Ken Olsen, the co-founder and long-time leader of minicomputer originator Digital Equipment Corporation died last week at 84. Perhaps more than any other man, all of us in the midrange owe Olsen for our paychecks and for the innovation that his engineering mind brought into being.
I was born the week before the first true minicomputer, the PDP-8, was brought into the world. At the time, the PDP-8 was noteworthy because a base configuration cost only about $18,000 with 4 kilowatts of 12-bit core memory and a teletype for input and output. After a bunch of DECies defected in 1968 to create Data General (which was eaten by EMC in the 1990s), arguably one of the best systems to come out of DEC, the 16-bit PDP-11 mini, and in 1976 DEC made the jump to 32-bits and virtual memory addressing with the VAX line, and 1978's VAX-11/780 and its VT smart terminals quickly became the king of the midrange. By the mid-1980s, DEC had added 10 Mbit/sec Ethernet and VAXcluster clustering software to the VAXes, allowing for clusters to be created and to run applications in a shared mode.
When Compaq was busy going broke in the wake of its 1998 acquisition of Digital, one of the secret sauces that Compaq sold off to Oracle was the VAXcluster and TruCluster microcode that was implemented in these VAXes; you know it today as Oracle's Real Application Clusters (RAC) extensions to the Oracle database, and it was in Digital's homegrown Rdb database way before Oracle got its hands on it. So, given all of this engineering excellence, driven by Olsen and some of the smartest people in data processing at the time, it was no surprise at all that Digital grew to over 120,000 employees, had stolen huge chunks of business away from IBM, and was the second largest computer company in the world in the late 1980s, peaking at $14 billion in annual sales.
That pressure from DEC, more than anything else, made IBM focus and create a little thing they called the Application System/400 in 1988. Digital went on to make plenty of mistakes--not taking the threat of open systems and application portability seriously enough until it was too late and ditto for PCs--but its VAX and Alpha processors were second to none and its systems software was also second to none. DEC was an old-school, engineering-driven IT company and it did not see the threat that volume PCs and their processors would have on all systems. Just like Intel, in public at least, is in denial about ARM processors.
My wife went to MIT, where Olsen got his degrees in electrical engineering, and I have walked the halls of the Lincoln Lab where Olsen and his peers, working for the U.S. Department of Defense, did their work in defining what a computer was in the wake of World War II. It must have been a lot of fun to be Ken Olsen most days.
Say what you will about Olsen, but he thought that PCs were toys and that Unix was a bunch of "snake oil" and that Digital's own proprietary operating system, VMS, was superior to Unix and then Windows. His company's vision of clustered minicomputers with smart terminals is not all that different from the cloudy future we all seem to be unavoidably moving toward. Olsen, no doubt, would have said that a graphical user interface was for sissies and that there wasn't anything you couldn't have done with a properly configured VT terminal.