The system approach to reliability

Abstract

The title of this paper is intended to indicate concern not so much with the individual techniques of reliability, even though a number of these will be mentioned, but rather with the total impact of these various techniques on the reliability of a data processing subsystem, especially techniques above the circuit-component level. Hence, this paper will discuss the approach to reliability taken in the design of one system which is now reaching the operational stage, and to indicate what sort of results that approach is giving. Then, some more recent developments will be reviewed for their implications to a system now being developed. The type of application being considered is one requiring continuous 24-hour-per-day operation in real time, such as a continental air defense or air traffic control system. Many of the same problems are found in other combat operations center or command control systems.