The design of real-time systems for safety-critical applications depends heavily on the normal operation of system, in critical conditions. In these applications, among of real-time tasks, a critical task must be immediately scheduled at its arrival time immediately; otherwise, it leads to a system failure and disasters in safety-critical applications. A major problem in real-time systems included critical tasks, is unpredictable arrival of these tasks. To resolve the problem, a kind of scheduler, called “super scheduler”, is employed. The problem can be more complex, in a hierarchical real-time system. A hierarchical real-time system consists of several real-time sub-systems, called “components” .Hence, using super scheduler for each component of the system, needs to special considerations. On the arrival of a critical task, the super scheduler preempts the currently running tasks and alters the priority of all existence tasks. When the critical task is completed, the preempted tasks are executed in their new priority order. This guarantees the completion of the critical and almost all other non-critical tasks before their deadlines, and therefore the stability of the component. To guarantee the stability of a hierarchical real-time system, all its components should be stable. This paper presents a model to guarantee the stability of a hierarchical real-time system included a critical task in each component. Moreover, a fault tolerance method has been applied for all components. Evaluation results show that the proposed technique improves the stability of a hierarchical real-time system included critical tasks by decreasing the number of tasks which miss their deadline.

Keywords