Efficient fault-tolerant scheduling on multiprocessor systems via replication and deallocation. (1st January 2014)