WITHDRAWN: Link delay based efficient search method for clustered peer to peer networks

Abstract

Due to self-organizing and autonomy nature of the Peer to Peer networks they are considered to be an important model for a wide range of application. The main challenge is to organize the peer nodes and perform search for a required data among the peer nodes. The clustering approach helps to efficiently organize and manage the network infrastructure. This paper proposed an approach to perform searching efficiently in a clustered network. The clustering can be implemented in a distributed fashion depending on the neighbouring nodes. Hierarchical clustering approach is adopted and each cluster is generated based on the link delay which reduces the transmission time between the cluster head or super node and the peer nodes in the network. In the clustered topology the congestion in the network especially in Cluster Controller (CC) is controlled by monitoring the state of the nodes including its wait queue length and processing capacity. An efficient searching method is proposed and compared with the flooding approach followed in the Gnutella network. The proposed approaches are most suitable for a file sharing network in which the bandwidth consumption while searching for a required data is reduced and hence the overall download time and the performance of the network is increased.