Never mind the industrywide computer slump and the high mortality rate for fledgling technology companies, Ken Fisher knows that the market is ready for his Encore.
"The minicomputer market is $15 billion today," says Fisher, chairman of Encore Computer Corp., "It will be $50 billion by the decade's end. What I'm basically saying is that 1 percent of that market puts me in the Fortune 500. We're going to get at least 2 percent.
"You don't have to dominate. Winning means, can you get that 1 or 2 percent?"
When Encore formally premieres its new Multimax technology at the National Computer Conference in Chicago this week, the company will quite literally be putting its chips on the table in a bid to establish itself as the next major computer manufacturer. Fisher is betting that Encore has the management team, the marketing savvy and precisely the right technology to catch the next wave of demand for raw computing power.
Under normal circumstances, Encore would be just another Route 128 computer start-up looking to hit the big time. What makes Encore special -- and has the analysts curious and the competitors nervous -- is that it just might be an encore.
The reason isn't simply business -- it's technology. Encore's technology exploits a new style of designing computers that takes them off the existing price performance curve. Where traditional computers improve in price and performance at roughly a 20 percent annual rate, Encore envisions computers that improve at double or even triple that.
That isn't idle speculation by budding entrepreneurs. The company's management is drawn from the very top ranks of the best minicomputer companies that line the Route 128 corridor. Fisher himself was chief executive officer and key marketing strategist of Prime Computer, a company he guided through phenomonal growth until a dispute with the company's chairman forced him out in 1981.
Encore's vice chairman of technology, Gordon Bell, is widely regarded as one of the gurus of innovative computer architecture. He designed the popular VAX minicomputer line for Digital Equipment Corp., the world's largest computer company after IBM.
Henry Burkhardt, the company's chief financial officer, cofounded Data General Corp. the minicomputer company of "The Soul of a New Machine" fame.
"We were a group of guys who had done it before and would do it again," said Fisher, "Nobody in this company is in a job he hasn't done before."
This level of management expertise, highly unusual in most start-up situations, encouraged a burst of venture capital from Hambrecht & Quist, a leading San Francisco venture group, as well as relationships with both Sperry Corp. and Schlumberger Ltd.
When Fisher left Prime, he originally envisioned Encore as sort of an umbrella-marketing organization for young computer hardware and software companies. But conversations with Burkhardt led to an examination of "the cost content of the computer industry," says Fisher, and that, in turn, led to the conclusion that microprocessors -- the silicon chips that run personal computers -- could be incorporated into a new architecture that made computers faster, more powerful and cheaper.
"We'd use microprocessors to create a new cost structure for the industry," Fisher recalled.
Fisher then called Gordon Bell, the architect of DEC hardware who doubled as a computer science professor at Carnegie-Mellon University, to see if he'd be interested in joining Encore.
"We spent 11 hours together," said Fisher, "We both agreed that multiple processing was the future."
Bell, who had dabbled with all kinds of computer architectures as both an academic and an engineer, had a specific multiprocessing architecture in mind and came on board in July of 1983.
For the most part, computers rely on a single processor to perform calculations. Computer designers look for ways to make that single processor calculate faster as well as ways to put data into the processor more efficiently. Most supercomputers are single processor devices that are designed to maximize processor speed -- which is usually measured in MIPS, for millions of instructions per second.
Logically, it would seem that two or three processors could perform calculations faster than a single processor. Technically, that's absolutely true. A big problem could be broken into little problems and each of the little problems assigned to a processor.
The catch is that the assumption is a large one. It's genuinely difficult for the computer to figure out when to use which processor in an optimal fashion. Obviously, it's uneconomical to buy a multiprocessor computer if most of the processors remain idle. Similarly, its difficult to craft a communications link between the processors to make sure that the right data goes from processor to processor and is correctly assembled into the final answer.
For these and other reasons, multiprocessor computers have never won widespread acceptance as a general purpose calculator.
What Encore has done, however, is figure out how rows of microprocessors can be placed on an ultrahigh-speed link -- called a bus -- to perform multiprocessing calculations. The virtue of microprocessors is that they're cheap and easy to obtain. The real hardware challenge was designing the data bus which enable the microprocessors to communicate with each other to share their data at high speed. The bus, which Bell designed, can carry up to 100 megabytes -- 100 million bytes -- a second.
What's more, Bell designed the computer so that additional processors could easily be added by simply placing another microprocessor board in a slot on the bus. Consequently, one can add to the processing power of the computer by simply adding another 1.5 MIP board, just like popping in another sparkplug.
What's critical is that this computer isn't idiosyncratic when it comes to software. It uses the UNIX operating system, a set of software instructions that are widely used by minicomputer manufacturers to run on their hardware. Thus, the Encore computer can use an existing software base.
The result is the Multimax, Encore's flagship product, which combines from two to 20 of National Semiconductor's 32032-model microprocessors. With two microprocessors to a board, the Multimax can run at speeds from 1.5 MIPS to 15 MIPS at prices from $115,000 to $600,000.
That's very competitive in price-performance terms with Data General, Digital Equipment and Prime minicomputers.
"I think they've really stabilized some new and exciting technology," said Tom Wilmott, an analyst with International Data Corp. "Gordon Bell's bus is really a tremendous technological achievement."
Technology aside, though, the multiprocessor approach using off-the-shelf microprocessors yield a tremendous price-performance plus. "When you're talking about the 7 MIP to 9 MIP version, it essentially doubles the processing power for the price with its competition," Wilmott asserted.
"By 1990, all minis and mainframes will be designed this way," asserts Bell.
Fisher is initially targeting the 2.5 percent of the computer-buying marketplace commonly described as "early adapters." These sophisticated computer users should have the expertise to write the software that takes advantage of the Multimax's multiple-processing capabilities.
But he is quick to add that Encore is not a "niche" company seeking a narrow band of users.
"Point product strategies are doomed to failure," Fisher contends, "To succeed, a company needs a family of products."
Encore is also introducing a local-area network capability and support software for its Multimax. Fisher stresses the company will be more than just a manufacturer.
Though the company clearly has both a technical and a management edge, there are significant concerns. One major technical concern is whether most users have problems that lend themselves to cost-effective solution by multiprocessors. Simply speaking, it's possible that single-processor machines are the most cost effective for solving most problems. Bell concedes this is a concern but not a probability.
On the business side, Encore's stock has slid nearly $1 from $5 since the company went public earlier this year. The computer business is a capital-intensive business. Fisher concedes that "I would have like to have raised more money." If Encore's sales lag and the cost of providing customer support stays high, Encore might have to seek new equity partners to survive.