SURVEY ON ARIMA Model Workloads in a DataCenter with respect to Cloud Architecture

Resource administration in a data center is a dire function that affects the operational cost (OPEX) of data center providers and the applications’ Service Level Agreement (SLA). Efficient resource management can maximize the resource utilization and guarantee each application’s SLA. Accurately predicting each application’s workload is a key requirement to provision an efficient resource management. Overestimating or underestimating the application workload results in the resource overprovisioning or under provisioning. In this paper, we apply ARIMA model specific workload application in the data center. ARIMA can automatically analyze the characteristics of the application workload time series and then applies different mathematical models to fit the corresponding application workload time series. The performance of the ARIMA based prediction model is tested through extensive simulations. By selecting the appropriate parameters of the ARIMA model by using the existing methods, the average application workload prediction errors during the day and the month are calculated (i.e., 7.01% and 6.73%, respectively) to provide the accuracy of ARIMA prediction model.