A multi-resource task scheduling algorithm for energy-performance trade-offs in green clouds

Previous studies on task scheduling for cloud computing mostly focused on either energy saving or upgrading performance only. However, with the development of cloud computing, users' demand becomes more and more diversified and difficult to meet with some previous studies' ideas. In this paper, we first introduce two good-quality algorithms: a time-aware algorithm and an energy-aware algorithm, which are designed for task scheduling in a heterogeneous environment. Secondly, based on the two algorithms, we propose to combine them and design a new algorithm named Energy-Performance Trade-Off Multi-Resource Cloud Task Scheduling Algorithm (namely ETMCTSA). Users can flexibly manage and control the energy and performance of a cloud system via tuning the probability parameterαof the algorithm. Thus, it realizes the goal of focusing on reducing energy consumption or saving time according to users' will. We conducted the simulated experiments based on MultiRECloudSim and evaluated ETMCTSA with several other task scheduling algorithms. The experimental results indicate that with a pre-specified α, users can tune ETMCTSA inclining to be more energy efficient or more time efficient.