2016 IEEE First International Conference on Data Science in Cyberspace (DSC) (2016)
June 13, 2016 to June 16, 2016
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/DSC.2016.11
The problems of energy consumption controlling of data center is the hot spot on big data, cloud computing and green computing. In this paper we construct an energy consumption controlling frame of Xen virtual machines on Hadoop from the perspective of Hadoop scheduler: first, we propose an energy calculation model to calculate the energy consumption of tasks running in a virtual environment via which we can implement the real-time energy consumption measurement of physical machine and virtual machine, then based on the above model, we design Green Scheduler which can allocate resources to jobs rationally and efficiently by perceiving the status of each resource in real time. In addition, this paper compares 3 different schedulers including Green Scheduler, FIFO Scheduler and Capacity Scheduler on Hadoop and the result shows that on the premise of similar total runtime, Green Scheduler has more advantages in energy consumption controlling.
Energy consumption, Virtual machining, Real-time systems, Power demand, Green products, Resource management, Power measurement
J. Zhai, H. Zhang, X. Zhong, W. Li, L. Wang and Z. He, "Energy-Efficient Hadoop Green Scheduler," 2016 IEEE First International Conference on Data Science in Cyberspace (DSC), Changsha, China, 2016, pp. 335-340.