<http://www.zdnet.co.uk/itweek/columns/2001/31/banks.html>
>
> The going down of the Sun The days of an industry stalwart could be
> numbered unless Sun stops resting on its laurels and recognises Intel's
> Itanium line as a real threat, says Martin Banks
> Here's a heretical question to take away with you on your summer holidays:
> how long will it be before Sun implodes?
>
> This is, I'll admit, absolute speculation on my part, but I am easily old
> enough to remember a time when we all thought Digital would go on forever,
> despite the best efforts of its idiosyncratic leader, Ken Olsen. But that
> industry giant died in the arms of Compaq, which has even seen fit to kill
> off one of Digital's biggest technological legacies, the Alpha processor.
>
> At present Sun seems impregnable. Its large share of the high-end Unix
> server market is likely to increase further, despite the best efforts of
> Microsoft to convince the world that it is a high-end enterprise supplier.
> But it may well be Microsoft's old partner in crime, Intel, that holds the
> key to Sun's future.
>
> The imminent arrival of the 64bit Itanium processor is going to change the
> world in which Sun operates, and I am not sure the company is ready for
> it, or even aware of the dangers. In the same way that Digital's Olsen
> scoffed at Unix, only to see it demolish his treasured VMS-based market,
> so Sun can dismiss Itanium all it wants this will not stop the
> inevitable.
>
> Yes, there are many press reports that suggest Itanium is not that good a
> performer. This might well be true although Intel would dispute it but
> it misses the fundamental point that Itanium is only the start of a chip
> family, and its successor, McKinley, looks like being a very impressive
> performer.
>
> Intel's 64bit processors will go from strength to strength. This cannot be
> said for Sun's own processor family, the Risc-based UltraSparc, which is
> beginning to show its age. Indeed, with Alpha and HP's PA-Risc family both
> under a death sentence, Risc-based processor architectures seem to be on
> the way out. Given the huge investment it now takes to be in the processor
> business at all, Intel could win this particular battle by default anyway.
>
>
> But the most important pointer to Sun's future can be found in
> applications. A platform can only succeed if enough useful applications
> are written for it. There may be a good business case for Linux being on
> corporate desktops, but unless more office software is written for it, it
> is unlikely to take off. The same is true for the IA-64 architecture
> underpinning Itanium.
>
> Luckily for Intel, serious enterprise-oriented applications are starting
> to appear. The announcement by BEA that it is working with Intel to
> optimise the WebLogic applications server for the processor is a case in
> point.
>
> Now BEA is a company that is growing into its future rather than trading
> off its past, and its future has to be on the most prevalent server
> platform to be found in the enterprise/ Internet arena. For now that is
> still Sun, and the two companies have a close relationship. But in two
> years' time?
>
> The IA-64 architecture may not be the best, but it will still be the one
> major server manufacturers like IBM, Compaq, Dell, and HP will focus on.
> Only Sun will refuse to play ball.
>
> Sun may well survive Apple has shown it is possible, but only until IBM
> gives up on the PowerPC. But like Apple, Sun seems doomed to drift into a
> cul-de-sac of its own making. |
