The choice of CPUs and operating systems make changing platforms a real concern
Dennis Allen
Lately, I've been getting mail asking if now is the time to buy desktop systems or if waiting for faster systems to become available is best. At first I was surprised--it's an age-old question with an age-old answer: There will always be something better and faster on the horizon no matter how long you wait. BYTE readers know that, and that they should ask that question at first confused me.
But at the heart of the question is whether now is the time to make a platform change. It's not just a matter of debating the merits of the next Intel 80x86--compatible machine; it's far more complicated and extends beyond CPUs to operating systems and videoconferencing.
On the CPU front, there are a lot of choices, and even more are comi
ng. Even staying with Intel doesn't necessarily make the choice easy. With 100-MHz 486s available, you may question the need to go to the Pentium. And if you are buying systems for dedicated, noncompute-intensive applications, a 100-MHz 486 may seem like overkill. If you've been considering Windows NT or the symmetrical-processing version of OS/2, your decision is further complicated by deciding whether two or more CPUs is the right answer, and whether those multiple processors ought to be 486s or Pentiums.
If your needs today are less critical than they'll be several months from now, you are probably also questioning whether to wait for one of the new Pentium competitors from Cyrix, AMD, or NexGen. Each promises to be a compatible alternative to the Pentium, with different twists for performance gains. And if you are considering one of the Pentium competitors, surely you're also wondering what Intel is doing with the P6, the reported Pentium follow-on. Little is known about the P6. That could be bad n
ews for Intel, because unless it starts releasing real technical information about the P6, it stands to lose significant market share to forthcoming competitor CPUs and to Intel's own 486 and Pentium alphabet soup (e.g., SX, DX, and P54C) of confusion.
Meanwhile, for most companies, staying with Intel-based machines is not etched in stone, and CPU offerings from DEC, Mips Technologies, Hewlett-Packard, IBM, and Motorola may be tempting. All promise higher-than-Pentium performance, and, generally speaking, they all require a move away from DOS and Windows. If you are replacing older Macs, it's no problem; you can run essentially the same operating system and applications on the Power Mac. Otherwise, you are faced with a gut-wrenching operating-system decision.
While the upcoming replacement for Windows--Chicago--brings welcome improvements, it won't support multiple CPUs, and it was not written for anything other than Intel CPUs. So the choice becomes focused on OS/2 and Workplace OS from IBM, or
NT and the promised Daytona and Cairo. And yes, there is still Unix in 31 flavors. The most notable ones are SunSoft's Solaris, SCO's Unix, and IBM's AIX. The latter is interesting because IBM has been demonstrating a version of AIX on its prototype PowerPC-based systems. The wild card is Apple's System 7, which the company plans to port and license to other platforms.
The big problem is that choosing a new operating system is a lot like getting married--if things don't work out, the separation and divorce can be long and painful. Then there's the video question. While most organizations have already made some commitment to multimedia-equipped systems, the step to systems ready for videoconferencing is a little bigger to make.
All in all, it is a complicated problem, and the answers are not easy. Remember that the CPU and operating-system wars are just heating up, and while the editors of BYTE predict that RISC CPUs will dominate the future, no one knows whose RISC CPU will be the clear winner.
The best strategy for the short term is to go ahead and replace systems that must be replaced now. Choose from what's available today. But as you consider systems that can wait until next year, my advice is to try out the different platforms in your environment. Base your decisions on how well a platform's technology fits into your specific needs and on whether that technology will continue to answer your enterprise-wide needs in the future. And we'll do our best to help you sort through it all.
DENNIS ALLEN, EDITOR IN CHIEF
(
dallen@bix.com
)