Reliability Assessment of Embedded Systems Using Stopwatch Petri Nets

In this paper, we propose a reliability approach in which feared events define reliability requirements and taking them into account allows to design systems which will be able to avoid the drift towards a feared state. The description of feared scenarios since the system design phase enables us to understand the reasons of the feared behavior in order to envisage the necessary reconfigurations and choose safe architectures. In order to face the increasing complexity of embedded systems and to represent the suspension and resumption of task execution we propose to extract directly feared scenarios from Stopwatch Petri net model avoiding the generation of the associated reachability graph and the eternal combinative explosion problem.