Binary Self-Adaptive Salp Swarm Optimization-Based Dynamic Load Balancing in Cloud Computing

Abstract

In the recent era of cloud computing, the huge demand for virtual resource provisioning requires mitigating the challenges of uniform load distribution as well as efficient resource utilization among the virtual machines in cloud datacenters. Salp swarm optimization is one of the simplest, yet efficient metaheuristic techniques to balance the load among the VMs. The proposed methodology has incorporated self-adaptive procedures to deal with the unpredictable population of the tasks being executed in cloud datacenters. Moreover, a sigmoid transfer function has been integrated to solve the discrete problem of tasks assigned to the appropriate VMs. Thus, the proposed algorithm binary self-adaptive salp swarm optimization has been simulated and compared with some of the recent metaheuristic approaches, like BSO, MPSO, and SSO for conflicting scheduling quality of service parameters. It has been observed from the result analysis that the proposed algorithm outperforms in terms of makespan, response time, and degree of load imbalance while maximizing the resource utilization.

A secure framework for effective workload resource management

“…The results have been evaluated in terms of resourceutilization. Parida et al [17], presented a binary self-adaptive salp-swarm-optimization for handling the dynamic load balancing in cloud computing. Results show better response time, makespan and has increase the resource-utilization.…”

An efficient load balancing in cloud computing using hybrid Harris hawks optimization and cuckoo search algorithm

“…Task scheduling involves giving task deadlines and completion times to cloudlets, and load balancing (LB) involves performing workload movement in the event of VMs violations, to keep the workload balanced in the cloud environment [4]. Task scheduling in the cloud is done by adhering to service level agreement (SLA) norms for customers and cloud developers, and it substantially aids in LB while performing tasks [5]. It doesn't take much time to cache these files because a dataset is regularly uploaded to the cloud environment.…”

“…The objective function was chosen after combining optimization of various functions as shown in Equation 5. (5) where indicates the generation's nest position, and L () Levy-flight random search pathways in Equation 6.…”