Research On Dynamic Adaptive Streaming Over HTTP

The technology of Dynamic Adaptive Streaming over HTTP is mainly based on network conditions and equipment performance to adjust the playback media rate dynamically,which can achieve seamless delivery and playback of the multimedia content.Rate adaptive algorithm as the key of dynamic adaptive streaming technology has been a hot topic in academia.Through the analysis of the existing rate adaptive algorithm,it is found that:(1)Low bitrate media segments can be selected by the bandwidth-based rate adaptive algorithm to be downloaded for a long time at the beginning of the playback,which may be resulting in the lower utilization of bandwidth;(2)Buffer-based rate adaptive algorithm is so sensitive to bandwidth varied that it can easily lead to frequent and significant bitrate switching,which is seriously affecting the quality of user experience;(3)The rate of media segment is increasing slowly at the beginning of playback by the buffer level based rate adaptive algorithm.And the buffer change of client is greatly under the environment with large bandwidth fluctuation.In view of the above problems,the research on rate adaptive algorithm seems like significant.Firstly,this paper studies the main characteristics of adaptive streaming technology,and it also introduces and analyzes the existing enterprise plans about adaptive streaming technology.Secondly,this paper focuses on the MPEG-DASH,which designed to provide a unified standard for dynamicadaptive streaming technology.Then,this paper studies and analyzes the existing rate adaptive algorithms,and in order to improve the issues of the algorithms,this paper also proposes a modified rate adaptive algorithm based on the buffer level.Under the condition of buffer state instability,the proposed algorithm can reduce the number of bitrate switching and improve the rate of segment at the beginning of playback.At the same time,according to the main factors influencing the user's quality of experience(Qo E)in the adaptive playback process,the Qo E evaluation indexes and evaluation function can be defined.This paper also utilizes the analytic hierarchy process to confirm the weight coefficient of the different indexes in the evaluation function,and realizes the objective evaluation of adaptivealgorithm.Finally,this paper implements the proposed algorithm and contrast algorithms in the Android client.The web server and network simulation tool also be used to verify the performance and analyze the QoE of the algorithm in the different network environment.The experimental results show that,the proposed algorithm in this paper is relatively stable in the restricted stable network environment and the continuous varied network environment.Under the condition of the stable buffer,the proposed algorithm has the same performance in the stair case network environment.At the same time,the Qo E value of the different algorithms are calculated by using the Qo E evaluation function in the continuous varied network environment and the Stair Case network environment.The results also show that,the proposed algorithm can improve the user's quality of experience in the fluctuating network environment.