FAULT TOLERANT COMPUTING

This paper is devoted to an overview of software fault tolerance by means of design diversity, i.e. the production of two or more systems aimed at delivering the same service through separate designs and realizations. The first section is devoted to a unified presentation of the approaches for software-fault tolerance; in addition to the recovery blocks and N-version programming methods, a third type of approach is identified from the careful examination of current, real-life systems: N self-checking programming. In the second section, the three approaches to software fault tolerance are analyzed with respect to two viewpoints: dependability and cost. The third section is devoted to the definition and analysis of architectures aimed at tolerating both hardware and software faults