Özet:

Abstract Cloud computing has many challenges, such as server failures, loss of confidentiality, improper workloads still limit the performance of cloud systems in real-world scenarios. Due to this, numerous research works are being carried out to improve the limitation of existing systems. Among them, load balancing seems to be a major issue that degrades the performance of the cloud industry, so optimal load balancing with optimal task scheduling is required. With the aim of attaining optimal load balancing by efficacious task deployment, in this manuscript Hybrid Elephant Herding Optimization and Flamingo Search Algorithm is proposed for effectual load balancing in cloud environment (LBS-CE-Hyb-EHO-FSA). The aim of proposed LBS-CE-Hyb-EHO-FSA is to enhance the population initialization and search space exploitation for activating the predominant load balance among the virtual machines (VMs) in the clouds. It includes the weighted task scheduling procedure depending on the optimization issue formulated utilizing the parameters of makespan, energy consumption and data center cost. Here, LBS-CE-Hyb-EHO-FSA is proposed for exploiting the merits of Elephant Herding Optimization (EHO) algorithm and Flamingo Search Algorithm (FSA) in order to achieve superior results in all dimensions of cloud computing. In this, LBS-CE-Hyb-EHO-FSA achieves the allocation of Virtual Machines (VMs) to incoming tasks of cloud, when the number of currently processing tasks of a specific VM is reduced than cumulative number of tasks presently processing by other VMs in the cloud.  It also attains potential load balancing process, then difference between the processing time of all individual virtual machine and the mean response time (MRT) incurred by the complete virtual machine. Finally, the simulation experiment of proposed LBS-CE-Hyb-EHO-FSA is conducted using Cloudsim platform. Here the proposed method provides 23.35%, 15.06%, 21.77%, 27.82%, 14.31%, 19.23% lower Mean Execution Time and 38.22%, 40.21%, 19.30%, 25.46%, 19.25%, 21.14% lower mean response time comparing to the existing models.

Anahtar Kelimeler: