A Novel Efficient Search Algorithm in Unstructured P2P Networks

There are two major problems of unstructured P2P systems, one is their heavy network traffic; the other is the problem of unpredictable delay. The popular search techniques in unstructured P2P networks such as flooding, dynamic query don´t deal with these problems efficiently for they are lack of heuristic. In this paper, we proposed a new search algorithm, called immune based search (IBS) algorithm, to solve the problem. Firstly, an immune query model is proposed, which is inspired from the concept of similarity-governed clone proliferation and mutation. Secondly, based on the immune query model, IBS algorithm is proposed. Lastly, some assistant strategies, that is, shortcuts creation and peer traveling are incorporated into IBS to develop ´immune memory´ for improving search performance, which can make IBS not be ´blind´ but heuristic. The simulation results indicate that IBS show good performance.