02-09-2017, 11:06 AM
P2P (peer-to-peer) networks lay application-layer overlays at the top of the Internet to facilitate efficient resource sharing. They can be roughly classified as structured or unstructured networks. Without strict restrictions on network topology, unstructured P2P networks can be constructed very efficiently and are therefore considered suitable for the Internet environment. However, the random search strategies adopted by these networks usually work poorly with a large network size. In this work, we seek to improve search performance in unstructured P2P networks by exploiting patterns of common interest captured by users within a probability-theory framework called the User Interest Model (UIM). A search protocol and a routing table update protocol are also proposed to streamline the search process through the self-organization of the P2P network in a small world. Both theoretical and experimental analyzes are performed and demonstrate the effectiveness and efficiency of our approach.