一個優質的隨選視訊伺服器除了提供使用者影片播放需求外，保持穩定不間斷的視訊服務也是相當重要的課題之一。本論文目的在設計一個具有容錯（Fault Tolerance）能力的MP4叢集隨選視訊伺服器，並提出一個影片配置策略來提高系統負載平衡效能。其中MP4叢集隨選視訊伺服器由負責分配系統負載的負載平衡伺服器及負責儲存影片並處理存取媒體資料請求的視訊伺服器構成。
　　本論文中分別針對負載平衡伺服器及視訊伺服器提出適當的容錯處理機制，在此機制中，負載平衡伺服器使用「備援負載平衡伺服器」方式回復錯誤，當錯誤發生時，啟動備援的負載平衡伺服器，並轉移故障負載平衡伺服器上的工作；視訊伺服器，則是利用系統中其他仍正常運作的視訊伺服器來有效分擔故障視訊伺服器工作的方式回復錯誤。
　　另外，本論文針對改善系統負載平衡效能方面，提出一個影片配置策略，將影片一熱門程度分類，不同等級的影片複製不同份數分散放置在叢集隨選視訊系統中的視訊伺服器上，此策略可以更有效分散熱門影片的需求，進而達到較佳的系統負載平衡效能。

　　For a high quality Video-on-Demand(VoD) server, it is important not only to support on-line playback operations, but also to provide stable and uninterrupted streaming service. In this thesis, we design and implement a PC-cluster based fault-tolerant MP4 Video-on-Demand server with a video stream placement strategy to improve the balance of system loads. Our architecture comprises load balance servers for balancing system loads and video servers for video file storage and video streams delivery.
　　In our architecture, we propose fault-tolerant mechanism to both load balance servers and video servers. Load balance server provides fault-tolerant capability by a redundant load balance server. When a fault occurs, the backup load balance server takes over jobs on the failed server to achieve fault-tolerance. Video servers provide fault-tolerant capability by transfering jobs of the failed video server to the rest of surviving video servers.
　　A video stream placement strategy is also proposed in this thesis to share and balance the loads among video servers. We classify videos according to video popularity into 3 categories (i.e. hot、normal、and cold). Then we replicate video files to different amount of copies according to the category they belong to and then distribute these replicas to other video servers. This strategy can effectively distribute requests for some very hot videos to more video servers and therefore can improve the balance of system loads.

[1] A. Dan, D. Sitaram, P. Shahabuddin, ”Scheduling Policies for an On-Demand Video Server with Batching”, Proceedings of ACM Multimedia, October 1994

[2] Apple Computer Inc., QuickTime Developer Connection, http://developer.apple.com/quicktime/

[3] Apple Computer Inc., “QuickTime File Format Specification”, http://developer.apple.com/documentation/QuickTime/PDF/QTFileFormat.pdf, 2000

[4] Apple Computer Inc., “QuickTime Streaming”, http://developer.apple.com/documentation/QuickTime/PDF/QTStreaming.pdf, February 2002

[5] Cristian F., “Understanding Fault-Tolerance Distributed Systems.”, Communications of the ACM, vol. 34 no. 2, Feb 1991

[6] David Singer, William Belknap, Guido Franceschini (Editors),” ISO Media File format specification”, ISO/IEC JTC1/SC29/WG11, 23 July 2001

[7] H. Schukzrinne, A. Rao, R. Lanphier, ”Real Time Streaming Protocol(RTSP)”, RFC2326,April 1998

[8] H. Schukzrinne, S. Casner, R. Frederick, V.Jacobson, ”RTP: A Transport Protocol for Real-Time Applications”, RFC3550-RTP, July 2003

[9] M. Handley, V. Jacobson, “SDP：Session Description Protocol”, RFC2327, April 1998

[10] P. Srisuresh, D. Gan, ”Load Sharing using IP Network Address Translation（LSNAT）”, RFC 2391, August 1998

[11] Richard B. Bunt, Dreak L. Eager, Gregory M. Oster, Carey L.Williamson, “Achieving Load Balance and Effective Caching in Clustered Web Servers”, University of Saskatchewan, January 20,1999

[12] Shang Rong Tsai, “Distributed Systems”, “http://turtle.ee.ncku.edu.tw/~homework/nfs/title.htm” Department of Electrical Engineering, National Cheng-Kung University

[13] “The Encyclopedia of Networking and Telecommunications”,
http://www.linktionary.com/re/fault_tolerance.html

[14] Yu-Pin Chen, “A Fault-Tolerant Video-on-Demand Server”, Department of Computer Science, National Chiao-Tung University, 1998

[15] Wen-Shih Hung, “An Effective Data Placement Scheme For Supporting Fault-Tolerance in Distributed Video Server Environment”, Department of Computer Science, National Chiao-Tung University, 1999

[16] Walter L. Heimerdinger, Charles B. Weinstock, ”A Conceptual Framework for System Fault Tolerance”, Software Engineering Institute, CMU/SEI-92-TR-33, October 1992

[17] 王春火, “具容錯能力之分散式Linux叢集伺服器系統”, 國立成功大學電機工程學系, 2002

[18] 林偉元, “電腦叢集上QuickTime Streaming Server（QTSS）研究”, 國立成功大學資訊工程研究所, 2003

[19] 邱柏凱, “建構於個人電腦叢集上MP4隨選視訊系統之設計與實作”,國立成功大學資訊工程研究所, 2003

[20] 顏培如, “具容錯能力之隨選視訊伺服器架構”,真理大學數理科學研究所, 2003