FireBat: a hybrid approach for load balancing in cloud computing

In cloud computing, resources are allocated over the Internet to deliver services to end users. However, owing to the exponential increase in the demand for cloud resources, data centres receive dynamic workloads with varying computational capacities, which leads to the unbalancing of servers. This imbalance results in the degradation of the quality of service and resource utilization, which may impact the processing time of the VMs or may cause a system crash. Therefore, an effective load-balancing approach FireBat is proposed in this work that integrates the exploration capability of Firefly and exploitation capability of BAT is implemented to address the issues in load balancing. This model considers task scheduling as a cost minimization problem, wherein task to VM mapping is evaluated based on several factors such as VM processing capability, task length and current load conditions. The proposed work not only improves the makespan and degree of imbalance but also focuses on maximizing throughput. The simulation was conducted using CloudSim to demonstrate that FireBat is significantly better than BAT, Firefly, and FHLBA. The results indicate that the proposed technique outperforms these algorithms, in the best case, the makespan is decreased by 23% while DOI by 12.66%, and throughput is increased by 35%.