Main ways to provide fault-free operation of multi-processor computation systems

Basic methods to provide the fault-free operation of computation systems, are considered. Brief reliability analysis of systems with the hardware redundancy (duplication), is presented. Technique of distributed diagnostics in homogeneous multiprocessor systems, and the distributed voting algorithms, are proposed.