Itanium was initially intended to be the next generation X86 platform, much like IBM intended Power to eventually replace X86 and Sun expected X86 to fail against SPARC.
Of these platforms you could argue that Power had the most success since it existed on both Apple systems and still exists in some form on some game systems today. But none of the platforms ever took the place of the Pentium-based platform now represented by Core on the desktop and Xeon on servers.
Volume rules, but that doesn’t mean there isn’t room for specialized systems targeting unique usage models, which is where these platforms continue today.
Volume parts have economies of scale that allow them to provide cost advantages that can result in their being used in areas that might have been dominated in the past by specialized parts like Workstations. But volume parts can’t be everything to everyone, which is why you have unique custom processors in the imbedded market at one end, and custom processors in mainframe class systems at the other.
Recent generation (Project Odyssey) HP Itanium systems did have one unique advantage: the ability to relatively easily switch them to Xeon-based architectures.
Had Oracle not falsely claimed Itanium was dead, it likely would have chased, or been chased, by Power and SPARC well into the next decade. But Oracle’s moves have severely hurt Itanium sales and HP is about to get what is likely to be a massive check from Oracle to make up for the pain. They could take this check and fund an effort to leapfrog IBM and Oracle with something very different.
The next generation of high end computing looks like it will be largely driven by massively multi-threaded applications, virtualization, and an effort to severely curtail energy use and heat generation. HP has been on the forefront with efforts surrounding both ARM and Atom with their project Moonshot.
While Power, SPARC, and Itanium are largely based on thinking from the 1990s, Project Moonshot is based on current thinking and anticipates the 2020s. With massive funding, something that HP could do with the anticipated Oracle judgment/settlement, HP could leapfrog Oracle in particular (since Oracle would be left financially drained by the judgment/settlement).
This funding could both accelerate Project Moonshot and force it to embrace the current Itanium base of customers much more quickly in order to prevent additional competitive erosion.
Clearly Intel would be willing to assist and NVIDIA (who has been the most active of the ARM vendors in the Supercomputer space) could be engaged as an alternative. But the end result could be a much faster transition to the next generation of technology for HP.
The cost of doing a clean slate project like HP’s Project Moonshot is daunting. And while this kind of thing was very common a few decades ago the money and risk involved appears to be very difficult for firms in HP’s class to accept. This means they often need a huge push in order to fund them fully and get to the next level first.
While it may take a while for HP to collect on Oracle’s bad bet with Itanium, the amount of money they do receive could be in the double digit billions given the massive damage Oracle did to HP. That money focused on moving Itanium customers to a new next generation architecture (and likely partially funding the migrations) could move HP to the front of this line. If it does, they’ll have Oracle to thank though I doubt Larry will appreciate the gratitude since he likely could have used this cash himself (there are still several more Hawaiian Islands he doesn’t own after all).
While this may not convince anyone to move to Itanium it could certainly convince many to stay there with the very real likelihood that the next migration choice from HP could be far better and partially funded by Oracle. Given how many IT folks feel about Oracle and its pricing; this last could be particularly sweet.