A Low Latency Resource Location Algorithm for Unstructured P2P Networks

In this paper we focus on building low latency resource location service over unstructured Peer-to-Peer (P2P) networks. In state-of-the-art search approaches in unstructured P2P systems ,they often evaluate efficiency by overlay hops, while two nodes logically adjacent maybe physically far away and from user´s QoS, clients always want to acquire as many as objects in a short time, so we propose an informed search algorithm to locate resources. The proposed algorithm selects optimal paths to send/forward the query messages according to the nodes´ weight. Experiment on the simulation proves that our method could be adaptive for topology changes and link state. Compared with random-walk system, it can dramatically reduce search average latency, enhance hit rate and improve performance up to 30%.