Time-sharing

In 1922, Thomas J. Watson Sr. had a notion that would lead to a lasting impact on computing. As president of the Computing-Tabulating-Recording Company, which would soon be renamed IBM, Watson perceived a large untapped market for the company’s tabulating machines. He knew that many businesses could benefit from using a tabulator to speed operations like billing and payroll yet realized they were too small to afford their own machine. Additionally, many companies that already owned one could certainly benefit from additional capacity — but they were reluctant to buy more machines just to satisfy peak demand.

Watson’s idea was to create a division that would enable such customers to “share” processing time. Users would send data on punched cards to the Service Bureau, which would run the cards through a tabulating machine, produce a report, and deliver the results back to the customer — all for a modest fee.

The process would evolve into a fundamental principle of advanced computing. Time-sharing, as it’s known, is a design technique that enables multiple users to operate a computer system concurrently without interfering with each other. Even though the central computer is accessed by multiple users, every user has the sense they are alone on the system.

Time-sharing not only made computing faster and cheaper but also greatly expanded the market for years to come.

Computers in the 1950s were expensive and highly inefficient compared with later models. They often remained idle while users entered code or waited for additional data or for peripherals like printers or hard drives to catch up. A technique called batch processing, in which previously collected jobs were processed in a single group, was developed to shorten downtime between the execution of one program and the next. But designing a computer that could support multiple users simultaneously was a completely different concept. The “state” of each user and associated programs would have to be maintained internally as the computer switched from one user to the next.

The concept of a time-sharing computer was first described by IBM’s John Backus at a summer session at MIT in 1955. Backus would go on to have a storied 41-year career at IBM and is best known as the father of Fortran, the first widely used, high-level programming language that helped open the door to modern computing. But that was after his time-sharing revelation. “By time-sharing, a big computer could be used as several small ones; there would need to be a reading station for each user,” Backus postulated.

In the early 1960s, an experimental time-sharing system was launched at MIT on a modified IBM 709. MIT added a typewriter input so that an operator could obtain additional answers from the computer on a time-sharing basis with other programs accessing the machine simultaneously. The system, which became known as the Compatible Time-Sharing System (CTSS), was one of the first widely used time-sharing operating systems. Service to MIT users began in 1963 and remained in use until 1973.

Computer time-sharing quickly went mainstream after CTSS. With System/360, a family of computer systems introduced in 1964, mainframes became powerful enough to enable time-sharing. System/360 Model 64 and Model 66 were the first IBM computers specifically designed to allow simultaneous usage. They employed automatic interrupt circuits that let the machines schedule job requests at the computing site and from remote terminals, with a supervisory program automatically controlling allocation. Dozens of users were able to work at the same time without interference — or even being aware of each other.

Time-shared systems provided a nearly instant response for users and opened the door to a whole new form of computing. With time-sharing, users would be able to work continuously on a problem and call on the computer at any time to make partial test runs, helping them to develop new approaches to a problem. Time-sharing resulted in more thorough and cheaper problem analysis.