由於越來越大量的運算需求增長，例如科學、商業以及網頁應用程式等等，對於大型資料中心來說意味著需要越多的能源。對於本篇論文來說，我們的目標便是希望最小化資料中心硬碟儲存系統的能量消耗，目前已存在許多儲存系統的省能技術，本篇便是根據能源認知排程演算法為基礎，該演算法的特點在於它能夠在不影響系統的效能以及不干擾儲存系統資料擺放方式的前提下，而能夠有效率地達到節省能源的效益。在這篇論文中，我們著重於尋找更好的即時排程演算法，我們的主要貢獻在於提出了幾項啟發式的即時排程演算法，且由以下列出四項在我們所提出的方法中我們所遭遇的挑戰：(1)預測未來的存取要求；(2)負載有變異性；(3)排程的資源有限；(4)處理寫入儲存系統的要求。本篇的實驗利用一個模擬器：Disksim 來模擬並分析我們所提出的演算法，而我們使用來評估的硬碟存取數據除了現實的硬碟存取數據也包含我們自己合成的數據，就整體結果來看我們的方法，我們所提出的即時排程技術確實達到了良好的省能結果而且也成功地實踐包含上述的四項挑戰。

Driven by the growth demand for computational power by science, business and web-application has led to the creation of large-scale data centers which consume enormous amount of power. In this thesis, our goal is to minimize the energy consumption of disk storage systems that used in those datacenters. Specifically, we investigate the approach of using energy-aware disk scheduling algorithm which has
been shown as a promising technique to reduce energy without causing significant system overhead and interference. In this work, we focus on the online scheduling solution. Our main contribution is to propose several heuristic online scheduling algorithms and address the following four four important challenges in our approach: (1)request prediction; (2)load variation; (3)limited scheduling resource and (4)write request. We extensively evaluate our solutions using diskSim simulator and workload traces from both real storage systems and synthetic workload generators. The results show our solution can effectively reduce energy using online scheduling techniques and overcome the four challenges in practice.

1 Introduction 1
2 Related Work 8
2.1 Finding Frequent Items in Data Streams . . . . . . . . . . . . . . . . 8
2.2 Time Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3 Other Energy Saving Techniques . . . . . . . . . . . . . . . . . . . . 12
3 Background 14
3.1 Scheduling Architecture and Models . . . . . . . . . . . . . . . . . . 14
3.2 Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4 Online Energy-Aware Scheduling 17
4.1 Online Scheduling Methods . . . . . . . . . . . . . . . . . . . . . . . 17
4.2 Irregular Loading and Limited Resource . . . . . . . . . . . . . . . . 21
4.3 Write Requests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
5 Experimental Setup and Result 24
5.1 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5.2 Energy Saving of Online Scheduling Methods . . . . . . . . . . . . . 25
5.3 Irregular Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.4 Limited Resource . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.5 Write Requests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
6 Conclusions 33

[1] Automaid [online]. Available from: http://www.nexsan.com/products/automaid.php.
[2] Copan [online]. Available from: http://www.sgi.com/products/storage/maid/.
[3] Disksim [online]. Available from: http://www.pdl.cmu.edu/disksim/.
[4] Dominion [online]. Available from: https://www.dom.com.
[5] Infrarati. Available from: http://www.infrarati.eu.
[6] National energy research scientific computing center [online]. Available from: http://www.nersc.gov/.
[7] Omnet [online]. Available from: http://www.omnetpp.org/.
[8] Pike research [online]. Available from: http://www.pikeresearch.com/.
[9] Renewable energy and data center projects. Available from: http://www.baryonyxcorp.com.
[10] Seagate cheetah 15k.5 [online]. Available from: http://www.seagate.com/www/enus/products/enterprise-hard-drives/cheetah-15k/.
[11] Storageio [online]. Available from: www.storageio.com.
[12] Umass trace repository: Oltp application i/o [online]. Available from: http://traces.cs.umass.edu/index.php/storage/storage.
[13] Dhruba Borthakur. The Hadoop Distributed File System: Architecture and Design. Apache Software Foundation, 2007.
[14] B. Boyer and J. Moore. A fast majority vote algorithm. Technical Report ICSCA-CMP-32, Institute for Computer Science, University of Texas, Feb 1981.
[15] Ronald S. Burt. Decay Functions. University of Chicago and INSEAD, 1999.
[16] Jerry Chou, Jinoh Kim, and Doron Rotem. Energy-aware scheduling in disk storage systems. IEEE Xplore, 2011.
[17] Graham Cormode and Marios Hadjieleftheriou. Finding frequent items in data streams. VLDB Endowment, 2008.
[18] E. Demaine, A. Lo ́pez Ortiz, and J. I. Munro. Frequency estimation of internet packet streams with limited space. European Symposium on Algorithms (ESA), 2002.
[19] M. Fischer and S. Salzburg. Finding a majority among n votes: Solution to problem 81-5. Journal of Algorithms, pages 376–379, 1982.
[20] R. Karp, C. Papadimitriou, and S. Shenker. A simple algorithm for finding frequent elements in sets and bags. ACM Transactions on Database Systems, (28):51–55, 2003.
[21] G. Manku and R. Motwani. Approximate frequency counts over data streams. International Conference on Very Large Data Bases, pages 346–357, 2002.
[22] G. S. Manku. Frequency counts over data streams. Very Large Data Base Endowmen, August 2002.
[23] A. Metwally, D. Agrawal, and A. E. Abbadi. Efficient computation of frequent and top-k elements in data streams. International Conference on Database Theory, 2005.
[24] J. Misra and D. Gries. Finding repeated elements. Science of Computer Programming, (2):143–152, 2008.
[25] Dushyanth Narayanan, Austin Donnelly, and Antony Rowstron. Write offloading: Practical power management for enterprise storage. Trans. Storage, 2008.
[26] Qingbo Zhu, Zhifeng Chen, Lin Tan, Yuanyuan Zhou, Kimberly Kee-ton, and John Wilkes. Hibernator: helping disk arrays sleep through the winter. SOSP'05: Proceedings of the twentieth ACM symposium on Operating systems principles, pages 177–190, 2005.
[27] Qingbo Zhu, Francis M. David, Christo F. Devaraj, Zhenmin Li, Yuanyuan Zhou, and P Cao. Reducing energy consumption of disk storage using poweraware cache management. 2003.
[28] Qingbo Zhu and Yuanyuan Zhou. Power-aware storage cache management. IEEE Trans. Comput., 2005.