The Ongoing Story of the Route 128 Minicomputer Industry
When we are riding a bus along Massachusetts Avenue in Boston, a man strikes up a conversation with my wife and me. “You know, I’m a ‘DEC:ie,’ that’s what we used to be called when I worked for DEC, Digital Equipment,” he tells us.
It’s been 10 years since the company ceased to exist, after being bought by PC manufacturer Compaq in 1998. When people still, a decade later, regard themselves as “DEC:ies,” it says something about the impact DEC once had, as the first, largest, and most successful firm in the once flourishing Massachusetts minicomputer industry.
When the cluster was at its heyday during the 1970s and 1980s, it was home not only to DEC but to a number of other minicomputer manufacturers, including Wang, Prime, Data General, and Apollo, as well as suppliers of everything from motherboards to software. Today, it may hard to believe that only 20 years ago Route 128, the beltway around Boston, rivaled Silicon Valley as a symbol of innovation and cutting-edge computing technology. Nobody talks about minicomputers any more, the firms have virtually all disappeared, either by being bought up or by going out of business. But the legacy of the minicomputer cluster is still important to the economy in several ways; in fact, some of the product lines that originated from it way back when continue to generate billions of dollars in revenues.
I have my own memories of Digital Equipment. In 1988, I interviewed the company’s founder and CEO Ken Olsen for the Swedish magazine Ny Teknik, where I will return to work after finishing my fellowship at Xconomy in a few days.
The company’s headquarters was in an old textile plant called The Mill—a huge maze of dark red brick buildings in Maynard, MA. The place had a certain austerity to it, with Ethernet cables running on naked metal cable troughs through corridors and offices, and a reception lobby furnished with rustic shaker chairs.
Olsen and Harlan Anderson, his coworker from MIT’s Lincoln Laboratory, had founded DEC in 1957 with $70,000 in seed money from venture capital pioneer general Georges Doriot. They put out their first minicomputer, the PDP-1, two years later. It was followed by an enormous range of other products, not only computers like the PDP-8, PDP-11, and the famous VAX models, but also terminals, printers, communication units, and software.
Ken Olsen was an impressive, tall man who downplayed his own importance during the interview, talking about helping other people grow and delegating responsibilities. But at the same time, he was known to be strong-willed and emotional.
I got a small taste of his well-known hot temper when I asked why DEC was so ambivalent in its support for the Unix operating system—a common complaint among users of the company’s machines. “Nobody supports Unix more than we do, it’s running on more of our machine’s than on any one else’s,” he told me. After the outburst, I decided not to confront him with his notorious characterization of Unix vendors as “snakeoil salesmen.”
When we met, DEC had grown to be the second-largest company in the computer industry. Olsen was generally regarded as one of the USA’s most successful managers, and many expected DEC to surpass number one IBM in sales within a few years. But in reality, the company had already reached it peak. The staff had expanded rapidly, but sales were stalling and profits had begun to erode. Ken Olsen abruptly left the firm in 1992, six years before Compaq bought the troubled company. DEC’s sale also heralded the end of the Massachusetts minicomputer era.
It is hard to define a minicomputer in specific technical terms. Firms like Apollo, Prime, and Data General each had different approaches, and it might be easier to regard “mini” as more of a cultural identity. The term set the industry apart from the traditional mainframe builders of that time such as IBM or Univac. A fascinating account of the inner culture of a minicomputer company is Tracy Kidder’s 1981 book The Soul of a New Machine, in which the author closely follows a development project at DEC spin-off Data General, a company that had been founded by PDP-8 chief designer Ed DeCastro.
Minicomputers were of course smaller and less expensive than mainframe machines, but they were also much more geared to real-time processing and direct user-interaction. Ideally, as a researcher or engineer you could have your own minicomputer instead of relying on a centralized computer department. In a way, minis were forerunners to the personal computers that would eventually supersede them. The machines even looked trendy, with colorful front panels giving a hint that of their personal nature and that it could even be fun to use them. Paul Allen, the billionaire Microsoft co-founder, writes about his memories of the mini-computer era on PDP Planet, a website featuring his big collection of DEC hardware: “I certainly had a ball using them in their heyday—from the late 1960s to the early 1980s! During that period almost all Microsoft development was done on these platforms.”
Much has been written about why DEC ultimately failed, especially in light of its so very successful early history. Explanations vary from Olsen’s shortcomings, particularly his inability to adjust to the new era of personal computing, the company’s engineering culture that has been said to disregard the business side, and so on.
Long before the demise, Gordon Bell, for many years DEC’s chief architect and later employed by Microsoft Research, identified the pressures the company would encounter from the new microprocessor technology, saying that “Technology advances in semiconductors, storage, interfaces and networks enable a new computer class (platform) to form about every decade to serve a new need.”
And when AnnaLee Saxenian, dean of the School of Information at U.C. Berkeley, compared Route 128 and Silicon Valley in her 1994 book Regional Advantage, she found that “Silicon Valley developed a decentralized but cooperative industrial system while Route 128 came to be dominated by independent, self-sufficient corporations.” In other word, the Massachusetts mini-computer industry was vertically integrated, with companies manufacturing a large part of their components and peripherals themselves instead of using a fully developed network of independent suppliers. That made it less able to adapt to changes than its counterparts in California, which could utilize the advances made by others, instead of developing everything on their own, according to Saxenian.
Not only DEC would disappear, by the 1990s so would most of the rest of the Route 128 minicomputer companies. Even The Computer Museum in Boston, which was founded with donations from DEC, has disappeared—moving its collections from the minicomputer era to the Computer History Museum in Mountain View, in the heart of Silicon Valley.
Nevertheless, that doesn’t mean the minicomputer cluster’s impact on the economy just went up in smoke. A recent report on the manufacturing industry in Massachusetts points out that the rapid expansion of the minicomputer industry was critical to the region reinventing itself as a high-tech manufacturer at the very time such low-tech sectors as textiles declined.
That high-tech manufacturing base, in its turn, has given the region some long-term economic advantages that continue to this day. “In terms of the gains in the 1980s, Massachusetts had a run up in its per capita income compared to other parts of the country,” says Lynn Browne, executive vice president and economic advisor at the Federal Reserve Bank of Boston. “That has not fallen back; in terms of capita income, those gains have stayed.”
For one thing, says Browne, “The people in the industry did stay and they showed up in other places even if for a while it was hard to see them, because there aren’t many large-scale employers in the region, with the exception of EMC. In the late ’90s Massachusetts was very active in the telecommunications sector, so my suspicion is that a large number of these persons went into telecomm,” says Browne.
What’s more, it appears that many Route 128 minicomputer companies didn’t disappear by simply going out of business—most were bought by other firms. In some cases their product lines continued to exist for a long time—it wasn’t until last year that Hewlett-Packard, which had acquired Compaq, stopped taking orders for new AlphaServer machines, a line originally launched by DEC in 1992 that had roots in the VAX architecture originally developed by Gordon Bell in the 1970s.
Another example is EMC, today one of the region’s largest companies. It started in 1979 as a supplier of memory boards for minicomputers. In 1999 EMC bought Data General and in that way got hold of the sophisticated software technology behind the minicomputer manufacturer’s Clariion hard-disk arrays. Since then, EMC’s Clariion line of disk arrays has accounted for billions of dollars in revenues, far more than Data General ever got from its computer business.
* * *
An exception to the extinction of minicomputer firms is Stratus Technologies. The Stratus headquarters is like a piece of Silicon Valley in New England, a complex of low buildings in a park-like environment that makes 110 Powder Mill Road on the outskirts of Maynard, MA, eerily similar to the offices and labs along Page Mill Road in Palo Alto. This used to be the head office of Digital Equipment Corporation in its final years, after it had moved out of The Mill.
Of all the minicomputer firms that once made up the Route 128 cluster, Stratus is the only one still operating under its old name. I visited the company to find out more about what made Stratus from its employee No. 5, software engineer Paul Green. He joined the company, then based in Marlborough, MA, in April 1980. Stratus started that year as a maker of fault-tolerant mini-computers, machines that had more or less every function doubled or quadrupled to ensure extremely high reliability. Since then, the company has gone through a series of transformations in the form of buyouts, mergers, and spinoffs before emerging in its present form.
“I think there are several reasons why we’re here and the others are not,” Green says. “It is an interesting question, and one that I have thought about a great deal. Innovation and consolidation are opposing forces. At some point an idea wins out and is adopted as a standard, a real standard or a de facto standard. When I started in the industry hardware was proprietary, the operating systems were proprietary, the programming languages were at least semi-standard. Now we take it for granted that there is all this standard software. It is still a lot of innovation going on, but it has moved into new spaces.
“This whole process is very brutal, there are great forces at play. The PC has changed the whole pricing scheme for computers. It has in many ways sucked the profits out of the hardware business, it is brutal to be a hardware firm. That is basically how you knock out the established companies as an upstart, you change the rules of the game.”
Instead of seeing itself as a hardware manufacturer, Stratus identified its core competence as being a provider of fault-tolerant, high-reliability information processing systems. The software can either run on Stratus high-end proprietary hardware or on standard server platforms. Last month, the company even launched its first pure software product, Avance, which transforms two ordinary servers, regardless of manufacturer, into a computer system with more than 99.99 percent availability.
Whether in the case of an individual company or an entire sector, the ability to reinvent oneself has been a hallmark of New England innovation. As Michael Best, retired co-director of the Center for Industrial Competitiveness at the University of Massachusetts at Lowell, observed last year in the Boston Globe, “Massachusetts has lost more industries than any region in the world, but has also created more industries than any other region.”