Analysis of round-robin load-balancing algorithm with adaptive and predictive approaches

An enormous increase in the number of internet users which will tend to rise further, cluster-based web servers (CBWS) are experiencing a dramatic increase in web traffic. Round-robin load-balancing algorithm (RLBA), is one of the most widely used for distributing loads among the web servers due to its simplicity. However, in the case of non-uniform web traffic, RLBA load distribution is inefficient. In this paper, we have developed novel approaches to optimize the RLBA with the necessary procedures. We propose Adaptive RLBA (ARLBA) and Predictive RLBA (PRLBA). We validate the effectiveness of algorithms through simulation results. Server load correlation and load variance are used as the performance measuring parameters. ARLBA and PRLBA outperform RLBA in all the cases. ARLBA performs slightly better than Modified round robin (MRR) and vice versa in case of PRLBA.