of the company's IA-64 sales comes from dedicated HP customers that buy into the company's single-platform strategy. Overall, the success of IA-64 means materially more to HP than to the other vendor members of the Itanium Solutions Alliance (Bull, Fujitsu, Fujitsu-Siemens, Hitachi, NEC, SGI, and Unisys), who reserve the platform for high-end Linux and Microsoft Windows solutions.

While Transitive's Solaris-to-Itanium migration demo at IDF was technically interesting, we do not believe the company's solutions will drive significant revenues toward Itanium vendors or customers away from Sun and IBM's UNIX solutions.

The reason for this has as much to do with customer entropy as technical performance. When a business invests significant resources in a specific server solution, it typically takes a game-changing boost in performance, an act of God, or gross vendor stupidity to drive them to another platform.

Itanium has never delivered on the first point, and with IBM's POWER strategy continuing to fire on all cylinders and Sun's troubles receding into the past, the notion of Itanium as the UNIX platform of choice is best left to the history books. While we expect HP to continue enjoying the benefits of migration sales among its PA-RISC, Alpha, Non-Stop customers, the Itanium/UNIX bucks, for the most part, stop there.

So where does that leave Intel and IA-64? A few suggestions:

* Leave Itanium/UNIX arguments to HP -- It's the company's case to make or break. Plus, it gives Intel an excuse to halt the IDF sideshow antics. Then again, maybe company executives enjoy a bit of slapstick.

* Pay more attention to other partners -- Sure HP is currently driving the Itanium bus, but if the platform has a wider future it will be due in large part to the efforts of other IA-64 vendors. Watch what they are doing and help them in any and every possible way.

* Focus on the future -- Pitch Itanium as the premier platform for 64-bit Microsoft Windows and Linux solutions. Given the ongoing rise of 64-bit x86 solutions and serious competition from AMD, the IA-64 value proposition needs serious work and clarification.

In essence, the arrival of Montecito and Itanium's growing market stability offers Intel and its partners a notable opportunity to realistically consider or reconsider their prospects and direction. With such an effort, they could reinvigorate and prepare the IA-64 platform for opportunities ahead. Without it, Itanium will likely remain what it is today: a platform of limited interest outside a small circle of dedicated Intel allies and customers.

Charles King is Principal Analyst, Pund-IT Research, which focuses on understanding product and technology evolution and interpreting effects these changes will have on business customers and the greater IT marketplace. The company also offers go-to-market products and consulting engagements, as well as vendor-focused communications and editorial services.