A Hybrid Fault-Tolerant Scheduling Algorithm of Periodic and Aperiodic Real-Time Tasks to Partially Reconfigurable FPGAs

FPGAs have been used widely in space related design engineers and the probability of fault occurring increases when they are subject to total ionization dose. In this paper, the problem of fault-tolerant is solved by task scheduling and a fault tolerant scheduling algorithm of hardware real-time tasks is proposed based on primary/backup copy. By scheduled backwards, the backup copy executes as late as possible before its deadline. So that the primary copy has no overlap or as small overlap as possible with the backup copy on the execution time. When the primary copy executes successfully, the backup copy is released. The experimental results demonstrate that the algorithm is effective and can make full use of FPGAs.