隨著網際網路的蓬勃發展,線上會議系統和線上數位教學平台的發展也跟著越來越興盛,在商業應用上較為著名的線上會議系統有:VSee、MeetingBurner、V-CUBE。線上數位教學平台則有edX、Big Blue Button Web Conferencing。在這兩種系統和平台上,可以歸類出幾種常用的功能:桌面共享、視訊分享、文件上傳與即時通訊等,桌面共享尤其重要,因為任何文字都沒辦法表現得比影像來得直觀與迅速。除了商業運用外也有開放源程式碼團隊著重於桌面分享技術,諸如JUV RTMP Client、BBB-ScreenShare和Red5-ScreenShare皆在桌面分享上做研究。 本論文針對桌面分享的整體架構與各階段流程做探討,接著提出優化區塊辨識法,希冀能提升桌面分享的整體效能,最後測試CPU、記憶體與頻寬使用率來當作依據,證明Red5-ScreenShare相較於JUV RTMP Client、BBB-ScreenShare兩套桌面分享軟體在CPU、記憶體與頻寬資源使用率上的各項數據評比都較為優異。
With the vigorous development of the Internet, online conferencing systems, and E-learning platforms that followed are also flourishing. The famous online conferencing systems in commercial applications are VSee, MeetingBurner, V-CUBE, Online E-learning platform edX and Big Blue Button Web Conferencing. These systems and platforms can be classified into several common features: desktop sharing, video sharing, file uploading and instant communications. Desktop sharing is particularly important, because text and image performance is more intuitive and accelerated. In addition to commercial products of the desktop sharing, JUV RTMP Client, BBB-ScreenShare and Red5-ScreenShare are the three popular open source softwares in the desktop sharing. In this paper, we first study the overall architecture for the desktop, and explore the various stages of the process, then scrutinize block identification methods to optimize and improve the overall performance of desktop sharing. And finally we evaluate the CPU, memory and bandwidth usage as a performance metrics to prove that the CPU, memory, and bandwidth resource utilization of Red5-ScreenShare is superior to JUV RTMP Client and BBB-ScreenShare.