Design and evaluation of a fault-tolerant multiprocessor using hardware recovery blocks
Attention is given to the design and evaluation of a fault-tolerant multiprocessor whose rollback recovery mechanism is based on hardware recovery blocks constructed by consecutive state-save operations and several state-save units in every processor and memory module. Upon detection of failure, the multiprocessor reconfigures itself to replace the faulty module and the process originally assigned to the faulty module retreats to one of the previously saved states, in order to resume fault-free execution. A mathematical model is proposed for the calculation of multistep rollback recovery coverage and the risk of restart when all available saved states are exhausted. The mean and the variance of task execution time with the occurrence of rollbacks and/or restarts are evaluated.