臺灣博碩士論文加值系統

隨著越來越多的新增資料與大量的分析運算需求，資料庫在整體資訊系統中扮演極重要的角色。資料庫系統處理資料的需求能力不斷的被提昇，以單一昂貴伺服器處理所有資料量的時代式微，漸漸改成以眾多機器平均承受負載。依照 netcraft.com目前的統計，到 2010年四月時，全球使用Apache網頁伺服器營運的網站有 205,368,103個[24]，而在開放原始碼應用搭配組合LAMP(Linux，Apache，MySQL，PHP)已居於領先地位。
因此我們希望MySQL資料庫管理者或系統規劃者能夠經由此研究報告資料，了解各種資料庫同步架構下之差異，並搭配負載平衡技術提升資料庫同步效能，讓建構的系統能以最少的代價得到系統最大效能。根據我們研究的成果顯示，導入負載平衡伺服器對於整個系統來說，網路頻寬的使用效率、伺服器的可靠度與安全性以及增強服務的可延展性都可獲得更好的效果，其中網頁完成時間的負載能力與使用者的平均等待時間都獲得了重大的改善。

Computer Information System has been fully adapted in all business lately. There is plenty of analysis being necessary for great amount of data. Such an environment made databases system the most important part in the whole Computer Information System. This result in sufficiently more processing power is needed for database system. A single super server to handle everything has been replaced by a group of servers working together to share the load.
This research uses the MySQL database replication with dynamic load balancing topologies to achieve the database performance enhancement effectively. Based on our evaluation, the MySQL database replication with dynamic load balancing topologies can reduce database replication time and web completion time, but also with more performance database system under high availability, good scalability and easy administration.

摘 要 I
ABSTRACT II
誌 謝 III
表 目 錄 VI
圖 目 錄 VII
第一章、 緒論 1
1.1 研究背景 1
1.2 研究動機與目的 2
1.3 研究方法 2
1.4 論文架構 4
第二章、 相關研究探討 5
2.1 虛擬機器解決方案 5
2.2 資料庫同步架構 7
2.2.1 單一MySQL資料庫伺服器 7
2.2.2 單一主資料庫及單一副資料庫 8
2.2.3 樹狀同步：單一主資料庫及多台同階層副資料庫架構 8
2.2.4 樹狀同步：主資料庫加多台副資料庫分佈於多階層架構 9
2.2.5 鏈狀同步：雙主資料庫架構 10
2.2.6 鏈狀同步：兩台以上多台主資料庫同步架構 11
2.3 負載平衡機制 13
2.3.1 Linux Virtual Server 14
2.3.2 LVS中的三種負載平衡技術 14
2.3.3 負載平衡演算法 18
第三章、 效能測試架構以及測試方法 20
3.1 效能測試架構 20
3.2 變數取樣及誤差值 21
3.3 效能測試方法 23
第四章、 研究成果 34
4.1 同步完成時間 34
4.2 網頁完成時間 38
4.3 網頁完成時間之分佈 42
第五章、 結論與未來展望 45
5.1 結論 45
5.2 未來展望 46
參考文獻 47
附錄一、各演算法之網頁完成時間分佈 50
附錄二、FEDORA 8 LINUX VIRTUAL SERVER安裝文件 55
附錄三、VIRTUAL SERVER VIA IP TUNNELING配置文件 58
附錄四、測試程式碼架構 61
附錄五、資料表架構 62

[1]Gustavo Alonso, “Introduction to Database Replication”, DS WS06/07, ETH Zurich, May 2007.
[2]Nicolai G. Andersen, Michael B. Herold, and Robin Sharp, “Hierarchical Load Balancing in Large Distributed Systems”, Informatics and Mathematical Modelling, Technical University of Denmark, January 2008.
[3]Gopal Ashok and Paul S. Randal, “SQL Server Replication: Providing High Availability using Database Mirroring”, Microsoft Corporation, August 2008.
[4]K. Daudjee , K.Salem, “Lazy database replication with snapshot isolation”, Proceedings of the 32nd international conference on Very large data bases , September 2006.
[5]S Drake, W Hu, D. M. McInnis, M. Skold, L. Thalmann, M. Tikkanen, O. Torbjornsen, and A. Wolski, “Architecture of Highly Available Databases”, Springer Berlin/Heidelberg, January 2005.
[6]S. Elnikety, F. Pedone, W. Zwaenepoel, “Database Replication Using Generalized Snapshot Isolation”, Proceedings of the IEEE Symposium on Reliable Distributed Systems, no. 1541186, pp. 73-84 , October 2005.
[7]G. Gheorghiu, “Performance vs. load vs. stress testing”, Agile Testing, February 2005.
[8]T. Groothuyse, S. Sivasubramanian, G. Pierre, “GlobeTP: Template-Based Database Replication for Scalable Web Applications”, Proceedings of the Sixteenth International World Wide Web Conference, May 2007.
[9]M. Y. Hwang, M. L. Chiang, and M. C. Wu, “Design and Implementation of Time-based Request Dispatching Algorithms for Web Clusters”, The National Computer Symposium (NCS2003), Taiwan, December 2003.
[10]J.R. Juarez, J. R. G. Mendivil, J. R. Garitagoitia, J.E. Armendariz, F. D. Munoz-Escoi, “A Middleware Database Replication Protocol Providing Different Isolation Levels”, In 15th Euromicro Intnl. Conf. on Parallel, Distributed, and Network-based Processing (WIP Track), Italy, February 2007.
[11]B. Kemme, “Implementing Database Replication based on Group Communication”, In Proc. of the International Workshop on Future Directions in Distributed Computing (FuDiCo 2002), Bertinoro, Italy, June 2002.
[12]M. Kruckenberg and J. Pipes, “Pro MySQL”, Apress, pp. 614, August 2005.
[13]Y. Lin, B. Kemme,P. M. Marta, J. P. Ricardo , “Applying database replication to multi-player online games”, In Proceedings of 5th ACM SIGCOMM Workshop on Network and System Support for Games (NetGames). ACM, October 2006.
[14]F. D. Muñoz-Escoí, J. Pla-Civera, M. I. Ruiz-Fuertes, L. Irún-Briz, H. Decker, J. E. Armendáriz-Iñigo, J. R. De González Mendívil, “Managing Transaction Conflicts in Middleware-based Database Replication Architectures”, In 25th Symposium on Reliable Distributed Systems, no. 4032502, pp. 401-410, October 2006.
[15]Baya Pavliashvili, “Introduction to Database Replication”, informIT, March 2004.
[16]J. Salas, R. Jiménez-Peris, M. Patiño-Martínez, B. Kemme, “Lightweight Reflection for Middleware-based Database Replication”, Proceedings of the IEEE Symposium on Reliable Distributed Systems, no. 4032500, pp. 377-390, October 2006.
[17]James E. Smith, Ravi Nair, “The Architecture of Virtual Machines”, Computer (IEEE Computer Society), vol. 38, no. 5, pp. 32-38, May 2005.
[18]Bhavin Turakhia, “Building a Scalable Architecture for Web Apps”, In Great Indian Developer Summit 2008, July 2008.
[19]Costa Walcott, “Taking a load off: Load balancing with balance”, linux.com, August 2005.
[20]Jeremy D. Zawodny and Derek J. Balling, “High performance MySQL: optimization, backups, replication and load balancing”, O''Reilly, 2004
[21]W. Zhang, “Linux Virtual Server for Scalable Network Services”, In Proceedings of the Linux Symposium, July 2000.
[22]V. Zuikeviciute and F. Pedone, “Conflict-Aware Load-Balancing Techniques for Database Replication”, In 23rd ACM Symposium on Applied Computing (ACM SAC 2008), March 2008.
[23]“MySQL 5.1 Reference Manual”, MySQL AB, Sun Microsystems, 2008.
[24]“2010 Web Server Survey”, available at http://news.netcraft.com/
[25]“Fast incremental file transfer tool”, available at http://samba.anu.edu.au/rsync/
[26]“File-synchronization tool”, available at http://www.cis.upenn.edu/~bcpierce/unison/
[27]“Linux Virtual Server Project Website”, available at http://www.linuxvirtualserver.org/index.h