A dependable system based on adaptive monitoring and replication

A multi agent system (MAS) has recently gained public attention as a method to solve competition and cooperation in distributed systems. However, MAS's vulnerability due to the propagation of failures prevents it from being applied to a large-scale system. This paper proposes a method to improve the reliability and efficiency of distributed systems. Specifically, the paper deals with the issue of fault tolerance. Distributed systems are characterized by a large number of agents, who interact according to complex patterns. The effects of a localized failure may spread across the whole network, depending on the structure of the interdependences between agents. The method monitors messages between agents to detect undesirable behaviors such as failures. Collecting the information, the method generates global information of interdependence between agents and expresses it in a graph. This interdependence graph enables us to detect or predict undesirable behaviors. This paper also shows that the method can optimize performance of a MAS and improve adoptively its reliability under complicated and dynamic environment by applying the global information acquired from the interdependence graph to a replication system. The advantages of the proposed method are illustrated through simulation experiments based on a virtual auction market.