Resource Optimization in Cloud Data Centers Using Particle Swarm Optimization

Introduction

Cloud computing is a paradigm where the IT requirements are fulfilled on a subscription-based model, cloud computing enables users to use a portion of the computing resources, storage, RAM from a datacentre that will host all the above-mentioned resources. Virtualization is a technique where a fraction of cloud datacentre resources is reserved for a given user for a limited period. Cloud computing deals with several aspects such as storing and retrieving the data from web applications to more complex scientific modeling problems where cloud customers no need to worry much about the cost associated with it.Number of dynamic scheduling techniques has been reassessed based on meta-heuristics and deterministic to map out with the challenges of resource provisioning (Malaisamy and Murali, 2020). To reduce the time complexity and to increase the efficiency heuristic and meta-heuristic algorithms are used.

The number of people using cloud resources is increasing at an exponential rate this necessitates efficient algorithms for resource sharing and allocation, many researchers worked in this area to bring out optimization in Cloud resource sharing and allocation.

Figure 1.

Virtual Machine Architecture

Figure 1. Shows the basic architecture of Virtualization over the cloud datacenters. IaaS is the platform where it provides infrastructure as a service to the end-users based on their need. Cloud virtualization allows user-customized applications. Cloud Computing serves both software and hardware applications based on user demands.

Our objective is to study all the existing algorithms and implement a new algorithm that will do the multidimensional optimization, less time consuming to reduce the number of running physical systems thereby increasing the power efficiency of the whole data center. Particle Swarm Optimization (PSO) (Kennedy & R Eberhart, 1995) is a population-based optimization method based on Swarm Intelligence (SI) technique. The basic idea is to find the particle's potential position according to its own experience and that of neighbors. PSO is one of the powerful optimization technique where only a few parameters to adjust when compared to other heuristic algorithms. PSO has been applied to a wide range of applications where finding an optimal solution is abundant. It has a considerable amount of interest from nature-inspired community computing that has been seen to many offers which influence solving the optimization problems in multidimensional search spaces. The common goal of the swarm is to find the source of food (Loganathan M.K & Gandhi 2016), individual particle position is represented as P_best and one of the particles in the swarm reaches best possible position g_best,is the global search position. The PSO algorithm searches in parallel as well as in a group of individuals. Individuals or particles over a problem space, approach the optimal value through their current velocity, with their previous experience, and with the experience of their neighbors. The total number of particles and the swarm size always affect the performance of the algorithm (Y. Shen et al 2018). The author proposed modified Particle swarm optimization (C. Wen and W. Jiang 2019) using the crossover to maximize the resource utilization by introducing two major models for load balancing as well as to know the resource wastage. CPSO model presents the scenario by adjusting the crossover parameters in each iteration.