Performance evaluation of P2P systems with hierarchical query routing

Abstract

P2P information retrieval systems have been changing from filename-based queries to contents-oriented queries. However, query messages occur frequently and consume the network bandwidth of communication systems. Instead of query broadcasting retrieval systems, we extended JXTA-like query routing protocols in order to handle characteristics of resources and queries, so that we tried to deal with topic-oriented query routing mechanisms. We also introduced hierarchical routing tables in order to reduce the storage and management costs. In our previous studies, we realized that the number of routing table hierarchy affect the trade-off between the cost of query forwarding and the amount of storage for routing information. In this paper, we consider analytical model in order to evaluate the characteristics of hierarchical peer clusters. We try to apply Palm-probability to calculate distribution of information resources. Additionally, we utilize Poisson-Voronoi-clustering to evaluate our hierarchical model.