Transmission Control In Multimedia Video

With the vigorous development of Internet technology represented by 5G,multimedia video applications have shown a rapid development trend.Video conferencing,live streaming,and other businesses have greatly enriched people’s lives,and their traffic has also shown explosive growth.The popularity of video applications has brought massive data transmission demands,posing more stringent performance requirements such as low latency and high quality to the underlying network.Video applications can be divided into two categories based on their requirements for network functions:real-time communication video applications represented by conferences and live streaming,which have more stringent requirements for data delay;and video-on-demand applications represented by short videos,which require advance caching for playback and thus have higher requirements for data transmission speed.Traditional network protocols are difficult to meet the multidimensional requirements of new multimedia videos,and conflicts with the rapid development of video applications,which urgently requires innovation in underlying transmission control technology.Therefore,this article focuses on research on transmission control technology for multimedia video,and the specific research results are as follows:(1)Data scheduling and congestion control for delay-sensitive video streams.In response to the current block scheduling algorithm’s single metric consideration and transmission waste problems,this paper designs a set of profit formulas that comprehensively consider information such as data block priority,size,and expiration date.By selecting the data block with the highest profit for transmission,the user experience is improved.To solve the problem of performance degradation of data transmission in high dynamic environments,this paper designs a congestion control algorithm based on delay and throughput utility functions.By using Kalman filtering and removing outlier values to more accurately estimate bandwidth,the algorithm detects network bandwidth in two steps to perceive the upper limit of bandwidth faster,and perceives network congestion faster by using delay.The simulation results show that the algorithm improves the user Qoe experience in different application scenarios,and the average Qoe is increased by 6.1%compared with STRF+Copa.(2)Multipath scheduling algorithm for large-scale video transmission scenarios.Short video content providers such as TikTok use peer-to-peer content distribution networks(PCDN)to reduce the cost of content distribution networks(CDN).In response to the current multi-path scheduling algorithm’s inability to solve the ordered transmission problem in the MS-Multi communication mode of multiple servers to a single client in PCDN,this paper designs a "route selection packet" multi-path scheduling algorithm that allocates data packets based on connection quality and redundantly transmits data based on data urgency and network quality.Simulation results and real-world A/B experiments show that compared with minRTT,the algorithm proposed in this paper reduces the degree of data disorder by 31.7%on average and increases the video transmission speed by 23.7%.