Article Preview
Top1. Introduction
The progression in technology has led to the emergence and evolvement of cloud computing from the basic computing paradigm of distributed, parallel, and grid computing. It has revolutionized the procedure for managing the data or information and the resources which have impacted human society socially and economically. The services are offered to the users in the resources form (e.g., storage, CPU, servers, network, and application). These resources are organized at one central point, the data center, from where accessing these resources comes at ease. The overall management of data centers is the responsibility borne by the service providers. The service providers serve the users’ interests with three basic services via the internet IaaS for hardware resources, PaaS for runtime environment, and SaaS for software resources. These former services are being accomplished with virtualization. Virtualization facilitates the creation and configuration of VMs possessing variant operating systems with varying resources (memory, CPU, and storage) that are then nailed on the host machine to furnish the services desired by users. It also expedites sharing of multiple VMs deployed on one distinct server and sharing the hardware resources. To accomplish the request or interest demanded by users, VMs are created and configured dynamically with variable configuration and resource demands. The key objective in this reference is to allocate resources in ways that they are effectively utilized, decreasing resource waste and resulting in low operational costs. Adopting an effective VMA algorithm is one way to accomplish this. VMA allows VMs to be placed on the servers such that computing resources must be efficiently utilized while also reducing the volume of active servers. With the decisive allocation techniques, there come challenges, including a reduction in QoS, reduction of performance with the curtailed energy usage, and then satisfying the user's QoS within the promised SLA. An effective VMs allocation narrows down the count of active and balanced multidimensional resource usage by servers, ultimately reducing the energy expenditure by the cloud. Figure 1 demonstrates the allocation of VMs in which all the VMs are allotted randomly to the server, while after optimization in figure 2, the VMs are consolidated in the first three servers, and the rest of the unutilized servers are turned off, thus saving the resources and energy. Identifying optimal VMs allocation belongs to the NP-complete problem (Lo, V. M. 1983), and achieving the optimum resolution of this is typically computationally infeasible, when the cloud involves multiple hosts and users (Widmer, T., Premm, M., & Karaenke, P. 2013). The issue can be addressed to minimize resource wastage and the number of active servers in the cloud data center. Various methods have been applied to resolve this issue in the literature. A theorem was given (Wolpert, D. H., & Macready, W. G. 1997), which stated that not a single meta-heuristic algorithm is available which can efficiently resolve all the optimization problems. In this regard, the WOA (Mirjalili & Lewis, 2016) has been successfully utilized for various optimization problems (Prakash, D.B. & Lakshminarayana, C.,2017; Sun, Wang, Chen, & Liu, 2018; Chen, H., Xu, Y., Wang, M., & Zhao, X., 2019; Too, J., Mafarja, M., & Mirjalili, S., 2021).