We all have our favorite mainframes. For many it will surely be the new and sophisticated mainframe Z9. For me there are four mainframes that standout among my memories. First was the GE 225 at Lehigh University where I was an electrical engineering student (1963-1967). The programming language used was called WIZ and it was very similar to BASIC. Programs were literally written on paper and then punched into "IBM Cards" using a keypunch. The deck of cards was then "submitted" through a plastic window. Hours later (sometimes days) the results of the program, known as a "printout" were placed in bins where students could pick them up.
I was fortunate to be able to go to graduate school at the University of South Florida part time while I was serving in the U.S. Army at the U.S. STRIKE Command in Tampa, Florida (STRIKE stood for swift tactical retaliation in any known environment). My masters thesis was in operations research and GPSS was the programming language I used to build simulation models. Like using the GE 225, programs were created on punched cards and submitted through a window — this time to an IBM System 360 Model 65. The model 65 was a giant of computing at the time — many times faster than the GE 225. Mainframe #3 is one I got to know up close and personal.
The first mainframe that I got to really know was the IBM 1410 at MacDill Air Force Base in Tampa, home of the STRIKE Command. Both the model 225 and the 65 had a "console operator" who would put the punched cards in the card reader and use commands at the console to tell the mainframe to read the cards and perform the necessary steps to execute the programs. The console operator also would mount magnetic tapes as necessary to make backup copies of data or to load databases. What was different about the 1410 for me was that I was the console operator. I had learned to type in high school, but it was in the army that I learned to type really fast — at the 1410 console. The skill has served me well to this day. The other skill I developed at the STRIKE Command was how to sift through all the decks of cards that had been submitted and pick the one which would take the longest to run. Typically this was an update to the global airfields database. The database was stored on seven reels of tape. I would go to the tape library, get the appropriate tapes, mount them on the six-foot high tape drives, put the cards in the reader, type in some commands at the console and then head for the desk where I did my grad school assignments while the 1410 churned away for hours.
Mainframe #4 was a System 390 installed at the Internet Technology lab at IBM in Southbury, Connecticut. My former colleagues there knew just about everything about Unix and about PC servers, but not much about mainframes. We convinced the general manager of the mainframe business to lend us a mainframe to experiment with and see if we could make it work well with the Internet. It took some getting used to but eventually the team had it integrated with everything. What was unique about this particular mainframe was that it was very compact — much bigger than a PC server but much smaller than most mainframes. It was a "starter" system. One day a customer engineering technician stopped by to help out with something. He took one look and said it couldn’t be a mainframe. We asked why and he said "if you can see over it, it can’t be a mainframe". I don’t know the numbers, but no doubt that the little 390 — about the size of a small refrigerator — was many times more powerful than the 225 or 65.
Perhaps the ultimate mainframe will be the Blue Gene. More on that another time.
Menu
