Test of fault tolerant distributed systems by fault injection

Fault tolerance protocols can be checked for design faults by injecting operation fault cases. We present a scheme for generating these operation fault cases on the basis of a reachability analysis. Thereby, we aim at a low number of cases by choosing an appropriate model representing the essential properties of fault tolerance techniques in detail and abstracts from others. Timing and selection of redundant data are expressed accurately, whereas application functions, iterations and recursions are coarsely covered. The semantics of digital signatures is fully included.