Elephants journey towards successful resource discovery in unstructured P2P networks

Abstract

This paper presents a resource discovery scheme for decentralised unstructured P2P file sharing applications. The scheme utilises the principles underlying elephants migration from dry place to green fields for food. Hence, the proposed approach is a bio-inspired swarm intelligence based search technique. The aim of the technique is to route a query from a node to suitable peers in the network to find the required object. For this, the scheme divides the resource available areas in the network as wet areas and dry areas. The wet areas are resource fertile areas and dry areas have very limited resources. When a node is located in a dry area, its queries are propagated to wet areas for increasing the performance of resource discovery process. At the same time, the powerful nodes try to make the dry area into a resource lush one so that nodes that have moved to wet area return home for resource discovery. Simulation results show that the proposed scheme significantly increases query success rate and reduces the network traffic as the resources are effectively distributed to well-performing nodes. © 2009 IEEE.