The death of Intel’s Itanium processors has been trumpeted by the media on many occasions over the last 10 years. Produced in conjunction with HP, the processor was intended right from the start to be a radical experiment, focused largely on solving the performance and scalability issues surrounding high availability servers and scientific computing workloads in the early 90s.
When Itanium first hit the market – very, very much later than originally planned – it became clear that while it had many powerful features, most businesses were more comfortable using low cost generic x86 servers. Leaving Itanium as a specialised server microprocessor, geared towards premium servers trying to deliver mainframe class reliability, availability, and serviceability.
So Itanium has been bucking the market trend for some time, and has been largely irrelevant in server commoditisation, and offered few performance advantages for mainstream enterprise software. As a result, vendor support and interest has dwindled on both the hardware and software side.
The 1990s saw enterprise software vendors typically offer and support several processor and OS platforms for their major products – because the high performance server market was very fragmented. Those days are long gone, with enterprise software vendors now focused almost entirely on developing and optimising for Xeon x86 servers.
While Oracle is taking some heat for recently dropping Itanium from their development roadmaps, they are hardly the first vendor to focus their efforts elsewhere. In 2010, Microsoft announced that Windows Server, Visual Basic, and Microsoft SQL Server would cease future development for Itanium. Intel is still developing new Itanium processors, so Itanium will continue as a hardware platform – but key enterprise software will be frozen at current revisions.
The question for businesses that rely on Itanium infrastructure becomes one of risk management, performance optimisation, and increasingly specialist support skillsets.
Supporting and performance tuning legacy software platforms is part of our daily business at Dataweave. It is not necessary to instantly junk your infrastructure just because a vendor has decided to stop supporting the platform.
However, if your key business applications depend on Oracle databases running on Itanium hardware platforms, when you have the opportunity to refresh your hardware, I would suggest crossing a new Itanium server off the shopping list. I would recommend considering Oracle’s new and hugely successful Exadata database machine, along with the other usual suspects – Intel x86 commodity servers, IBM’s P series, and Oracle’s SPARC M-Series.
If any business critical legacy platform is achieving service availability goals, then securing support resources and skillsets becomes a key factor for your risk management strategy. As transaction rates or storage needs outgrow hardware capabilities, the ability to smoothly plan and execute migration strategies becomes critical to ensure business continuity.