Unstructured Peer-to-Peer (P2P) overlays are the most widely used topologies in P2P systems because of their simplicity and very limited control overhead. A P2P overlay specifies the logical connections among peers in a network. Such logical links define the order in which peers are queried in search for a specific resource. The most popular query routing algorithms are based on flooding, thus they do not scale well as each query generates a large amount of traffic. In this paper, we use heuristics to improve overlay search in an unstructured P2P file sharing system. The proposed heuristics effectively decide replica locations for popular resources based on the availability of computing and storage at a given peer, its neighborhood information, and the used routing strategy. Simulations performed over two different types of unstructured P2P network topologies (i.e., power law and random graphs) show significant improvements over plain flooding in terms of reduced network traffic and search time. © 2013 Springer-Verlag.
|Title of host publication||Grid and Pervasive Computing|
|Subtitle of host publication||8th International Conference, GPC 2013 and Colocated Workshops|
|Publication status||Published - 2013|
|MoE publication type||A4 Article in a conference publication|
|Series||Lecture Notes in Computer Science|
- overlay networks