Abstract: Security-critical systems tend to be subjected to much more serious security threats, therefore it poses a great challenge to reduce the risk of security-critical applications. This article proposes a risk-driven task scheduling mechanism by incorporating risk assessment with real time scheduling. Based on the concept of synthetic utilization of aperiodic tasks, an online task scheduling algorithm is designed to schedule security-sensitive aperiodic tasks with minimal security risk. This algorithm is an approximate algorithm, which can yield security performance guaranteed near-optimal solution in low time complexity without violating the security-critical and time-critical constraints of tasks. Simulation results show that the proposed algorithm can significantly reduce the risk of security-critical real time applications and satisfy the bound of risk approximation ratio. Specifically, maximal security risk reduction of up to 22.2% is achieved, and minimal approximation ratio of security risk can decrease to 102.4% for the bound of 110%.

Key words: security-critical, risks, real time systems, aperiodic task, scheduling algorithms