Arik Hesseldahl

Recent Posts by Arik Hesseldahl

Oracle Drops New Documents in Itanium Trial, and They’re Juicy

A new trove of previously redacted emails and other documents submitted as evidence in the Oracle-Hewlett-Packard Itanium trial fill in a lot of the blanks on the state of play among HP, Oracle and Intel before the lawsuit over the Itanium chip began last summer.

The documents were released as part of a new offensive by Oracle to ratchet up the pressure on HP and take its case to the marketplace — that HP and Intel had already planned to bring an end to the Itanium chip’s life, and that HP lied to its customers about the chip’s long-term future.

HP has long denied Oracle’s contention, and has tried to portray this as Oracle’s failure to live up to its end of a contract.

In an open letter to affected customers, Oracle said it was releasing the documents in order to allow customers to “make your own decision” on the matter:

Dear Customer:

A little over a year ago we announced that we would no longer be developing new versions of Oracle’s database and other products on the Intel Itanium platform due to our strongly held belief of Itanium’s imminent end of life. We ensured our Itanium customers would have an easy transition to the platform of their choice by committing to 10 years of support for existing Oracle software running on Itanium.

Hewlett Packard strongly disagreed with our characterization of Itanium’s future and launched an immediate campaign designed, in their words, to foment “customer outrage.”

At this time, there are many documents that have been disclosed through litigation that describe the true state of Itanium in Hewlett-Packard’s own words. Rather than us interpreting this situation for you, we thought we would give you access to the public HP documents so you can make your own decision regarding your investments in Itanium technology.

After reading these documents we are confident that you will agree with our decision, taken with the best interest of our joint customers in mind.

Sincerely,
Jeb Dasteel, Senior Vice President and Chief Customer Officer
Oracle

Bloomberg Businessweek got its hands on some of these trial exhibits, but in its posts this morning — and presumably in the tech section of this week’s magazine — stuck to a fairly limited set of highlights. And in fact there’s not so much that’s surprising, if you read my story on the unredacted version of Oracle’s cross-complaint from January.

Basically, the new documents add more color and a lot more tension to the state of the HP-Intel relationship over the production of the Itanium chip. It also lends a lot of weight to Oracle’s narrative heading into that trial: That HP relied heavily on profits earned from multiyear sales and support contracts with customers who bought its Integrity servers that run Intel’s exotic and expensive Itanium chip. In support of that, it paid Intel nearly a half-billion dollars to keep the chip alive, despite the fact that, outside of HP, there was no other single vendor using Itanium chips.

In the emails, Intel, for its part, certainly looks like it wants out of the business of making the chip, but is willing to accept HP’s money to keep churning them out. Asked at one point what would happen if HP didn’t pay a certain amount to Intel, Intel would — in the words of Martin Fink, then-head of HP’s Business Critical Server business — shut down the teams producing certain chips that were in the process of being designed, and slap “high fives all around.”

On its face, there’s nothing inherently wrong with HP paying Intel to keep making a particular chip. HP had customers willing to buy these servers, and it made money supporting them, so paying Intel to keep them coming — remember, HP for all intents and purposes, is the only vendor buying this chip — was more or less a cost of doing business.

However, Oracle’s argument has been that HP refused to play straight with the wider marketplace, insisting that Itanium would be around for many generations to come. Even Intel itself insisted that was true last year, when Oracle first announced that it would stop making new applications that support the chip, which of course led HP to sue last June.

AllThingsD has compiled all of these documents into a single 75-page document; you can read the entire collection below. They’re ordered by exhibit number rather than chronological order. I’ll probably have more to say as I go through them.

HP just sent a statement:

“Intel has provided unequivocal and repeated statements to the marketplace that Itanium is not at an end of life. The undeniable fact is there is committed support for Itanium that extends out toward the end of this decade. Statements that Itanium was at or near an end of life are false. With the unsealing of court filings, the public can see the undisputed facts of Intel’s Itanium roadmap clearly showing a long and sustained future for Itanium.”

An Intel spokesman had no comment, saying it is not a party to the lawsuit.

Update: So let me try to curate these documents a bit. They cover a time period beginning in August 2007, and go through to April 2011. In the first, (Ex. 6) Fink writes to another HP exec and says, “Intel dropped a bomb on us last night” during a meeting that included talk of “canceling Poulsen,” a future version of the Itanium chip that was on internal road maps. Poulson, for the record, is Intel’s code name for a 32 nanometer version of Itanium that was to be released sometime this year. The response: “Call Pat G,” referring to Pat Gelsinger, the once very senior Intel executive who was at one time widely considered to be a possible successor to current CEO Paul Otellini, but is now the COO at EMC, and likely to be its next CEO. Fink’s response: “I did, spent an hour … This was a high tension call.” Intel was worried that an HP server was being built using a competitor’s chip, presumably one from Advanced Micro Devices.

From there, skip forward to Exhibit 43. The email to Fink from Scott Stallard, now a retired HP exec, details a discussion with Intel’s Tom Kilroy, then VP of its enterprise business. The message to Intel: “Don’t possibly signal to the world” the end of the Itanium road map.

The next document is notes from a meeting between Intel and HP led by Otellini and then-HP CEO Mark Hurd in September of 2007. According to those notes, Intel’s Kilroy concedes that the then-current core used to build the Itanium chip has reached its end of life, and that there are two paths forward, one expensive, the other painful — a “crash landing.” Otellini then says that Intel can’t continue to lose money on the product. Hurd goes on to say that HP had by that time sold a combined $9 billion worth of Itanium-based servers, and that it “would be hard to walk away from those customers.”

Otellini then says, according to the notes, that “we need to address the inevitable on the future of Itanium.” Both sides then agree that over the next several years they’d like to glide toward using Intel’s more plentiful and more mainstream Xeon server chips, and essentially keep Itanium alive until 2013. Intel then proposed a new business model that would turn Intel essentially into a contractor for HP. Its goal would be not so much to make money on the Itanium business, but to stop losing money on it. Hurd agrees to take a serious look at the numbers.

Next in line is Exhibit 55, in which Stallard outlines the November 2007 Intel proposal to HP. The key point: HP would pay Intel $488 million over five years to keep building Itanium chips. It includes provisions for an annual “true up,” where Intel gets paid for any difference between its costs and what HP has already paid it for that year. Stallard writes: “So you ask, why should that be that we are forced to true up Intel to break-even? lt is because Tom [Kilroy] says Paul [Otellini] has been consistent on one thing all along, that if we do any other scenario than Tukvale (e.g. shut down the business early) then “Intel can’t lose any (more) money on this thing.’”

Next is Exhibit 15, a PowerPoint deck outlining the strategic rationale for HP to consider buying Sun Microsystems. Key phrase: HP-UX, its version of Unix developed specifically for Itanium servers, “is on a death march” because of Itanium’s inevitable demise. The slide also shows HP’s worries concerning a scenario in which IBM acquires Sun. The key phrase there: Such a deal “Isolates and exposes HP-UX as 3rd tier player, accelerates our decline (product/service) as customers look to consolidate vendors.”

We all know what happened instead: Oracle acquired Sun in 2010 and, in a stroke, took over its server business, giving Oracle an obvious motivation to cut its support for Itanium and hurt its new rival HP in the process. I’ve hit a few high points here, but I think you can get the gist from a careful reading of the documents below.

Update: I’ve re-uploaded the exhibits to Scribd in more logical order, so you don’t have to skip around. Sorry if that was confusing before.

Oracle Itanium Exhibits Chronological


Latest Video

View all videos »

Search »

I think the NSA has a job to do and we need the NSA. But as (physicist) Robert Oppenheimer said, “When you see something that is technically sweet, you go ahead and do it and argue about what to do about it only after you’ve had your technical success. That is the way it was with the atomic bomb.”

— Phil Zimmerman, PGP inventor and Silent Circle co-founder, in an interview with Om Malik