Resource discovery in location-aware effective P2P network model

Abstract

Resource discovery is a key component of P2P network. The resource discovery algorithm complexities of classic models, such as Chord and Pastry, are always equal or even bigger than O(logN), which means while the network scale increasing, the time delay will also increase rapidly without a ceiling. In this paper, we present a new P2P network model for resource discovery. Firstly, we take location-aware into account to divide the whole network into several domains, so that the topology mismatch problem between the P2P logical topology and the physical topology can be solved. Furthermore, in order to control the algorithm complexity as a constant, we assign each domain a super peer to help normal peers to do some operations, such as resource publishing and searching. Comparing the performance of this model with Chord, the results show the new model can reduce the resource discovery complexity to a constant, and significantly shorten the latency of the searching process.