Peer-to-peer networks

Summary form only given. The term peer-to-peer (P2P) system encompasses a broad set of distributed applications which allow sharing of computer resources by direct exchange between systems. The goal of a P2P system is to aggregate resources available at the edge of Internet and to share it cooperatively among users. Specially, the file sharing P2P systems have become popular as a new paradigm for information exchange among large number of users in the Internet. They are more robust, scalable, fault tolerant and offer better availability of resources than the traditional client server model. Depending on the presence of a central server, P2P systems can be classified as centralized or decentralized. In decentralized architecture, both resource discovery and resource download are distributed. Decentralized P2P architectures may further be classified as structured or unstructured networks. In structured networks, there are certain restrictions on the placement of contents and the network topologies. In unstructured P2P networks, however, placement of contents is unrelated to the topologies of the network. Unstructured P2P networks perform better than their structured counterparts in dynamic environments. However, they need efficient search mechanisms and they also suffer from numerous problems such as: fake content distribution, free riding (peers who do not share, but consume resources), whitewashing (peers who leave and rejoin the system in order to avoid penalties) and lack of scalability in searching. Open and anonymous nature of P2P applications lead to complete lack of accountability of the contents that a peer may put in the network. The malicious peers often use these networks to do content poisoning and to distribute harmful programs such as Trojan Horses and viruses. In this talk, we will discuss various architectures of the current p2p systems and the major challenges in deployment and operations of such networks such as scalability, quality of services of applications, security and privacy of users and applications etc. After a generic introduction of these issues, we will present a search algorithm for p2p networks which is able to combat the problem of inauthentic file downloads. The searching algorithm is also able to improve search scalability while protecting the privacy of the users. The scheme is based on construction of an overlay of trusted peers where neighbours are selected based on their trust ratings and content similarities. The search efficiency is improved by taking advantage of implicit semantic community structures formed as a result of topology adaptation. We will also present detailed experimental results to demonstrate the performance efficiency of the searching scheme for p2p networks.