Method and device for enhancing the reliability of a multiprocessor system by hybrid checkpointing

A device for enhancing the reliability of a system comprising a plurality of processors and a memory, the device comprising:

• means for grouping processes into a plurality of groups, said means grouping the processes by data dependency spheres, each data dependency sphere grouping processes that access areas of the memory, wherein:
• a union of the access areas is logically connected, and
• each of the access areas has a non-empty intersection with the union of all other areas among said access areas;
• means for saving individually for each group of processes the data stored in the memory, the data being used by at least one of the processes belonging to said group,
• to restore an error-free global state of the system following an error occurring in one of the processors executing one of the processes belonging to said group, without restoring the entire memory, and,
• means for confining accesses to the memory by data dependency sphere, the means for confining comprising:
• means for managing access rights to the memory areas for each process to prevent propagation of a logic error occurring in the processor executing said process;
• and
• means for managing indicators of memory areas belonging to each data dependency sphere.