Stochastic automata networks for dependability modelling

Stochastic automata networks have been used in the past with success to model parallel and distributed systems. It is our contention that they can also be used effectively to solve dependability models. These models possess characteristics that the SAN approach can exploit. In particular, the SAN approach completely eliminates the problem most often associated with this type of modelling, namely, the need for large amounts of computer memory. The only remaining difficulty is to show that the computation time is not excessive and this we attempt to do by providing a number of modelling strategems and numerical techniques designed to ensure that the time needed to compute solutions remains modest