Key Technology Research Of Multimedia Streaming Delivery

In recent years,Internet has witnessed the explosive growth of video services.Video content traffic accounted for almost half of the entire Internet traffic.With the development of streaming media services,streaming media transmission technology becomes increasing important.The design of new and efficient streaming media delivery technology is the guarantee to provide better network quality of service(QoS)and video quality of experience(QoE)for video users.This dissertation is based on broadband wired networks,and chooses streaming media delivery technology as the theme.Also,it gives an exhaustive and comprehensive study on the key technology of streaming media delivery,and aims to solve the problem of key streaming media delivery technology.This dissertation thoroughly studies the streaming media delivery technology in current international and domestic industry applications and standards,on this basis,the dissertation puts a particular effort on several efficient protocol optimization and the resource allocation algorithms,which are analyzed and demonstrated profoundly.The specific work and innovations of this dissertation are as follow:1.Research of RTP streaming media transmission technology based on router buffersBuffer size plays an important role in real-time streaming media delivery system.Meanwhile,RTP streaming media delivery technology is one of the most interesting key technologies in real-time streaming media transmission system.This dissertation proposes a method to improve user perceived performance under minimized buffer size,based on which system capacity and user perceived performance are thoroughly investigated and validated by simulation experiments.Through this analysis,it is shown that the method achieves much better system capacity and user perceived performance.In under-provsioned network,the proposed method can obtain 4dB to 7dB improvement in the PSNR value and get a quality level increase in the MOS.,2.Research on TCP algorithm in HTTP adaptive streaming technologyThis dissertation mainly analyzes the effect of interactions between HTTP adaptive streaming and TCP on system capacity,particularly,the dissertation focuses on the system performance from link average throughput and loss rate perspective.Firstly,the intial phase pacing algorithm is proposed to smooth the burst in the initial burst phase and skip the slow start after retransmission timeouts occurred in the trailing ACKs phase.Then the reducing tail loss algorithm is proposed to reduce the probability of RTO in the trailing ACKs phase by sending redundant packets.Results from both theoretical analysis and simulations indicate that the method of improving TCP algorithm can improve the whole system capacity.The average throughput gets a maximum of 1 time increase.3.Research on bottleneck problems in multi-user HTTP adaptvie streaming delivery systemThis dissertation focus on the whole system capacity and user perceived performance while lots of clients launching requests simultaneously causes bottleneck problems.This dissertation proposes an effective method to solve the bottleneck problems.This method includes two efficient traffic shaping algorithms,namely,Global perceptual video quality maximization algorithm and Revenue maximization algorithm.Simulation results proves that these two algorithms achieve much better system capacity and higher user perceived performance under a reasonable lower computational complexity.The research productions which are involved in this dissertation are published in IEEE journals and conferences.