Itanium in Japan and Emerging Markets
One of the more interesting twists in the evolution of Itanium is the relatively high adoption rates in the developing world. Anecdotal wisdom (and most analysts) said that IPF would only be a clear success in situations where customers already had legacy systems being replaced by Itanium. The biggest examples are clearly the HP-UX, OpenVMS and Nonstop user communities, and to a lesser extent, IRIX, GCOS and ACOS users. Despite this common sense, IPF is doing extremely well (by the system revenue metric) in Japan, and emerging markets.
The explanation for IPF’s success in emerging markets, such as Korea, Russia (where it out ships all RISC competitors) and China, is fairly straight forward. Unlike the US and EU, there is no highly entrenched market of Solaris, OS/400 or AIX users to compete with. Moreover, since many of these customers are purchasing without any legacy constraints, they are relatively uninterested in proprietary UNIX based systems (and any associated vendor lock in). Linux is really the UNIX of choice, based on developer momentum and the cost and localization advantages. Windows is also an option, since it is ubiquitous and easy to use. Linux on SPARC is not very popular and handicapped by almost no developer attention (as one example, there is no modern JVM available for SPARC/Linux) and IBM has had relatively little success stirring up Linux customers for the POWER5+ systems. As a result, Itanium becomes the obvious candidate for a scalable and highly available system; both Windows and Linux are supported and popular, and Itanium based systems are often available from local vendors, which is an added bonus.
In Japan, the success of IPF is largely due to political goals and cultural and business norms. There is a very large entrenched customer base in Japan, but the system integrator/customer relationship is very different from in the US or EU. Companies like NEC, Hitachi and Fujitsu really act like ‘trusted advisors’ to customers, rather than simply a merchant hauling out the latest in technical wares. This full service approach means that customers are more likely to accept the recommendations of their IHVs. The three major Japanese vendors all have their own in-house Itanium systems, and are marketing these as the future of high-end, scalable and reliable computing. So it is hardly surprising to see a strong uptake. Moreover, the Japanese government has established that national supercomputing capabilities are a political priority, and heavily subsidizes bids from Japanese companies, many of which position IPF for HPC. Together, the close relationship between system integrators (who all back IPF) and customers, and Japan’s emphasis on home-grown super computing have pushed IPF into the lead in the Japanese market, where it out ships both PPC and SPARC by system revenue.