The acronym CMS originally stood for Cambridge Monitor System, after the IBM laboratory (the Cambridge Scientific Center) where it was developed; it was later renamed by IBM when it became an official IBM product.
CMS was developed in concert with VM to provide a time-sharing system; its original form was heavily influenced by CTSS, which was built by people at the same location (at MIT) as CSC. CTSS was actually used to develop initial versions of CMS software, using cross-assemblers.
Users actually log in to VM and then proceed to boot their own virtual machine, usually by executing a file named 'PROFILE EXEC' or by issuing the command 'IPL CMS'. ("IPL" stands for "initial program load"; standard IBM jargon for booting a machine.)
While CMS started in the era of paper line terminals, by the late 70's most users were connecting with 'full-screen terminals' such as the IBM 3270. Unlike the terminals most Unix users are familiar with, full-screen terminals actually have local storage and minimal processing abilities to deal with an entire screen of information at a time. This approach has many advantages for displaying and editing text and data entry. Because the terminal deals with information in its buffer, users feel that the system is typically more responsive, and overall system performance is enhanced because it doesn't have to deal with every keystroke typed by a user, unlike a typical Unix system.
Two of the commonly used CMS tools include the editor XEDIT and the REXX programming language. Both of these products have been ported to other platforms and are now widely used outside the mainframe environment.
CMS is still in development and wide use today.
See also: VM/CMS for the history of the development of the entire system, including VM