An Efficient and Scalable Search Mechanism in Unstructured Peer to Peer Network

Peer-to-peer (P2P) network systems gain a huge popularity due to their scalability and reliability in architectures and search facilities. Basically, most of the real world P2P network is unstructured. Due to their unstructured nature it is often impossible to pre-define the searching criteria. As a solution, flooding scheme is used in most cases. But one major limitation of flooding is its query overhead and unnecessary use of bandwidth. In this paper, we propose a novel mechanism to improve the search efficiency in unstructured P2P networks. The method is based on feedback biased walk. Instead of keeping only the feedback report a peer keeps measuring its rank among the neighbors, this algorithm maintains another rank index for a peer that indicates the reliability to make a successful query hit through that peer. Cumulative Feedback-biased Walk (CF-walk) removes the limitations of unstructured P2P networks based on flooding point and cache based searching, improves the hit rate in case of search resources and reduce the query delay. Extensive simulations show that it improves the workload and reduce query traffic. Keywords—P2P Networks, Unstructured P2P, Feedbackbiased walk, Scalable, Rank based.