Using neural search approach for resource discovery in P2P networks

Resource discovery is one of the main concerns of Peer-to-Peer (P2P) networks. This is due to the distributed nature of P2P networks and that there is no centralized indexing to look for the information of available resources and their places. Previously, to search for one requested resource, algorithms were invented that exploited the local knowledge about the network. This approach is inaccurate and cannot be scaled well in presence of large P2P networks. On the other hand, replicating each node’s local knowledge and using it to build up a comprehensive knowledge base is very expensive. In this paper, we describe a method that benefits from the flexibility of Neural Networks and not only takes into account most of the measures previously used to optimize the searching process, but also performs efficiently and is designed simple. Result of running the algorithm on a sample network is evaluated and its performance is compared with another common method to prove its efficiency