Scalable model for dynamic configuration and power management in virtualized heterogeneous web clusters

This paper presents a model for dynamic configuration of Virtualized Application Servers (VAS) based on an algorithm of linear complexity, achieving the scalability needed for big data centers and cloud computing platforms. The goal is to minimize the power consumption of the processing environment and maintenance of the quality of service (QoS) of the applications. The experiments evaluate our model as compared to the Linux CPU governors and a state-of-the-art approach based on optimization. The results show that our model conserves up to 49.6% of the energy required by a cluster designed for peak workload scenario, with a negligible impact on the applications performance.