So, it appears that there were no significant announcements about server developments at Intel’s Developer Forum, while at the same time the biggest story from IBM is that the p-Series and i-Series systems are being merged into a rebranded entity. Of themselves neither story is worth too great an emotional outburst, but they are perhaps a sign that, at last, server vendors are starting to realise that they are now updating hardware faster than the users actually need. What they need is the ability to make better – probably re-architected and possibly upgraded – use of the applications and tools they already run on the platforms they already have.  Having announced a batch of new server-oriented developments a few weeks ago, Intel’s biggest server news at the IDF was sight of Tukwilla, the 4-core Itanium processor. That is of only marginal interest to most server users, though HP will pleased to see it and others, such as PSI, could be interested in its ability to run mainframe software on big Itanium boxes a more interesting proposition. There are obvious strong economic arguments for IBM to want to rationalise the hardware of the Unix-running p/Series boxes and the old AS/400 i/Series. They have been using the same processor for some time, so the more the internals of the hardware can be the same assembled circuit boards the bigger the manufacturing savings that are possible. There are some question marks as to how far the rationalisation can go as each system is, historically at least, aimed at different operational and application models and therefore may still require different I/O stages to meet those application needs. Branding issues obviously come in here as a necessary digression, for bringing the p- and i- families together requires a new name. IBM has opted for the safe – and under the circumstances undoubtedly sound – answer of `Power’. After all, both use Power family processors. This does actually also leave the door open to ultimately include the z/Series mainframe family, which runs on a close relative (definite first cousin material) of the Power processor line. That filial relationship is likely to get stronger as well.  

But the more that IBM can bring these application areas towards running on a totally common hardware platform the more sense there is in it for users. In short, it makes opting for the platform that bit easier – and is the essential argument behind common platforms and virtualisation: the users get far greater flexibility without having to revert to too many alternative platform options. They also spend less time – and money – on continually updating to the latest gee-whizz hardware across multiple platform types. IBM looks as though this is where it is heading – certainly a common hardware platform for Linux,Unix (AIX) and OS/400, and an outside, longer term place for z/OS as well. Long term, this may prove to be no bad idea either, as I do suspect the market for the z/Series mainframes is changing. There is growing scope for their application in areas such as web services, as well as others well outside of the scope of the traditional `mainframe’. To catch the wave, however, IBM has to be ready to move with it. The current announcements are not yet sufficient to say this is happening, but they are pointing in the right direction and getting a lot closer to where the future looks as though it is heading.  In fact, a single common hardware platform is a good theoretical goal – but in practice it is probably unachievable. At least two alternates are needed, probably three, which is likely to be the best compromise between real operational flexibility and manageability for both future and legacy systems. Three will cover most O/S and applications bases, yet should not cause too many operational and infrastructure management headaches. And this latest announcement from IBM makes it possible to speculate that the three could be x86 multicore, Itanium multicore and the `Power’ family. That would allow most users to perm any two from the three. In the end, however, much of this will depend upon IBM’s ability to look beyond the comfort of large, established but static markets for its systems and on to the potential that exists for real growth in new market sectors. A slow decline through atrophy may otherwise be a result.