近幾年來,由於電腦和通訊技術的快速發展,使得在網路上提供互動式多媒體服務變成是可行的。有越來越多的人除了可從報章雜誌、收音機、電視來獲得流行資訊及熱門新聞外,還可經由連上網際網路來取得這些資訊。然而,人們最想要獲得的熱門資訊是那些即時發生的事情。因此,他們想要連上網際網路來立即收看那些現場直播的節目。但是,由於提供影片的伺服器端必須要提供大量的頻寬才能滿足人們的需求。因此,在近幾年內是不可能實現的。所以,如何在有限的頻寬之下來提供越多人的服務變成是很關鍵性的問題。 由於人們收看的幾乎都是相同的熱門直播影片,所以影片所需的頻寬是有可能被共享的。假如我們能增加頻寬共享的機率,相信我們能減少頻寬的需求量。然而,現場直播節目的性質是:即時的播放出來並且它的產生率和播放速度是一樣的。我們不能夠事先取得影片並且存放到視訊伺服器內。更進一步地,使用者端也不能快速下載影片的後面部份。因此,雖然有很多的演算法被提出來減少頻寬的需求,但是都是為了那些可事先儲存在伺服器內的熱門影片而設計的,它們都不能很有效的拿來支援現場直播的節目。 在這裡,我們提出一個新的演算法,叫做“調適性現場直播演算法“。不但能支援時光平移播放,也能夠做到人數少時低頻寬的需求,人數多時保持在一定的頻寬之內。所謂的時光平移播放可以被分成兩種:第一種是可以允許晚到的人能夠從頭收看到正在進行的現場直播節目。第二種是面對即時性非常重要的節目,它們每小時播放的內容是會變動的,就像是每小時播放的即時新聞一樣。舊的新聞會被刪掉不播放,而新的新聞會被加入。 In the recent year, the developments in computer and communication technologies have made it possible to provide interactive multimedia services. More and more households are connecting to the Internet to get popular information and hot news in addition to newspaper, magazine, radio and TV. However, real-time information is the hottest one that people want to access. People could connect to Internet to watch the live video immediately. But it is impractical in the near future due to the video server has to prepare or reserve a huge bandwidth. Therefore, how to use the limited bandwidth to accommodate as many users’ requests as possible is vital. Since many people want to watch the same hot live video, the required bandwidth of live video is possible to be shared. If we increase the probability of sharing the bandwidth for the same video and using local disk to buffer posterior video data, we can reduce the require bandwidth of the video server. However, the characteristics of live video are playing/recording in real-time and its production rate is equal to playback rate. We cannot pre-fetch the video and store them in the disk of video server beforehand. Furthermore, the clients cannot fast-download the posterior portion of video. Many schemes are proposed to reduce the bandwidth requirements for stored popular videos, but they cannot be used to support live video broadcast perfectly. Herein, we propose a new broadcasting scheme, called Adaptive Live Broadcasting (ALB) scheme, which supports time-shifting playback and perform well over a wide range of request arrival rates. Time-shifting playback can be classified into two playback types: the first one is it allows the late user to watch an on-going live video from the beginning while recording is still in progress. Another one is its timeliness is very important. The broadcast contents per hour of the live video are different. The older news would be abandoned, and the recent news would be added.