Veitch Speaks: Intel doesn't own the world

Although Intel has a huge influence on most system developments, computers will evolve and Intel will become more of a back-seat driver, says Martin Veitch
Written by Martin Veitch, Contributor

It is true that Intel architecture is a huge force in enterprise computing and one that is expanding its domain from the desktop through to the data centre. However, it should be noted that Intel is not and never will be the way forward for all of us.

Even devout iconoclasts should admit that Intel microprocessors are the direction for the vast majority of the world's computer makers. That includes those that have previously opted for proprietary routes and those that have selected alternative architectures.

In PCs, Intel remains the dominant force and there isn't any great prospect that that will change despite the best efforts of AMD, Cyrix, IDT and start-ups such as Transmeta. But the real interest is in workstations and servers. A blind man can see that the direction of all hardware makers that don't have a grouse with Intel is towards the Californian firm's chips. The sheer volume with which Intel can pump out parts makes competing resemble the task of Sisyphus -- the Greek chap condemned to keep rolling a boulder up the mountain only to see gravity have its inevitable effect.

Delays to IA-64 mean that Silicon Graphics' MIPS and Hewlett-Packard's PA-Risc get one or two last verses and both firms are smart enough to have acknowledged that. Sun Sparc and Compaq Alpha get longer because of the requirements of the installed base, but it's still a death and taxes probability that they'll have to go sooner rather than later.

But even with Pentium III just around the corner, Intel won't have everything its own way. There's a simple reason for this - you've got to have chips with everything but they don't do a lot on their own. Intel chips are very likely to form the heartbeat of the next generation of desktop, server and mainframe computers, but that still leaves plenty of scope.

I'm thinking particularly of communications, interconnect technologies, storage and graphics. As computer architectures develop and processing power becomes a given, Intel will become more of a back-seat driver. Intel absolutely knows this and that's why you've seen it ooze towards any field it sees as key to driving industry momentum and sales of its core microprocessor business -- motherboards, networking, graphics, communications, bus technologies and even software. For all its mountains of bucks, the level of success has been moderate.

It all comes down to what Americans call sticking to your knitting. Chipset and board makers may be smaller than Intel but that's all they do and they're getting plenty of juice out of the Socket7 layout. There is a similar explanation for the continued success of the giants of graphics and networking. Bandwidth is the biggest technology issue facing the industry and Intel is a bit player in that ongoing drama.

A subsidiary reason for continued competition is that most of the biggest brands, notably Compaq, don't want their futures governed by a dominant Intel. That may explain why the Compaq-HP-IBM PCI-X effort for a next-generation peripheral interconnect technology is gathering pace.

It's a new generation of Intel-centred hardware that will be truly differentiated. Look at Silicon Graphics' new workstations - they're Wintel machines but they're in no way replicas of an Intel-Microsoft blueprint. The same applies to the recent mainframes from Unisys, Fujitsu, ICL, Amdahl and the rest that make Intel chips the pulse of machines that run a variety of operating systems. Linux, the BeOS and Monterey Unix will offer operating system alternatives to Windows.

Intel will remain but the variety of computers based on its chips will expand for a long time yet.

Editorial standards