Improving Resource Discovery and Query Routing in Peer-to-Peer Data Sharing Systems Using Gossip Style and ACO Algorithm

摘要

With the far-reaching significance of the Internet and the drastic advances in computer technology, more and more news and data are available on the Web. Peer-to-Peer (P2P) systems have become a popular way of sharing these data, and have drawn much attention of both academia and industry. The key and one of the most challenging design aspects in data sharing in P2P systems is how to find flexible, scalable and efficient mechanisms for searching and retrieving data. In the proposed approach, we use mobile agents to take advantage of a distributed system. To optimize the migration strategy of the mobile agents, we first resort to the biological behavior of ant colonies; the agents use trails leaved by other agents on peers they have visited. Then, to enhance the quality of the search results, reduce the randomness of peers and preserve their autonomy, we also introduce a social aspect, i.e., the friends list is used to gather peers having similar center of interest. To discover friends, the peers rely on the gossiping algorithm. We find that our contribution has three originalities distinguishing it from other approaches. The first one takes into account the two principal issues in data sharing in P2Pdatabase systems. The second one has the advantage to be totally independent of the centralized management. Finally, the third one is inherent to the resource discovery mechanism, which includes a social aspect using the friendship links.