Dynamic Voltage Scaling Scheduling on Power-Aware Clusters under Power Constraints

Energy consumption of computing systems and especially of large-scale systems has a growing concern nowadays due to economic and ecological reasons. To reduce energy consumption, a variety of mechanisms are proposed in the literature. One of them is Dynamic Voltage Scaling (DVS). Many systems impose energy consumption restrictions when tasks are executed. We propose an energy saving mechanism based on DVS that takes into account energy consumption restrictions imposed by the system during the execution of tasks. The proposed DVS mechanism maximizes performance without violating energy consumption restrictions. Furthermore, it achieves great energy savings by lowering processors, frequency and executes tasks under a lower power budget than the one imposed by the system. Simulation experiments show promising results when hard real-time CPU-intensive tasks are executed. Success ratio is kept at satisfactory levels (97%-99.2%) while energy consumption is reduced from 13.7% up to 21%.