Fault-Tolerant Real-Time Scheduling using Passive Replicas

In hard real-time multiprocessor systems, it is necessary to have a fault-tolerant task execution scheme to meet the tasks' deadlines even in the presence of a processor failure. In this paper, we propose a delayed scheduling algorithm using a passive replica method. This scheme has relatively a small overhead for backup processes. For the purpose of high schedulability, we allow the primary copy of one task to overlap with the backup copy of another task. And, if all processors are busy to execute a task, the schedule of an arriving task is delayed until at least one processor is available. A heuristic cost function is used to select the task to be scheduled next among the delayed tasks. Simulation results show that the proposed algorithm achieves higher schedulability compared to other conventional approaches and a single spare node scheme.