Improving energy efficiency for transactional workloads in cloud environments

Research on energy efficiency in data centers has been focusing on reducing energy consumption, and state-of-the-art techniques have been emphasizing on optimizing power and energy consumption at hardware and infrastructure levels of data centers. Although these techniques have achieved significant improvement in reducing the energy consumption of data centers, the increasing heterogeneity of the current workloads call for more holistic approaches to enable optimization at higher levels. the goal of this work is to look for new opportunities to further improve energy efficiency at the level of applications with a focus on transactional workloads. In particular, we propose the model to characterize the energy per job of transactional-based applications. the model is experimentally validated on a real federated cloud infrastructure. Alternative policies to optimize the energy consumption of transactional applications are evaluated on the basis of the model.