A Two Dimensional Clustering Search Protocol for Peer-to-Peer Networks

Peer-to-peer networks have evolved rapidly due to the flexibility and reliability in sharing data objects among participating users. However, without a centralized service provider, locating data objects is not a trivial task for P2P applications. In this paper, we propose a semi-structured protocol that explores the benefits both of unstructured small world models and consistent hashing techniques used in structured methods. The proposed methodology uses clustering to form a 2-dimensional overlay network topology, to localize search and to speed up lookup. As demonstrated by extensive simulation experiments, the proposed protocol is able to deliver a consistently better search performance than many other techniques presented in recent research literature, particularly meeting the needs of large scale P2P networks.