Forget about IBM; it looks like HP might need to buy Novell
I sat down at my desk this morning to find that Mary Jo Foley had already blogged about my planned topic for this morning, the demise of Itanium support in future generations of Windows Server. While I'm sure that generated a collective yawn from the x86 and x64-centric universe, I'll admit that my first thought was "what the heck does this mean to HP?"
A few years back HP shelved their own RISC architecture processors in favor of the Itanium and has built an entire product line and high end strategy, the HP Integrity systems, on the Itanium processor. The end of Windows Server on the processor line is bad enough, but the end of Red Hat Linux support on it as well, bodes ill for the HP strategy. Especially since other Linux vendors are currently quiet on the subject at this point in time.
HP may need to adjust their marketing campaign...
This leaves HP with a very limited selection of enterprise general purpose operating systems for their big datacenter hardware; HP-UX and Open VMS. In a world where broad third-party support is taken as a basic requirement for anything other than special purpose computing, HP-UX and OpenVMS enjoy only a small percentage of the support found for Linux or Windows. It also offers the unattractive situation for HP that the IBM Power Systems will enjoy broader OS support and hence offer a more attractive option to customers in the market for these classes of servers.
Last week Christopher Dawson posted a list of reasons why IBM should buy Novell, but with this announcement by Microsoft and Red Hat it would seem that HP buying Novell (or perhaps HP partnering with Intel to make a joint purchase) would make far more sense right now. Everything that Chris said about IBM buying Novell remains true with HP doing the buying, with the added bonus of guaranteeing that there would be continued development of Linux for HP's flagship processor. HP has made a significant investment in the Itanium-based Integrity servers and the limiting of future options for the processor isn't a recipe for long-term success.