Timed Hazard Analysis of Self-healing Systems

Self-healing can be used to reduce hazards in embedded real-time systems which are applied in safety-critical environments. These systems may react to failures by a structural reconfiguration of the architecture during runtime. This means the exchange of components or the modification of the components’ connections, in order to avoid that a failure results in a hazard. This reaction is subject to hard real-time constraints because reacting too late does not yield the intended effects. Consequently, it is necessary to analyze the propagation of failures in the architectural configuration over time with respect to the structural reconfiguration. However, current approaches do not take into account the timing properties of the failure propagation and the structural reconfiguration. In this paper, we present a hazard analysis approach which specifically considers these timing properties. We illustrate our approach by an example case study from the RailCab project. Further, we demonstrate the scalability of the approach by experiments.