In the cloud environment, the tasks are assigned to the virtual machines based on their requirements without considering their long-term and overall utilization. Also, in many cases, load balancing and scheduling process are computationally cost expensive and also affects the performance of the Virtual Machines (VMs). Task scheduling is a major issue in cloud computing. An efficient load balancing and scheduling algorithm should focus on satisfying all the Quality of Service (QoS) parameters such as task execution time, task waiting time, resource utilization, execution cost, etc. In order to improve the performance of task scheduling and also to meet the above mentioned QoS parameters, a VM scheduling algorithm that takes already running VM resource usage over time by analyzing past VM utilization in order to schedule the tasks among the VMs by optimizing performance by using K-Nearest Neighbor (K-NN) and Naive Bayes (NB) classification technique is proposed. Thus, by using the historical VM utilization, the task waiting time and the execution time have reduced, and also the resource utilization has increased. The count of the physical machine gets reduced by four using K-NN and NB classifier. The task performed by 28 physical machines when using Support Vector Machine (SVM), JAYA is reduced by using K-NN and NB classifier algorithm, and also the error got decreased by 0.025%.

Keywords: Scheduling, Execution time, Machine learning, Optimization, Resource utilization.