Profiling Energy Usage of Software Configurations for Virtualized Environment

A Software Product Line development configures consolidation of components that include modules and artefacts of a software ecosystem to aim its goals where different combinations may consume different amount of energy. In a virtualized system like Cloud Computing energy efficiency means a cost reduction that gives more revenue to their stakeholders. Moreover, many software in Cloud computing system were developed in a typical configuration way where reusability gives benefit to the ecosystem. This work combines Software Production Line Engineering, Cloud Computing and Energy Measurement technique to profile the consumption of energy that taking into account the workload of the system. In order to reveals the relationship between energy usage and processes within the system under measurement, we make use of Gaussian Process Regression to charac-terize the impact of workloads to software configurations in a virtualized environment. To demonstrate our method, a case study that combines of 12 software configurations for Cloud Computing environment, that retrieved from the Ubuntu software repository, will be presented. In order to reveals the information of energy usage, we developed a method to utilize the infrastructure of the virtualized system. Our profiling energy usage has open a chance to extend further research on dynamic energy modeling for virtualized environment.