Dynamic autonomous redundancy management strategy for balanced graceful degradation

This paper proposes a redundancy management method, ARM (Autonomous Redundancy Management). The method dynamically allocates redundant computer modules to tasks by an autonomous algorithm based on shared fault information, considering the importance of the tasks necessary to balance their reliability margin. It allocates more computer modules to a task which is endangered because of the number of fault occurrences, or which is granted as a very important task requiring very high reliability