The Big Squeeze Facing Digital

By John Markoff
Special to the New York Times

April 5, 1988

Maynard, Mass. -- With an average annual growth of 20 percent in the last four years, the Digital Equipment Corporation has been forcing the computer industry's leader, the International Business Machines Corporation, to look over its shoulder constantly.

Yet despite its success, Digital, too, finds it is spending more and more time watching the rear-view mirror. Digital first grew to prominence by changing the definition of a computer from a room-sized machine to a small, easily maintainable office and laboratory appliance that sat in a corner of the room. But it is facing new uncertainties in which new companies making powerful and often even smaller computers are challenging Digital in the same way it challenged I.B.M.

This development is becoming familiar in the computer industry: small companies using newer semiconductor technology nibble larger and larger chunks of business away from their mature competitors..

Digital counters that anyone can build a ''little hot box'' - its term for fast and inexpensive machines made by competitors like Sun Microsystems Inc. But it says that none of its challengers have the comprehensive computer networking capabilities that are standard features of Digital systems.

''They're just selling snake oil,'' said Kenneth H. Olsen, Digital's founder and president. ''The real job is to tie everything together.''

The minicomputer business, which Digital pioneered, partly displaced large mainframe computers in the late 60's and 70's. Then in the 80's aggressive new high-tech companies used the the new 32-bit microprocessor to attack the markets of both mainframes and minicomputers. The 32-bit chip led to a new type of machine, the superminicomputer, which offered high-performance at a lower cost.

''Digital is caught in a two-way squeeze play,'' said Chris Drahos, a vice president at the Elxsi Corporation, a computer maker in San Jose, Calif., that is one of more than a dozen companies hoping to steal market share from Digital. Analysts say that at the high end of the market Digital still does not offer the power to compete directly with I.B.M.; at the low end it is confronted with the competitive and less profitable desktop computer market, a problem highlighted by its lackluster entrant in the personal computer market, the Rainbow.

A New No. 2

After years as the nation's second-largest computer manufacturer, Digital dropped to third place last year when the Sperry Corporation and the Burroughs Corporation merged and formed the Unisys Corporation, providing additional strong competition.

I.B.M. still dominates the mainframe market and is considered the leader in personal computers as well. However, this year, Digital will control almost a quarter of the market for the mid-sized computers used by scientists and engineers for number crunching and increasingly used by individual departments in the nation's largest businesses.

Digital today is remarkably different from the minicomputer maker whose customers once were primarily engineers and scientists.

Analysts said the company, which grew out of computer research projects conducted at the Massachusetts Institute of Technology in the 1950's, transformed itself by offering superior networking and software compatibility among its products -an area in which I.B.M. has lagged. And it has continued to make inroads against its larger rival by providing the same high-quality service and support that I.B.M. is known for.

Better Communications

The key factor, however, has been better computer communications technology, which has permitted Digital to penetrate some of I.B.M.'s largest corporate accounts. In contrast to I.B.M., which is hampered by product lines that run different software and do not communicate, Digital's computers, from desktop work stations to clustered minicomputers, all run the same software and are based on the same hardware design.

Digital's timing has been fortunate. As corporations have moved to automate and integrate business functions as diverse as manufacturing and accounting, the company's distributed computing design, which ties together large and small computers in networks, has proved to be the wave of the future.

The strategy has paid off handsomely. By keeping costs under control and taking market share away from I.B.M., Digital increased profits to $1.14 billion last year, up 88 percent from 1986, on sales of $9.39 billion.

Some analysts now believe that Digital has surpassed I.B.M. in providing high-quality cradle-to- grave support for its machines, from programmers writing special software to service personnel at customers' office. Sales from service-related functions have risen to 33 percent from 24 percent over the last decade. Until last year's stock market crash, the company had been expanding its sales force as well as its service and support staffs.

Some Slippage

However, the shift in emphasis from technology to marketing and sales has meant that Digital has lost some of its edge as a leader in advanced computing technology.

The company has fallen behind in offering high-performance computers at low cost. Its processors are more expensive and slower than those of a number of smaller computer makers, analysts say. And only in the last several weeks has Digital introduced multiprocessing, an important technology that permits computer makers to chain several processors together so that they can work in tandem. Competitors have offered that feature for several years.

But the company's designers contend that critics are measuring the wrong things. ''Our fundamental strategy is distributed computing; we believe that you should pick the right size machine and put it as close to the user as possible,'' said William D. Strecker, Digital's vice president of product strategy and architecture.

New Partnerships

However, to protect itself in the short term, at both the low end and the high end of the market, Digital has formed partnerships with Cray Research Inc., the leading supercomputer manufacturer, and Apple Computer Inc. But such alliances are only stopgap measures, analysts say. They believe that Digital must respond soon with its own small desktop computer and its own supercomputer if it is to remain competitive.

Makers of computer work stations, like Sun Microsystems and Apollo Computer Inc., and start-up manufacturers of superminicomputers, like Sequent Computer Systems Inc., Multiflow Computer Inc., the Convex Computer Corporation and Alliant Computer Systems, are applying pressure that threatens to stall Digital's spurt.

Work stations are the next level up from personal computers and have generally been used for scientific and engineering work. The superminicomputers approach the power of a mainframe at the cost of a minicomputer.

''The real problem for D.E.C. is that they've built themselves on high prices and high margins,'' said William Joy, vice president for research and development at Sun Microsystems. ''It's a problem presented by high technology. When you're a large company, innovation works against you.''

Orders Down

Despite Digital's optimism, there are signs that it faces tougher going. Last month company officials told financial analysts that uncertainty about the economy had slowed orders for large computers. While lower-cost work station orders are up significantly, they have not compensated for lost sales at the high end of the market.

Digital officials also concede that some portion of the slowdown may be due to what is known as ''downsizing'' - the arrival of speedy and inexpensive 32-bit microprocessor machines that offer the same power as minicomputers but at a far lower price.

For Digital, the microprocessor invasion signifies a need for increased innovation. During the last several years Digital has set up new computer research facilities in Silicon Valley to tap the best computer designers in that region.

The company's computer designers are confident that they will soon provide some surprises. ''We believe that we can make substantially more powerful machines in the very near future,'' Mr. Strecker said.

New Work Station

Industry sources said Digital would soon raise the stakes in the market for transaction processing equipment used in such applications as airline reservation and electronic banking systems. And later this year, Mr. Olsen said, Digital will offer a low-cost, high-performance work station, a move he believes will return the company to the top of that market.

Yet Digital still has a fundamental problem. Its most serious competitors, Sun Microsystems, the Hewlett-Packard Company and Apollo Computer, are all betting on a new computer design philosophy known as reduced instruction set computer, or RISC - a simplified system that increases performance by eliminating instructions that are seldom used. But Digital is still wedded to its Vax computer design, now more than a decade old. Vax's strength is that it has been extended to almost all Digital computers. Its weakness, critics say, is that it may prove too expensive to stay in the performance race with its competitors' stripped-down RISC designs.

''I would run, not walk, from the Vax,'' said Gordon Bell, who during the 1970's led the design of the Vax as Digital's vice president of research and development. Mr. Bell now holds the same position at the Ardent Corporation, a Silicon Valley start-up that last month introduced a graphics supercomputer that, according to Mr. Bell, has 10 times the performance of Digital's computers at one-fifth the cost.

To achieve a breakthrough in computer speed, RISC designers argue, it is necessary to simplify computer hardware design with a smaller set of computer instructions that run faster. As a consequence, what has been Digital's greatest strength - a single computer design across its entire product range - may soon become its greatest liability.

'Having All the Pieces'

Mr. Olsen argues that far too much attention has been paid to new computer designs. ''RISC is not the key thing,'' he said. ''The key thing is having all of the pieces that make it all fit together.''

Yet despite the the economic uncertainties and the specter of the 32-bit microprocessor, nobody is counting Digital out. The company has learned vital lessons and achieved a measure of its success by imitating I.B.M., Mr. Olsen said. ''Our position in the world has changed quite a bit,'' he added. ''We've been working awfully hard to look just like I.B.M.''

Still, industry analysts agree that continued growth of 20 percent annually at Digital will be difficult. ''The last five years have been wonderful,'' said David Wu, a financial analyst at S.G. Warburg & Company. ''The next five will be more challenging.''

GRAPHIC: graphs of Digital's stock and market share over past 10 years (Source: Shearson Lehman Hutton); photo of Kenneth H. Olsen of Digital Equipment