Reliable processor coordination for distributed resource management in hypercubes

Under distributed resource management, processors in a multicomputer system must coordinate with each other through interprocessor communication to exchange state information, synchronize management activities, allocate and deallocate computing resources, and detect component faults. In this paper, we propose a reliable processor coordination scheme based on a communication mechanism, called recursively synchronized broadcast (RSB), to support these critical functions in a failure-prone distributed hypercube system