A Fault-tolerant Scheduling Algorithm Based on Grouping for Real-time Multiprocessor

In order to improve the fault-tolerant scheduling efficiency of aperiotic tasks in a real-time multiprocessor system, a fault-tolerant scheduling algorithm based on grouping for real-time multiprocessor is proposed. This algorithm determines the primary-backup scheduling forms of aperiotic tasks based on processor grouping technology. The algorithm uses Improved Spring algorithm to finish the scheduling of task copies in primary-backup concur and overlap form, and it uses Application Parameter overloading strategy to optimize the scheduling of task copies in primary-backup exclusive form, so that the scheduling efficiencies of aperiotic tasks are improved considerably. It is shown from the simulation that this algorithm has important effect on improving the fault-tolerant scheduling performance of aperiotic tasks.