AntSearch: An Ant Search Algorithm in Unstructured Peer-to-Peer Networks
- Kai-Hsiang YANG†a
This paper presents a new replica optimal location strategy based on ant colony optimization algorithm. Through the improvement of ant algorithm, it can effectively choose the optimal replica node from a great number of nodes in peer-to-peer (P2P) system, which will help to realize the global optimization. The strategy fully considers that P2P nodes are heterogeneous, so the replica of a high degree of popularity will be placed in the high-performance nodes. It increases the high availability of the popular files. At the same time, owing to the increase of the high popular replica, it reduces the number of inquiring nodes when searching a file and decreasing the network traffic. This strategy also takes full consideration of the deferent factors that affect replica location, like path load, delay and so on. The simulation by using P2P tool PeerSim show that this strategy can effectively reduce request response time. Therefore, it improves performance of overall system.