Probabilistic Search in P2P Networks with High Node Degree Variation

Zhang Haoxiang,Zhang Lin,Shan Xiuming,Li Victor O. K.
DOI: https://doi.org/10.1109/ICC.2007.286
2007-01-01
Abstract:A novel adaptive resource-based probabilistic search algorithm (ARPS) for P2P networks is proposed in this paper. ARPS introduces weighted probabilistic forwarding for query messages according to the node degree distribution and the popularity of the resource being searched. A mechanism is introduced to estimate the popularity and adjust the forwarding probability accordingly such that a tradeoff between search performance and cost can be made. Using computer simulations, we compare the performance of ARPS with several other search algorithms. It is shown that ARPS performs well under various P2P scenarios. ARPS guarantees a success rate above a certain level under all circumstances, and enjoys high and popularity-invariant search success rate.
What problem does this paper attempt to address?