A design tool for fault tolerant systems

Complex systems may have to meet severe availability objectives related to the importance of the service being provided; such systems must be fault tolerant. Designers of fault-tolerant systems try to implement diagnostics to detect as many faults as possible because, in complex systems, uncovered faults lead to latent highly undesired situations. Unfortunately, diagnostics themselves may fail. Starting from the basics of FMECA, a design methodology and a tool have been developed. It is called DIANA (DIagnostic ANAlysis). The basic idea of DIANA is to perform coverage analysis during hardware and firmware design together with reliability engineering analysis. To this purpose, DIANA has been integrated into the computer aided design (CAD) tools in the same way that logic simulation timing analysis and analog transmission simulation are performed. Two main results have been obtained by the DIANA project: the first is to give the designers a tool that helps them to think in such a way as to prevent uncovered fault situations; the second is to calculate the effects of faults on diagnostics in order to provide transition rates to system availability models when real, rather than ideal, cases are taken into account