A scalable and adaptive distributed service discovery mechanism in SOC environments

Abstract

Current researches on service discovery mainly pursue fast response and high recall, but little work focuses on scalability and adaptability of large-scale distributed service registries in SOC. This paper proposes a solution using an agent based distributed service discovery mechanism. Firstly an unstructured P2P based registry system is proposed in which each peer is an autonomous registry center and services are organized and managed according to domain ontology within these registry centers. Secondly, an ant-like multi-agent service discovery method is proposed. Search agents and guide agents cooperate to discover services. Search agents simulate the behaviors of ants to travel the network and discover services. Guide agents are responsible to manage a service routing table consisting of pheromone and hop count, instructing search agents' routing. Experimental results show that the suggested mechanism is scalable and adaptive in a large-scale dynamic SOC environment. © 2008 Springer Berlin Heidelberg.