Fault Tolerance in High-Performance Computing and Grids