Alpha multipliers breadth-first search technique for resource discovery in unstructured peer-to-peer networks

Abstract

Resource discovery in unstructured peer-to-peer (P2P) networks is important in the field of grid computing. Breadth-first search (BFS) is widely used for resource discovery in unstructured P2P networks. The technique is proven to return as many search results as possible. However, the network cost of the technique is high due to the flooding of query messages that can degenerate the performance of the whole network. The objective of this study is to optimise the BFS technique, so that it will produce good search results without flooding the network with unnecessary walkers. Several resource discovery techniques used in unstructured P2P networks are discussed and categorised. P2P simulators that are used for P2P network experiments were studied in accordance with their characteristics such as, scalability, extensibility and support status. Several network topology generators were also scrutinised and selected in order to find out the most real-life like network generation model for unstructured P2P experiments. Multiple combinations of five-tuple alpha multipliers have been experimented to find out the best set to make -BFS. In our test, the -BFS increases the query efficiency of the conventional BFS from 55.67% to 63.15%.