Load Prediction Energy Efficient VM Consolidation Policy in Multimedia Cloud

Abstract

Multimedia cloud computing is a dynamic development on could computing by upgrading various kind of media services including text, image, video on the Internet. The most significant challenges of the rising popularity of multimedia cloud computing are energy utilization and green cloud computing. To enhance the efficiency of energy on multimedia cloud data centers, we enhance the virtual machine consolidation (VMC) framework. There are two phases in VMC: the virtual machine (VM) selection and the virtual machine allocation. There are many researchers who proposed the solution with VM allocation and VM selection separately. Related on these two policies, we proposed the fast up and slow down (FUSD) load prediction-based energy-efficiency VMC for data-intensive jobs in multimedia cloud infrastructure. According to the simulation results in CloudSim with real trace data of cloud platform illustrated that the proposed load prediction policy shows up to better performance for efficiency of energy consumption, service level agreement violation (SLAV), and a number of VM migrations (VMM), respectively, compared to existing traditional VMC. The proposed VMC policy can use for large-scale multimedia cloud platform where minimal QoS assurance, SLAV, and energy consumption is inevitable.