An efficient load balancing method for ad hoc networks

• Published in: International journal of communication systems 2018-03, Vol.31 (5), p.n/a
• Main Authors: Singh, Jaspreet, Rai, C. S.
• Format: Article
• Language: English
• Subjects: Batteries; Computer networks; Computer simulation; Confidence intervals; efficient routing; Factorial design; Load balancing; load distribution; Load distribution (forces); MANETs; Mobile ad hoc networks; multipath routing; scalable routing; Stress concentration; Wireless networks
• ISSN: 1074-5351; 1099-1131
• DOI: 10.1002/dac.3503

Summary Routing is the most basic and essential operation of any ad hoc network. A mobile ad hoc network presents many challenges, because of the severe resource limitations such as dynamic and varying topology, lack of centralized control, insecure medium, and limited battery power, among others. Therefore, optimization and conservation is the key to success of any ad hoc network operation. In this paper, we propose and define 2 new metrics for ad hoc networks: bandwidth utilization ratio and load index. These metrics can be used as an indicator to measure and monitor the network usability and to improve its efficiency by efficient load distribution. They can be used to predict the additional load that can be accommodated in the network, without causing any congestion or overflows. We also propose a new load balancing routing scheme for ad hoc networks, called efficient load balancing method. This method tries to offset the load on different paths using load index as a metric. Load index is defined as a measure of a node's degree of involvement in the message routing process, which is indicative of its load. To make this algorithm efficient, we limit our routes to a few efficient ones only. This number of alternate routes used, out of the pool of all available routes, is defined as degree of distribution. Simulation results adequately prove the efficiency of proposed method, vis‐à‐vis 2 other load balancing approaches, and these are verified statistically at 99% confidence interval. A p × q factorial design is used to verify that simulation results are the actual measurements and not due to some unknown errors. In this paper, Two new metrics, bandwidth utilization ratio (BUR) and load index (LI), are proposed, which can be used to predict the additional load that can be accommodated in the network, without causing any congestion or overflows. A load balancing routing scheme for ad hoc networks called ELBM is also proposed, on the basis of LI metric. Simulation results are verified statistically at 99% confidence interval using a p × q factorial design, and they prove the efficacy of ELBM.