In Mainframe We Antitrust: System/360 Compels System/3
April 14, 2014 Timothy Prickett Morgan
The mainframe, as first embodied in the System/360 and carried forward to the System z, celebrated its 50th birthday last week. And while we technically do not cover the mainframe in this publication, the very fact that IBM created the mainframe and was paranoid about losing this fast-growing and lucrative business is perhaps the very reason that the IBM i platform as we know it exists.
I am not going to go through the entire history of the IBM mainframe line. So many others did that last week, and that is all interesting and fun reading. Over at EnterpriseTech, I reminded readers that the System/360 was not just a revolutionary machine embodying some of the core ideals of, well, enterprise technology, that we think of as normal, but it was a system with business practices that were created under the constraints of the 1956 Consent Decree that IBM signed with the U.S. Department of Justice to end a lawsuit from 1952 accusing IBM of monopolistic control of the punch card tabulation machine market. It is intuitively obvious that IBM had a monopoly back then, and with the System/360, the wild popularity of the mainframe gave IBM much more control of a much larger market. In fact, it was the mainframe that made IBM the blue chip stock that it has been since 1960s and that turned IBM into a giant of computing, manufacturing, research, and sales. IBM didn’t have to do marketing back then so much. People were lining up for these machines, which had a common set of operating systems and application runtimes and which allowed for third-party peripherals and third-party financing, among many other innovations.
Here’s the fun bit. IBM was always worried that its precious mainframe business would be broken up in some way if the Antitrust Division or the U.S. Federal district judge David Edelstein, who presided over the IBM case, decided that IBM was misbehaving. And so, when IBM decided to go into the midrange business, it did so with the System/3, a box that was incompatible with the System/360–thus violating the core tenet of the venerable mainframe. The System/3 hailed from Rochester, Minnesota, nowhere near the upstate New York stomping grounds of the punch card and mainframe businesses, although that was close to supercomputer makers Control Data and Cray and some very innovative computer science programs at the Big Ten schools.
Yes, I know that IBM had Report Program Generator on earlier mainframes as well as the System/360 and some of its follow-ons, but I just wanted to point this out: If IBM was not paranoid about protecting the mainframe, you would all have been COBOL and CICS programmers and your companies would have been paying at least twice as much per transaction over the many decades years.