Diagnosis of computer systems by stochastic Petri nets

A fault-tolerant computer system is represented as a stochastic Petri net, and discrete-event simulation methods are introduced to study its behavior. The power of stochastic petri nets with general firing times for formal representation and quantitative diagnosis, from the customer point of view, is exhibited. The results show that simulation methods for stochastic Petri nets can be used advantageously to diagnose quantitatively whether or not a fault-tolerant computer system meets the requirements of the environment