臺灣博碩士論文加值系統

由於企業電腦主機的使用數量與日俱增，電子資料在企業營運管理上的應用，已是不可或缺的趨勢。為了提供多樣性的資訊服務，企業增加電腦硬體的採購、維護管理與電力成本的增加已是必然。在雲端運算的環境下，我們可以藉由虛擬化技術的應用在一台實體機器上可以模擬多個虛擬機器來提供多個服務，以降低營運成本。由於提供服務的虛擬機器存在有尖峰與離峰的負載狀況，為了有效利用運算資源與降低電力成本，我們在離峰時段，將虛擬機器整併，並關閉閒置的實體機器，而在尖峰時段再進行開啟。此外，透過系統資源使用的監測，在尖峰時段，我們進行系統負載分散到所有實體機器的動作，達到負載平衡的效果。
本論文研究與實作在企業應用雲端運算技術建置私有雲(Private Cloud)的應用上，提出一套結合負載平衡與虛擬機器整併的即時遷移機制，除了可充分利用實體機器資源，同時亦可降低電力成本，達到綠能科技的期望。

The electronic applications in the management of business operations is inevitable, due to the increasing number of computer hosts in an enterprise. In order to provide a diversity of services, the enterprises have to increase the procurement cost, maintenance cost and electricity cost.
The cloud computing has brought an opportunity to reduce operating costs. Through the virtualization technology, we are able to deploy multiple virtual machines on a single physical machine to provide multiple services, in order to efficiently use of computing resources and reduce the cost of electricity, during the off-peak hours, we consolidate the virtual servers and close the idle physical machine, while activate more machines at peak times. In addition, through the system resources monitoring, at peak times, system loads are distributed to all physical machines to achieve load balancing.
In this thesis, we apply cloud computing technology to build a private cloud. By integrating the features of both load balancing virtual machine live migration, our proposed system is able to make better use of the physical resources and reduce the cost of power consumption, which leads to the expectation of green technology.

第 1 章 緒論 1
1.1. 研究背景與動機 1
1.2. 論文貢獻 1
1.3. 論文架構 2
第 2 章 相關研究 3
2.1. KVM/QEMU 3
2.2. Libvirt 4
2.3. Live Migration 4
2.4. Back Propagation Neural Network (倒傳遞神經網路) 5
2.5. GlusterFS 5
2.6. MongoDB 6
2.7. Zigbee Power Meter 7
第 3 章 決策機制設計 9
3.1. Monitor Module 11
3.2. Load Predication Module 11
3.3. Load Calculation Module 13
3.4. Migration Module 14
3.5. Load Balance Module 15
第 4 章 相關的基礎實驗 17
4.1. 實驗設備與環境 17
4.2. 實體機器基礎實驗 18
4.2.1. 實體機器開關機實驗 18
4.2.2. 實體機器CPU負載耗能實驗 19
4.3. 虛擬機器相關實驗 19
4.3.1. 即時遷移的耗能影響實驗 19
4.3.2. 實驗(1)─大檔案傳輸對即時遷移的效能影響實驗 21
4.3.3. 實驗(2)─網站傳輸對即時遷移的效能影響實驗 23
4.3.4. 實驗(3)─虛擬機器的CPU負載預測實驗 25
第 5 章 系統效益評估 27
5.1. 機器開啟的決策實驗 27
5.2. 機器整併的決策實驗 28
5.3. 實體機器負載平衡的實驗 29
第 6 章 結論與未來展望 31

[1]Cloud computing – Wikipedia, http://en.wikipedia.org/wiki/Cloud_computing#Private_cloud, referenced in April 2013.
[2]Virtual machine – Wikipedia, http://en.wikipedia.org/wiki/Virtual_machine, referenced in April 2013.
[3]Mayank Mishra, Anwecha Das, Purushottam Kulkarni and Anirudha Sahoo, “Dynamic Resource Management Using Virtual Machine Migrations”, IEEE Communications Magazine, Vol.50, Issue.9, pp. 34-40, September 2012.
[4]Lanzheng Liu, Rui Chu, Yongchun Zhu, Pengfei Zhang, Liufeng Wang, “DMSS: A Dynamic Memory Scheduling System in Server Consolidation Environments”, 2012 15th IEEE International Symposium, pp. 70-75, April 2012.
[5]Anton Belogiazov, Rajkumar Buyya, “Energy Efficient Resource Management in Virtualized Cloud Data Centers”, 2010 10th IEEE/ACM International Conference, pp. 826-831, May 2010.
[6]KVM Main Page, http://www.linux-kvm.org/page/Main_Page, referenced in April 2013.
[7]QEMU, http://wiki.qemu.org/Main_Page, referenced in April 2013.
[8]libvirt: The virtualization API, http://libvirt.org/, referenced in April 2013.
[9]Frontpage Gluster Communtity Website, http://www.gluster.org/, referenced in Apr 2013.
[10]ZigBee, https://zh.wikipedia.org/wiki/ZigBee, referenced in April 2013.
[11]x86 virtualization - Wikipedia, the free encyclopedia, http://en.wikipedia.org/wiki/X86_virtualization, referenced in April 2013.
[12]Andrew J. Younge, Robert Henschel, James T. Brown, Gregor von Laszewski, Judy Qiu, Geoffrey C. Fox, “Analysis of Virtualization Technologies for High Performance Computing Environments”, Cloud Computing(CLOUD), 2011 IEEE International Conference, Indiana Univ., July 2011.
[13]Takahiro Hirofuchi, Hidemoto Nakada, Satoshi Itoh, and Satoshi Sekiguchi, “Enabling Instantaneous Relocation of Virtual Machines with a Lightweight VMM Extension”, Cluster, Cloud and Grid Computing (CCGrid), 2010 10th IEEE/ACM International Conference, May 2010.
[14]Yoshihiro Sugaya, Hiroshi Tatsumi, Mitiharu Kobayashi and Hirotomo Aso,“Long-Term CPU Load Prediction System for Scheduling of Distributed Processes and Its Implementation”, 22nd International Conference on Advanced Information Networking and Applications, pp. 971-977, 2008.
[15]Florin Pop, Alexandru Costan, Ciprian Dobre, and Valentin Cristea, “Prediction based meta-scheduling for grid environments”, In CSCS-17, 17th International Conference on Control Systems and Computer Science, pp. 128-136, Bucharest, Romania, 2009.
[16]Principles of training multi-layer neural network using backpropagation, http://home.agh.edu.pl/~vlsi/AI/backp_t_en/backprop.html, referenced in April 2013.
[17]MongoDB, http://www.mongodb.org/, referenced in April 2013.
[18]NoSQL, https://en.wikipedia.org/wiki/NoSQL, referenced in April 2013.
[19]Virt-Top, http://people.redhat.com/~rjones/virt-top/faq.html, referenced in April 2013.
[20]Wake-on-LAN, https://en.wikipedia.org/wiki/Wake-on-LAN#External_links, referenced in April 2013.
[21]ab - Apache HTTP server benchmarking tool - Apache HTTP Server, http://httpd.apache.org/docs/2.2/programs/ab.html, referenced in April 2013.
[22]lookbusy -- a synthetic load generator, http://www.devin.com/lookbusy/, referenced in April 2013.