As Hewlett-Packard and Oracle battle in court over whether Oracle broke an agreement with HP to continue to port its software to HP's high-end hardware platforms, one thing is clear: Customers who love Oracle on HP-UX/Integrity are going to have to make the hard choice of giving up one or both platforms. If the stronger affinity is to Oracle, then customers have lots of choices. If the affinity is to HP-UX, customers are pretty much out of luck.
HP's arrangement with Intel for Itanium is that Intel works as a contractor, designing and manufacturing the chip for HP. According to court documents, HP has shelled out hundreds of millions of dollars to Intel in this relationship. While HP could continue doing that, without Oracle software there's virtually no point, and without Itanium, there's no path forward for HP's NonStop, OpenVMS, and UX operating systems.
To make matters worse for HP, SAP has discontinued support of its Business Objects software on Itanium. HP could port its operating systems to Intel's Xeon chip, but there's no chance now that Oracle would support such a move, and it is unlikely SAP would either.
Note that HP's external messaging directly contradicts what I'm saying here. Here's what HP said in documents relating to its Kinetic strategy: "HP is committed to the continued innovation of NonStop, OpenVMS, and HP-UX on Itanium-based Integrity servers. At the same time, HP is developing enhanced services, highly scalable and resilient x86-based servers, and is working with Microsoft, Red Hat, and open source communities to deliver a Unix-like mission-critical experience for the x86 environment. HP has no current plans to bring HP-UX to the x86 architecture."
That strategy probably won't stand up to shareholder scrutiny. The loss of HP-UX hurts HP much more than the loss of Itanium since HP supplies the entire ecosystem, including consulting, for HP-UX. A normal end-of-life process for the chip could have resulted in a lot of consulting for HP. Indeed, at one time HP had a plan to port HP-UX to x86, which would have at least provided an avenue to a safe landing for its customers. The time for such a port was long ago, like shortly after Sun ported Solaris to x86, so at least the decision now not to port HP-UX to x86 makes sense.
Instead, HP is embarking on a three-pronged strategy to bolster its server business, called Odyssey, Voyager, and Moonshot.
Moonshot is by far the most interesting, forward-looking, and potentially disruptive of the three projects. It calls for developing high-density servers using low-power CPUs such as Intel's Atom. Earlier this month, HP and Intel announced that the Gemini line of servers would use an Atom chip called Centerton specifically made for the job. The idea is to put thousands of 64-bit x86 servers in a rack. Think of this approach as undoing the virtualization trend, which broke large servers into small servers. So-called microservers cut out the virtualization middleman and provide scads of cheap servers to use as you will. As interesting as Centerton and Gemini are, they really aren't an answer for HP's Business Critical Systems (BCS) customers.
The Voyager project is, in my mind, the least interesting of the three projects. The first instantiation of Voyager are the Proliant Gen8 servers HP released earlier this year. The notion is to make servers smarter so that they're easier to manage--an approach that was disproven more than a decade ago by Compaq, which nearly ruined itself by throwing lots of stuff into the server that did nothing for the system's actual performance.
HP says the Proliants can monitor up to 1,600 system parameters so that system admins will know about potential failures long before they occur. The problem is that almost no one wants to pay for much "added value." After all, the system is still running the same Xeon with the same motherboard, bus, memory, and disk subsystems as everyone else. The value add comes from Intel, Microsoft, and the Linux community. Most buyers would prefer that HP just slap in a power supply, add sheet metal, and call it a day. Buyers of these systems are sensitive to price, so there's only so much innovating HP can do without pricing itself out of competition. This project also doesn't do much for BCS customers.
That leaves us with Odyssey, which is aimed directly at BCS customers. The goal for Odyssey is provide a path for NonStop, OpenVMS, and HP-UX customers to Windows and Linux running on high-end HP systems such as the Superdome2 and HP BladeSystem C. As part of Odyssey, HP says it will continue to release new versions of NonStop, OpenVMS, and HP-UX on these systems, but that's predicated on the availability of Itanium chips. Given all that's gone on, we expect that the two versions of Itanium now under development will be the last. At least in the long run (past 2016), the Windows and Linux components of Odyssey will be about all that matters.
HP's goal for Windows and Linux is to beef them up with some of the features that its users like best on HP-UX and NonStop. In its press release on Odyssey, HP talks about supporting nPars (its proprietary virtualization system) and a number of diagnostic systems and fault tolerant features, a bit like the Voyager project on steroids but with a more receptive audience.
While the audience may listen, they'll also shop around. We suspect that HP will have a hard time holding on to this high-margin business no matter what it does.
At this year's InformationWeek 500 Conference C-level execs will gather to discuss how they're rewriting the old IT rulebook and accelerating business execution. At the St. Regis Monarch Beach, Dana Point, Calif., Sept. 9-11.