Still Patching After All These Years
Emulator connects to terminals, POC efforts

HP's expert estimates Itanium's end-date

We return you to California's Santa Clara County Superior Court, where the future of Itanium and HP-UX is already in progress. HP and Oracle continued their battle over the future and value of Itanium yesterday, with each side trying to wring dollars out of their dispute over whether Itanium is finished at HP. The lawsuit's final phase addresses damages. Oracle hopes to prove HP's public and partner strategies cost them sales of Sun servers where Integrity had already lost the business.

StopwatchOracle's expert estimated the company lost $95 million in profits, working on the premise that HP lied about the future of its only HP-UX processor line. The Integrity servers have been a popular platform for Oracle's database. A lawsuit that wrapped up in September forced Oracle to continue its development for the server line. The database vendor wanted to stop enhancing Oracle for HP's platforms including HP-UX, all tied to the Itanium chip.

HP's expert Jonathan Orszag of the consulting firm Compass Lexecon had to counter by estimating the lifespan of HP's Itanium business. Orszag said the ending date for Itanium looked to him like 2020. HP would have surely reviewed Orszag's testimony before he offered it to judge James Kleinberg. HP's expert witness in the damages phase of the suit said he based his testimony on Itanium road maps from HP as well its chip partner, Intel.

If Orszag and Hewlett-Packard are on target, then 2020 would mark about two decades of actual service to the enterprise computing customer. That's a mark that HP's initial chip family for the 3000 didn't achieve. But the period of 1974-1989 was nothing like the 21st Century. For one thing, Intel didn't have competing versions of an enterprise business processor on sale during the '70s and '80s. That split focus for Intel showed up again last month, when the chip maker announced a couple of downgrades to Itanium's future.


Those announcements looked like signals that Intel is shifting its R&D resources away from the only processor that drives HP-UX applications. Many migrated HP 3000 sites are using HP-UX. Those who have a migration in their future, nearby or otherwise, still have Itanium on their menu of choices.

Intel said in a modest post on a webpage that it won't be re-engineering Itanium to use the same sockets as the mainstream Intel chipset Xeon. HP's partner added that it will design and manufacture the Kittson family of Itanium using the 32 nanometer architecture that's in the current Poulson family. Intel has 28-nm designs in store for Xeon. Both the smaller architecture and the socket-compatible design have been withdrawn from Intel's Itanium plans.

Kittson will be manufactured on Intel’s 32-nm process technology and will be socket compatible with the existing Intel Itanium 9300/9500 platforms, providing customers with performance improvements, investment protection, and a seamless upgrade path for existing systems. The modular development model, which converges on a common Intel Xeon/Intel Itanium socket and motherboard, will be evaluated for future implementation opportunities.

Evaluated for future implementation opportunities means, in summary, that the engineering to bring Itanium closer to an assured position at Intel is halted. The HP-UX chip is still a unique design at Intel -- which might be okay if Itanium sales were growing, or even maintaining market share. HP says that's not true. The Intel decision forces HP to gamble that the installed base of Itanium users will buy enough to maintain HP's promises on the roadmap which Orszag used. Winning new customers gets harder with every improvement to Xeon which is denied to Itanium.

Hewlett-Packard was in a similar situation, with the then-fledgling Itanium designs, near the end of its HP 3000 futures. A few years before HP's MPE/iX "opportunities" ended, the 3000 division stood at the crossroads of adopting what was called IA-64. First the division would design toward IA-64 with future 3000s. Then it said it would wait, and make do with existing PA-RISC chip family.

The outcome of the California court's damages phase will result in a material impact on the futures of the Business Critial Systems operation at HP. Just not enough. Even if HP wins its $4.2 billion demand, the money would only provide $600 million a year in lost profits for the group. It would offset the losses BCS will post as it continues its spiral. HP CEO Meg Whitman said last month that BCS "was a big and profitable business, and you see that it's declined by 24 percent year over year. The good news is that we've got the best product lineup we've had in a long time in [the Enterprise Group.]"

The bad news? HP's CEO now refers to BCS-Itanium business in the past tense.

Last year the entire Enterprise Servers, Storage and Networking unit at HP -- whose BCS-Integrity line is a fraction of the Xeon-based Linux servers, ProLiant sales, disk sales and networking products -- posted just $2.1 billion in profit. Whitman noted that HP is making investments behind the Enterprise lineup.

But those Hewlett-Packard and Intel roadmaps, plus the chip-maker's announcements, show the money is going elsewhere. Based on what Intel and that HP expert say about Itanium chip dies in the future, it looks like the final act has been scheduled at Hewlett-Packard. Is six more years enough for today's customers? Probably so -- if they don't arrive on HP-UX with 3000-like expectations for investment protection.