Virtual Machine Provisioning Within Data Center Host Machines Using Ensemble Model in Cloud Computing Environment

In the digital age of exponential data proliferation and growing computing demands, efficient resource management within data centers is crucial. A key challenge is the provisioning of Virtual Machines (VMs), which significantly impacts performance and scalability. Traditional static provisioning methods are increasingly inadequate. This inadequacy leads to resource underutilization and/or over-provisioning. Consequently, these inefficiencies result in compromised operational effectiveness. Dynamic provisioning, predictive modeling, and machine learning are essential for efficient resource management in the increasing demand for cloud computing and datacenter services. This manuscript presents a novel ensemble approach named ARIMA, GA, and ANN (AGA) that combines predictive modeling techniques with data centre resource management to enhance virtual machine provisioning within host machines. The proposed ensemble AGA model aims to improve resource allocation and scheduling decisions, leading to better performance by reducing the. performance metrics, CPU Utilization (%), Memory Consumption (%), Network Bandwidth (%), Storage Availability, Mean Absolute Percentage Error, and average value of forecasting accuracy. The experimental results reveal that the proposed AGA ensemble model outperforms the other state-of-the-art techniques, ARIMA (Forecasting model), a hybrid model of ARIMA with GA (Evolutionary optimization model), and ANN (Predictive model).