Mathematical Model for Improving Cloud Load Balancing Using Scheduling Algorithms

Cloud computing is an environment that provides processing and executing requests in the form of tasks on the Virtual Machines (VMs). The resource scheduling and load-balancing mechanisms of the cloud are important to keep an easy flow of task processing and execution at any instance of time, irrespective of the task size. Therefore, it becomes crucial to study the behavior of resource scheduling algorithms with respect to the load-balancing mechanism. The main objective of this research paper is to process and execute different-sized tasks in the WorkflowSim environment on the cloud VMs by using the resource scheduling algorithms Max–Min (MX–MN), Minimum Completion Time (MCT), and Min–Min (MN–MN) under different scenarios with the purpose to study the behavior of these resource scheduling algorithms with respect to the load-balancing mechanism. A mathematical model is also proposed to calculate the amount of load balanced by a specific VM. The mathematical model of linear regression is used to provide an empirical analysis and differentiate how the resource scheduling algorithms behaved under these various scenarios. From the experiment, results obtained, and empirical analysis, it can be said that the performance of the MX–MN is best, followed by the performances of MN–MN and MCT, respectively. The machine learning (ML) method of reinforcement learning (RL) is also proposed at the end to enhance the resource scheduling and load-balancing processes of the cloud.