Secure Information Sharing in Mixed – Criticality Systems

In this paper we discuss the application of integrity models in a mixed–criticality system to enable the secure sharing of information. The sharing of resources and information enables cost savings. It is the main driving factor in integrated architectures which are used to implement mixed–criticality real-time systems. The major challenge of these systems is simple: low criticality applications must be prevented from interfering with high criticality ones which execute in the same system. An example for such an integrated architecture is the the ACROSS MPSoC architecture which facilitates the implementation of hard real–time systems. We present an integrity model for the secure exchange of information between different levels of criticality within ACROSS. Our approach is based on Totel’s integrity model which proposes to upgrade information from low to high by rigorously validating this information. We use different anomaly detection algorithms for these validation objects. Furthermore, we developed an automotive case study for the hardware–in–the–loop simulation of ABS and odometer subsystems in order to provide a proof– of–concept implementation and to evaluate these algorithms within an automotive case study. We were able to show that the encapsulation mechanisms of the ACROSS architecture support the implementation of the proposed integrity model. Moreover, all of the selected anomaly detection algorithms validate the information flow correctly with respect to the defined integrity model. For some of these algorithms, we are able to propose tuning parameters. Summarizing, we are able to show that the secure sharing of information is feasible in a mixed– criticality system. Integrating several subsystems in a single Multi–Processor System–on–a–Chip (MPSoC) not only reduces the number of required hardware units but also enables new ways to implement services.