Load Balancing in Peer-to-Peer System Using Fuzzy C-Means Clustering

Load Balancing in Peer-to-Peer System Using Fuzzy C-Means Clustering

Rupali Bhardwaj (Ajay Kumar Garg Engineering College, Ghaziabad, Uttar Pradesh, India), V. S. Dixit (Department of Computer Science, Delhi University, Delhi, India) and Anil Upadhyay (Bipin Chnadra Tripathi Kumaon Engineering College, Ranikhet, Uttarakhand, India)
Copyright: © 2013 |Pages: 12
DOI: 10.4018/ijfsa.2013010105
OnDemand PDF Download:
$37.50

Abstract

Objective of load balancing algorithm is to keep all nodes normally loaded through migration of modules from heavy weighted nodes to light weighted nodes. In addition, load balancing must involve low communication overhead and respond quickly to load imbalance in the system. In previous load balancing algorithms, classification of nodes is done by using threshold value; which is fixed and predefined. In this paper, the authors proposed load balancing algorithm using fuzzy c-means clustering which changed the status of nodes dynamically according to the state of system. The proposed algorithm is compared with other existing algorithms and is found to be fast and efficient in reducing load imbalance in peer-to-peer system.
Article Preview

Literature Review

Timothy (1995) proposed that fuzzy logic is very useful to solve uncertain problems. Karimi et al. (2009) presents a new dynamic load balancing algorithm with fuzzy logic and it show better response time than Round Robin algorithm by 30.84%. Ally et al. (2002) presents load balancing in distributed computing systems with help of fuzzy functions and solve uncertainty problem in state information and in task selection with the help of fuzzy functions of variable membership values. Clustering motive is to distribute cases (people, objects, events etc.) into groups, so that the degree of association to be strong between objects of the same cluster and weak between objects of different clusters (Olson et al., 1996). Bezdek et al. (1980, 1981, 1987) present some fuzzy techniques for clustering numerical data. Horng et al. (2002) proposed convergence theorem for a fuzzy ISODATA clustering algorithm. Ruspini (1970) presents numerical methods for fuzzy clustering. Huang et al. (2003) improves performance of computer network by balancing load among computers with the help of receiver-initiated fuzzy logic control method. Ahn et al. (2007) presents load balancing in distributed systems using an intelligent fuzzy grouping approach. In the round robin algorithm (Pradeep et al., 1996), processes are equally divided among all processors. Each new process is assigned to a new processor in a round robin fashion. Whenever number of processes larger than number of processors, round robin algorithm works well. No need of inter-process communication in case of round robin algorithm. In the bidding algorithm (Xu et al., 2003), when a node is heavy loaded, it multicasts a request for bids to the other nodes in the system. After collection of all bids by heavy loaded node, the best bid is chosen as light weighted node. If none of the node is found in the group for load transfer, the bidding procedure starts over again.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 6: 4 Issues (2017)
Volume 5: 4 Issues (2016)
Volume 4: 4 Issues (2015)
Volume 3: 4 Issues (2013)
Volume 2: 4 Issues (2012)
Volume 1: 4 Issues (2011)
View Complete Journal Contents Listing