In the 1970s and '80s, Digital Equipment Corp. was the world's top minicomputer company. Then, late in the '80s, DEC found its sales stalling under pressure from powerful desktop workstations, against which DEC's aging VAX computer design couldn't compete. So, the company set to work on a product that would impress its customers and keep them loyal.
The result is Alpha, an all-new design introduced in February. Promoted as the foundation for DEC's most important future computers--from workstations to mainframe-class machines--the technology is designed to sustain DEC for 25 years. At the heart of Alpha is the latest in computer design--reduced instruction-set computing (RISC). That's a way of designing computers so that they can crunch information very quickly, by reducing the number of commands needed to complete a given task. Alpha is at the apex of the RISC performance curve.
Indeed, at first blush, the Alpha microprocessors look scorching hot. DEC says the chips will grind through 150 million instructions per second (MIPS), vs. the competition's processors, which today run in the 25- to 78-MIPS range. What's more, Alpha chips will attack 64 bits of data at a time, not the 32 bits most RISC processors handle today.
SLOW SLOG. Already, the new computer family has gained customers' attention. Last year, Raymond V. Sasso, chief information officer at J.R. Simplot Co., ruled out buying any more VAX minis. But with Alpha, he says, DEC stands a chance of selling Simplot the RISC machines it now prefers.
Just the same, Alpha can't turn DEC's sagging fortunes around overnight. The first Alpha-based computers aren't due out in volume until early 1993, which could give competitors a chance to match Alpha's performance. And those early models will be pricey minicomputer-class systems, not cheap workstations that DEC can use to stave off workstation leaders Sun Microsystems Inc. and Hewlett-Packard Co.
Moreover, Alpha's super number-crunching power doesn't help customers unless they can run their software on the computers. And DEC must catch up with software designed for Alpha. That gives Sun and HP a big advantage: They have been shipping their RISC-based computers for several years, and there are sizable software libraries that work on the machines.
To solve the problem, DEC says it will provide customers with a smorgasbord of services and programming tools to smooth their transition between VAX and Alpha. And DEC also will make Alpha run Microsoft Corp.'s Windows NT operating system, a core program that's expected to attract lots of attention from the PC software market. Indeed, on Apr. 24, Microsoft will disclose plans to convert its own PC applications, such as spreadsheet and word processing programs, to run on Alpha-based computers.
Microsoft's backing should help DEC with another challenge: persuading other computer companies to use Alpha in their machines. That's a must for DEC's bottom line, since it will help offset the tremendous costs required to keep a computer architecture such as Alpha up to date over time. DEC, for instance, has begun work on a $500 million plant to manufacture Alpha chips. So far, DEC has signed up supercomputer leader Cray Research Inc., which plans to build a so-called massively parallel computer that will gang together hundreds of chips to tackle big scientific problems using Alpha technology. Another Alpha licensee: Japanese workstation maker Kubota Pacific Computer Inc.
CENTER STAGE. For now, Alpha is more paper tiger than killer chip: "I don't believe Alpha will be a factor in the marketplace until the 1994 to 1995 time frame," says HP Vice-President Willem P. Roelandts.
That means DEC's VAX will remain at center stage for some time. And in the meantime, the Alpha hoopla is denting VAX sales, which fell more than 20% last quarter, partly because customers are worried Alpha will make the VAX obsolete. "People do not want to buy the old technology, so they're waiting," explains Howard L. Niden, Chicago regional director for Price Waterhouse.
Here's Alpha's ultimate challenge: Capturing customers beyond the current installed base of 500,000 VAXes. As First Boston Corp. analyst Curt Rohrman puts it: "It's hard to grow your company if you don't bring new people into the fold." Indeed, for DEC, that's just about the Alpha and the omega of the situation.