隨著雲端計算的需求增加，資料中心冷卻的花費越來越被重視。然而，如何權衡性能與溫度管理使得熱管理的議題格外棘手。為了提供在互動性雲端服務的品質保證，本論文探索異質伺服器系統的熱感知MapReduce即時排程問題。本研究提出資料區域性感知功率控制器進行動態電源管理與動態頻率管理，使其減少系統總體耗能並滿足溫度和服務截止期限的限制。

With the increased requirements of cloud computing, the cooling cost is getting serious in data centers. However, thermal management has proven to be challenging due to the tradeoff that occurs between performance requirements and overheating. To provide quality of service for interactive web services, this study explores thermal-aware MapReduce real-time scheduling in heterogeneous server systems. A data-locality-aware power controller with thermal consideration is proposed to dynamically switch the power state and to switch the executing frequency of each server. The thermal efficiency of the proposed method was evaluated using a series of workloads, and impressive results were obtained.

[1] “Nvidia tesla gpu accelerators.” http://international.download.nvidia.com/pdf/kepler/TeslaK80-datasheet.pdf.
[2] G. Liu, M. Zhang, and F. Yan, “Large-scale social network analysis based
on mapreduce,” in Proceedings of the Computational Aspects of Social Networks (CASoN), pp. 487 – 490, 2010.
[3] Y. Li, H. Zhang, and K. H. Kim, “A power-aware scheduling of mapreduce applications in the cloud,” in Dependable, Autonomic and Secure Computing (DASC), 2011 IEEE Ninth International Conference on, pp. 613–620, IEEE,
2011.
[4] Y. C. Lee and A. Y. Zomaya, “Energy conscious scheduling for distributed
computing systems under different operating conditions,” IEEE Transactions
on Parallel and Distributed Systems, vol. 22, no. 8, pp. 1374–1381,
2011.
[5] N. B. Rizvandi, J. Taheri, A. Y. Zomaya, and Y. C. Lee, “Linear combinations of dvfs-enabled processor frequencies to modify the energy-aware scheduling algorithms,” in IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGrid), pp. 388–397, IEEE, 2010.
[6] Z. Du, H. Sun, Y. He, Y. He, D. A. Bader, and H. Zhang, “Energy-efficient scheduling for best-effort interactive services to achieve high response quality,” in IEEE International Symposium on Parallel & Distributed Processing (IPDPS), pp. 637–648, IEEE, 2013.
[7] T. V. T. Duy, Y. Sato, and Y. Inoguchi, “Performance evaluation of a green scheduling algorithm for energy savings in cloud computing,” in IEEE International Symposium on Parallel & Distributed Processing, Workshops and Phd Forum, pp. 1–8, IEEE, 2010.
[8] J. Shi, J. Luo, F. Dong, and J. Zhang, “A budget and deadline aware scientific workflow resource provisioning and scheduling mechanism for cloud,” in Computer Supported Cooperative Work in Design (CSCWD), Proceedings of the 2014 IEEE 18th International Conference on, pp. 672–677, IEEE, 2014.
[9] M. Alrokayan, A. V. Dastjerdi, and R. Buyya, “Sla-aware provisioning and
scheduling of cloud resources for big data analytics,” in Cloud Computing in Emerging Markets (CCEM), 2014 IEEE International Conference on, pp. 1– 8, IEEE, 2014.
[10] E. M. Elnozahy, M. Kistler, and R. Rajamony, “Energy-efficient server
clusters,” in International Workshop on Power-Aware Computer Systems, pp. 179–197, Springer, 2002.
[11] J. Leverich and C. Kozyrakis, “On the energy (in) efficiency of hadoop clusters,” ACM SIGOPS Operating Systems Review, vol. 44, no. 1, pp. 61–65,
2010.
[12] Y.-C. Kao and Y.-S. Chen, “Data-locality-aware mapreduce real-time scheduling framework,” Journal of Systems and Software, vol. 112, pp. 65–
77, 2016.
[13] H. Sun, P. Stolf, J.-M. Pierson, and G. da Costa, “Multi-objective scheduling for heterogeneous server systems with machine placement,” in Cluster, Cloud and Grid Computing (CCGrid), 2014 14th IEEE/ACM International
Symposium on, pp. 334–343, IEEE, 2014.
[14] H. Duan, C. Chen, G. Min, and Y. Wu, “Energy-aware scheduling of virtual machines in heterogeneous cloud computing systems,” Future Generation Computer Systems, 2016.
[15] L. T. X. Phan, Z. Zhang, Q. Zheng, B. T. Loo, and I. Lee, “An empirical analysis of scheduling techniques for real-time cloud-based data processing,” in Proceedings of the IEEE International Conference on Service-Oriented Computing and Application, pp. 1–8, 2011.
[16] Z. Tang, J. Zhou, K. Li, and R. Li, “A mapreduce task scheduling algorithm for deadline constraints,” Cluster Computing, vol. 16, pp. 651–662, 2013.
[17] C.-W. Lee, K.-Y. Hsieh, S.-Y. Hsieh, and H.-C. Hsiao, “A dynamic data
placement strategy for hadoop in heterogeneous environments,” Big Data Research, vol. 1, pp. 14–22, 2014.
[18] T.-Y. Chen, H.-W. Wei, M.-F. Wei, Y.-J. Chen, T.-s. Hsu, and W.-K. Shih, “Lasa: A locality-aware scheduling algorithm for hadoop-mapreduce resource assignment,” in Collaboration Technologies and Systems (CTS), 2013
International Conference on, pp. 342–346, IEEE, 2013.
[19] M. Khan, Y. Liu, and M. Li, “Data locality in hadoop cluster systems,” in Fuzzy Systems and Knowledge Discovery (FSKD), 2014 11th International
Conference on, pp. 720–724, IEEE, 2014.
[20] L. T. Phan, Z. Zhang, Q. Zheng, B. T. Loo, and I. Lee, “An empirical analysis of scheduling techniques for real-time cloud-based data processing,” in 2011 IEEE International Conference on Service-Oriented Computing and Applications (SOCA), pp. 1–8, IEEE, 2011.
[21] G. Caruana, M. Li, M. Qi, M. Khan, and O. Rana, “gsched: a resource
aware hadoop scheduler for heterogeneous cloud computing environments,”
Concurrency and Computation: Practice and Experience, 2016.
[22] Y. Mao, H. Zhong, and L. Wang, “A fine-grained and dynamic mapreduce task scheduling scheme for the heterogeneous cloud environment,” in 2015 14th International Symposium on Distributed Computing and Applications for Business Engineering and Science (DCABES), pp. 155–158, IEEE, 2015.
[23] S.-J. Yang and Y.-R. Chen, “Design adaptive task allocation scheduler to improvemapreduce performance in heterogeneous clouds,” Journal of Network
and Computer Applications, vol. 57, pp. 61–70, 2015.
[24] M. Zhou, H. Chen, X. Dong, and Z. Zhu, “Dynamic token based improving
mapreduce performance in cloud computing,” in Big Data and Cloud Computing
(BDCloud), 2015 IEEE Fifth International Conference on, pp. 180– 186, IEEE, 2015.
[25] “Apache, mapreduce.” http://hadoop.apache.org/docs/r1.2.1/mapred_tutorial.html.
[26] “File transfer time - data transfer speed calculator.” http://www.t1shopper.com/tools/calculate/downloadcalculator.php.
[27] “Project recs.” http://shared.christmann.info/download/projectrecs.pdf.
[28] M. Vor Dem Berge, G. Da Costa, M. Jarus, A. Oleksiak, W. Piatek, and
E. Volk, “Modeling data center building blocks for energy-efficiency and
thermal simulations,” in Energy-Efficient Data Centers, pp. 66–82, Springer,
2014.
[29] S. Baruah and N. Fisher, “The partitioned multiprocessor scheduling of sporadic task systems,” in 26th IEEE International Real-Time Systems Symposium (RTSS’05), pp. 9–pp, IEEE, 2005. [30] Y. Chen, S. Alspaugh, D. Borthakur, and R. Katz, “Energy efficiency for large-scale mapreduce workloads with significant interactive analysis,” in Proceedings of the 7th ACM european conference on Computer Systems, pp. 43–56, ACM, 2012.
[31] A. Verma, L. Cherkasova, and R. H. Campbell, “Two sides of a coin: Optimizing the schedule of mapreduce jobs to minimize their makespan and improve cluster performance,” in 2012 IEEE 20th International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, pp. 11–18, IEEE, 2012.
[32] “Dells next generation servers: Pushing the limits of data center cooling cost savings.” http://www.dell.com/downloads/global/products/pedge/data_center_cooling_fresh_air.pdf.
[33] Y. Peng, S. Wu, and H. Jin, “Towards efficient work-stealing in virtualized environments,” in Cluster, Cloud and Grid Computing (CCGrid), 2015 15th IEEE/ACM International Symposium on, pp. 41–50, IEEE, 2015.
[34] C. Bassem and A. Bestavros, “Network-constrained packing of brokered workloads in virtualized environments,” in Cluster, Cloud and Grid Computing (CCGrid), 2015 15th IEEE/ACM International Symposium on, pp. 149– 158, IEEE, 2015.