REAL-TIME SYSTEMS

Aprender e ́ a única coisa de que a mente nunca se cansa, nunca tem medo e nunca se arrepende. —LEONARDO DA VINCI Mechatronics systems are characterized by their ability of integrating components and systems to control real world entities. To do so, real-time systems have increasingly been used, since controlling real world objects requires considering both their logical and timing constraints. In this context, fault tolerance mechanisms play an important role since faults in such a system may imply in considerably losses. For example, an error in a flight control system may incur in loss of human life. Several fault-tolerant approaches for real-time systems have been derived. Most of them restrict the system and fault model in a particular way or are strictly linked to a specific recovery model or scheduling policy. Moreover, to the best of our knowledge, there is no systematic metric, which enables an effective comparison among the existing fault-tolerant approaches. The main goal of this work is to fill this gap by providing a fault resilience comparison metric for different fault-tolerant scheduling approaches. To do so, a simulation-based framework was derived to compute resilience values while simulating specific time windows. Such values are then used to infer system resilience based on statistical analysis techniques. Results have shown that the derived metric can be used to compare, for example, two scheduling policies for real-time systems from fault resilience viewpoint, which demonstrates that our approach is reasonably independent of the system model