Task reallocation for fault tolerance in multiprocessor systems

A simple, yet effective, method to improve multicomputer/multiprocessor system reliability via redundant allocation of tasks to computers (processors) is described. Given any known (nonredundant) scheduling strategy, tasks are allocated to processors statically and redundantly using the k-circular shifting algorithm, so that if some processors fail during the execution all tasks can be completed on the remaining processors (but in a longer time). Due to static preallocation of tasks this method is simpler and thus more practical than well-known dynamic reconfiguration and roll-back recovery techniques in multiprocessor systems. The authors discuss in detail the redundant allocation of independent tasks to identical processors, subject to real-time constraints on total execution time, and derive analytic reliability estimates for this case. The longest processing time (LPT) scheduling is given as an example of nonredundant deterministic scheduling for independent tasks. The authors discuss processor utilization for redundant task allocation, and compare it with the standby redundancy technique. This comparison shows that the proposed method achieves much higher processor utilization than standby redundancy.<
>