寬頻網路的普及與MPEG-4 壓縮技術的發展促使網際網路上的資訊趨向更多元化的發展，而隨選視訊（Video-on-Demand）便是其中最熱門也最被看好的一項服務。然而，要將現有的影音格式轉換成MPEG-4規格需要非常長的壓縮時間，另外，雖然MPEG-4提高了壓縮比，隨選視訊系統仍需大量的儲存設備來存放影音資料，且目前與MPEG-4 相關的硬體設備的價格仍然居高不下。而個人電腦叢集（PC Cluster）系統的架構可同時解決這三種問題。 本論文中，我們利用Linux個人電腦叢集（PC Cluster）來達到高效能影像壓縮。並且利用Diskless Cluster技術使我們在系統管理上更加方便，在影像壓縮方面我們利用dvd::rip這軟體工具來平行壓縮影像，並且透過不同分切影像方式使我們能達到最佳的執行時間,使每個結點都能發揮最佳處理效能。 The wideband network popularization and the MPEG-4 compression technology development urge the Internet information to tend to the multiplex development. Among these techniques, Video-on-Demand is the most popular service, which potentially wins the public’s favors. However, it needs extremely long compression time to convert audio and video data into MPEG-4 format. Moreover, although MPEG-4 enhanced the compression ratio, it still needs massive storage equipment to deposit the audio and video material. In addition, currently the price of MPEG-4 related hardware equipment still stays at a high level. However, these problems can be solved simultaneously by using the cluster computing (PC Cluster) technology.In this thesis, we use the Linux PC cluster to achieve the high performance video conversion. We also use diskless cluster computing technology to make it more convenient in the system administration. In video conversion aspect, we use the software tool called “dvd::rip” to parallel process the video compression to achieve the best execution time. Meanwhile, it enables each node to perform in its best processing potency.
