Distributed Diagnosis and the System User

Consider a computing system in which the diagnosis is carried out by the Kuhl and Reddy distributed diagnosis algorithm. When using this algorithm, each fault-free unit correctly determines the status of every unit in the system when the number of faulty units is not larger than the connectivity of the directed graph representing the system. However, this may be useless to an outside observer (or user) who needs to know the status of the system in order to repair the faulty units and/or obtain the results of computing tasks from the fault-free units.