Functional approach to decentralized search engine for P2P-network communities

Abstract

In this paper we propose a simple powerful method to increase search performance in large p2p-network communities, especially, but not limited to Internet. Our approach is based on introducing common functional language methods "map" and "reduce" in distributed network environment. The method exploits different connectivity levels of p2p nodes, dividing them into two groups – master nodes, performing preliminary search processing and worker nodes, performing actual search. We show that approach could be effectively used in Grid network, built over common P2P-network utilizing benefits of routing in Grid environment. An experimental implementation of the method confirms the theoretical analysis of the method.