MoDiVHA: A Hierarchical Strategy for Distributed Test Assignment

Distributed diagnosis allows a set of fault-free nodes to monitor the state of all nodes of a given system. Diagnosis is based on the results of tests, which are assigned among system nodes. Hierarchical testing assignments lead to efficient and scalable diagnosis algorithms, both in terms of the number of tests executed, and the latency. In this work we present a novel hierarchical testing strategy, called MoDiVHA. By obtaining as much diagnostic information as possible from each test, and avoiding tests on nodes about which information has been already obtained, MoDiVHA executes less tests in comparison with previously proposed assignments. Experimental results are presented from three series of simulations conducted to compute the number of tests and diagnosis latency for various system sizes and fault situations.