The Transmission Technology Research Of Adaptive Streaming Based On Client-driven

The streaming media means that the continuous media data transfers over Internet using streaming technology. The streaming technology's main characteristic is that, the consumers do not need to download the entire file to the local to play the media, just only the initial few seconds of data need to be downloaded to the local buffer. At the same time, in order to maintain the continuity of playback, the remaining data is still ongoing from the server to the consumers. The most important thing during the development of streaming media technology is the technology of streaming. Up to now, the technology of streaming mainly has RTSP/RTP, HTTP progressive download and adaptive streaming.The adaptive streaming means that, the media file has the same content can be encoded into multiple copies with different code rate, each copy has different quality level; each copy consisting of two to five seconds segments called chunks that are separated. Because each copy has a different quality, the chunks from these copies differ in quality as well. The adaptive streaming technology allows the client to request one data chunk a time and different quality level to react to varying network conditions.Due to the competing traffic or changing to a different routing path, the channel throughput may not be constant bit rate and may vary over time. This may cause increased data delays or even data loss over the network. The result is that the decoder buffer may underflow or overflow, leading to unwanted pauses and noticeable perceptual quality loss during playback. Thus, adaptive streaming regimes need to be considered.The major work of this paper includes the following aspects:Firstly, the paper studies the adaptive streaming technology. The major work is to categorize and compare the available adaptive streaming technology of domestic and abroad, analyze the most advanced adaptive streaming technology of abroad currently, e.g. Microsoft smooth streaming and Apple live streaming.Secondly, a adaptive streaming prototype system based on client-driven is designed in this paper. The client-driven means that chunks'quality level choice and adaptive stream switching is completed by client. A simple adaptive streaming regime is designed in the client of this prototype system, the regime uses average bandwidth as the reference and compares the current bandwidth with it, thereby, the quality level can be chose as the network conditions be estimated.Thirdly, as the adaptive streaming regime used in the prototype system has some disadvantages, the Markov Decision Progress model is introduced into the adaptive streaming regime and the Markov decision adaptive streaming regime is designed. The advantages of this regime have three points, chunks'average quality level is advanced as high as possible, chunks'quality level changing amplitude and frequency is decreased as low as possible, the video is played as smooth as possible.Finally, the adaptive streaming improved system is designed and realized by the NS2 simulation software. The Markov adaptive streaming regime is introduced into this improved system , which is designed based on the prototype system mentioned before. The experiment did on the improved system shows the availability of the regime and achieves the expected results.