Leader based adaptive fault diagnosis algorithm for distributed systems

This paper proposes a new algorithm named Leader Based Adaptive Fault Diagnosis (L-AFD) algorithm for distributed systems. This algorithm detects all faulty nodes in the network where, the network is not fully connected. This algorithm works for the arbitrary network. The t-diagnosibility of a system under consideration is (n-1) where n is total number of nodes or computer systems in the network. This algorithm supports new node entry in the network. It allows re-entry of the repaired faulty nodes during the next diagnostic cycle. This algorithm can also work with more than one leader. This algorithm executes periodically on each node.