Mixed-Critical Systems Design with Coarse-grained Multi-core Interference

International audienceThose autonomic concurrent systems which are timing-critical and compute intensive need special resource managers in order to ensure adaptation to unexpected situations in terms of compute resources. So-called mixed-criticality managers may be required that adapt system resource usage to critical run-time situations (e.g., overheating, overload, hardware errors) by giving the highly critical subset of system functions priority over low-critical ones in emergency situations. Another challenge comes from the fact that for modern platforms-multi-and many-cores-make the scheduling problem more complicated because of their inherent parallelism and because of "parasitic" interference between the cores due to shared hardware resources (buses, FPU's, DMA's, etc.). In our work-in-progress design flow we provide the so-called concurrency language for expressing, at high abstraction level, new emerging custom resource management policies that can handle these challenges. We compile the application into a representation in this language and combine the result with a resource manager into a joint software design used to deploy the given system on the target platform. In this context, we discuss our work in progress on a scheduler that aims to handle the interference in mixed-critical applications by controlling it at the task level