Research On Dynamic Adaptive Streaming Optimization Algorithm Based On HTTP/2 Protocol

With the rapid development of streaming media transmission technology in recent years,various kinds of adaptive streaming media transmission protocols have been put forward one after another.In order to unify and standardize various protocol formats,DASH emerges as the times require.The development of DASH protocol is mainly divided into two directions:a variety of bit rate adaptive algorithms based on the parameters of bandwidth and cache,and the development of new protocols and mechanisms in the HTTP protocol transmission,such as HTTP/1.1,HTTP/2,etc.By analyzing and measuring the existing algorithms,the following problems are found:(1)At the beginning of the playback stage,the bandwidth strategy of the existing algorithm will spend a long time downloading resources to avoid the interruption,which will seriously affect the bandwidth utilization and the user's experience.(2)The existing algorithm based on the buffer strategy is unstable in face of the scene of bandwidth fluctuation,fast bit rate switching appears frequently,which also has great influence on the user's experience.(3)The existing algorithm has a slower rate switching at the beginning of the playback stage and a larger amplitude of buffer change.In view of the above problems,this paper proposes an adaptive algorithm based on HTTP/2 protocol to solve the problems.This paper introduces the characteristics of dynamic adaptive HTTP streaming technology and the latest development of HTTP/2 protocol at first.Then,an adaptive algorithm is designed for the push mechanism of HTTP/2 protocol.By determining the number of K in advance in each push cycle,the algorithm will not cause the waste of resources at the same time of improving the throughput.In addition,this paper proposes an improved algorithm combining congestion control theory with bandwidth strategy and cache strategy to solve the existing problems.Finally,a streaming media transmission system is designed to implement the proposed algorithm,and the bandwidth and network scene are simulated and controlled through the network simulator and the bandwidth shaping device to observe the differences of performance between the existing algorithms and the improved algorithms.Through the analysis and comparison of the experimental data,it is found that the perfermance of the improved algorithm proposed in this paper is better than that of the existing algorithm under different network scenarios.In general,the improved algorithm has a higher bandwidth utilization and a stronger anti-interference ability,which can improve the user's experience to a certain extent.