Combining search and trust models in unstructured peer-to-peer networks
The Journal of Supercomputing
| Hi-index | 0.00 |
Peer-to-peer networks have gained a tremendous popularity in sharing huge volumes of data. Success of such networks highly depends on the performance of their search algorithm. We propose a semantic based search algorithm for unstructured peer-to-peer networks. In the proposed method we establish ontology based indexes for outgoing links of each node, which are utilized in routing the query through the network. In contrast to available approaches, we maintain limited size indexes and also consider number of documents accessible via each link and the distance between source and destination nodes, to improve the accuracy and efficiency of our algorithm. Our design is scalable and adaptable, to satisfy the requirements of dynamic peer-to-peer networks. We show that the proposed search method performs well whether the requested object is a popular or rare concept. Our technique is simple in design and implementation. We simulate our algorithm on an unstructured peer to peer network and show that while preserving the quality of search results, it significantly reduces network traffic in comparison to randomized routing scheme.