HP continues to cling fast to its “industry standard” message. “You have this huge, huge installed base and continued growth, with our Proliant line, the world’s largest industry standard line,” Cox said, “If you also look at the new industry standard, which was just released in the last few years ר Itanium — HP leads in that space as well.”
Customer enthusiasm for the Proliant line is unquestionable, though some analysts find contention with touting Itanium-2 as an industry standard.
“Frankly, the market acceptance of Itanium has not been particularly robust,” said Charles King, senior industry analyst for the Sageza Group. “The promotion of Itanium as an industry standard by Intel and HP has been a left-handed attempt to capitalize on the market share that IA-32 has.”
Slow adoption raises HP’s Itanium stakes. “They’re not the only vendor out there creating Itanium-based servers, but they are the only major U.S. vendor that’s really bet the farm on them,” King said.
While Opteron and Nocona Xeon (Intel’s extendable 32-bit chip and answer to Opteron) cloud the 64-bit picture, Opteron supports only an 8-way server, and Nocona is being marketed as a logical extension of Xeon. In effect, Itanium stutter-steps could extend PA-RISC’s working retirement: “If the demand’s not there, HP really won’t have an option other than extending the PA-RISC support or development for another generation or two,” King said.
King also notes that Intel’s marketing has not yet brought Opteron and Nocona head to head. “They’re really positioning [Nocona] as a low-end migration platform for small companies, or as an experimental 64-bit environment, but there’s no mention of the HPC cluster applications where Opteron’s been shining,” King said, leading him to speculate that high performance computing is either being kept reserved for Itanium-2 or Nocona does not stack up well in performance compared to Opteron.
To that end, HP’s addition of Opteron servers makes strategic sense.
So let’s get to the question of the hour: Should HP break up? King’s Sageza Group has seen clues that the vendor may well have been considering it for the past six months. HP’s R&D has recently been skewed toward printing, imaging, and consumer products, he says. And, “a few months ago they created a shadow listing on Nasdaq. [HP] said at the time that they were simply make it easier for customers to buy stock,” King said, but “that also gives them an umbrella [for an offshoot].”
King sees the appointment of the “very bright” and “very, very able” Ann Livermore (who was also a likely candidate for CEO in the pre-Carly-Fiorina days) as executive vice president of HP’s technology solutions group as significant and paving the way for change, should a new CEO be needed. If HP were to split, he postulates, “some trusted partners” might be interested in investing — perhaps Intel or Microsoft?
Place Your Bets
Going forward, Cox sees a “two-horse race” with IBM on the Unix side. Sun, in his book, has fallen by the side of the track. “We used to run into Sun all the time. [Now] We actually don’t see them that often in deals,” he said. He claims Sun still hasn’t regained a footing with its x86 offerings, “Now you see Sun offering Linux, and even, say nominally, that they would accept customers putting Windows onto their x86 platforms.”
As for the main event, HP sees its Itanium gamble as more of a sure bet against IBM’s POWER architecture. “IBM is continuing to go forward with a proprietary approach. That doesn’t give them much flexibility,” Cox said.
Whether Itanium-2 will catch on in earnest or HP jockeys its veteran PA-RISC horse back to the competition remains to be seen. As long as it continues to adapt, however, HP will be in, or at least near, pole position.