Energy Efficient Speed Scaling and Task Scheduling for Distributed Computing Systems

With the rapid development of the performanceof computer systems, energy consumption hasbeen increasing dramatically as well. Energy efficiency hasbeen paid much attention, especially in large‐scale distributedcomputing systems. We propose an effective approachfor energy reduction, by dynamic speed scaling andtask scheduling simultaneously. Markov models for distributedcomputing systems are proposed, and detailedanalyses of the models are provided. Markov decision processes(MDP) are applied for problem formulation, andMDP algorithms for obtaining the optimal solutions areintroduced. The efficacy of our approach is further validatedby simulation experiments.