Reliable Computing Systems (1977)

Author(s): Randell B, Lee PA, Treleavan PC

    Abstract: The paper presents an analysis of the various problems involved in acheiving very high reliablility from complex computing systems, and discusses the relationship between system structuring techniques and techniques of fault tolerance. Topics covered include (i) differeing types of reliability requirement, (ii) forms of protective redundancy in hardware and software systems, (iii) methods of structuring the activity of a system, using atomic actions, so as to limit information flow, (iv) error detection techniques, (v) strategies for locating and dealing with faults, and for assessing the damage they have caused, and (vi) forward and backward error recovery techniques, based on the concepts of recovery line, commitment, exception and compensation. A set of appendices provide summary descriptions and analyses of a number of computing systems that have been specifically designed with the aim of acheiving very high reliablility.

      • Date: May 1977
      • Series Title: Computing Laboratory Technical Report Series
      • Pages: [118]
      • Institution: Computing Laboratory, The University of Newcastle upon Tyne
      • Publication type: Report
      • Bibliographic status: Published

        Professor Brian Randell
        Emeritus Professor, and Senior Research Investigator