Evaluating Energy Consumption in a Different Virtualization within a Cloud System

Abstract

Cloud computing manages the isolation of processes within Virtual Machines by utilizing resources via hypervisor. In the last decade, specific application such as Big Data that mostly run in the Virtualized environment has attracted many Cloud providers. This opportunity has increased the growth of alternative technology such as Docker and Kubernete that made enhancement on top of a single host virtualization. However, diverse virtualization technology may have different characteristics, which may also affect the consumption of energy. This work provides information on the impact of workload into Cloud computing services by measuring the energy usage of three different virtualization technologies namely Virtual Machine, Docker and Kubernete in an Openstack system, Opensource Cloud computing environment. In order to trigger processes in the system under measurement, we use a workload technique that simulates incoming users to the system. A method of energy measurement without physical powermeter for several Virtualized systems is reported. Our experiment result shows that using similar workloads, for different virtualization technology, consumes a significant distinctive amount of energy.