Research And Implementation Of High Concurrent Live Broadcast System Based On SRS Streaming Media Server

With the rapid popularization of 5G technology in recent years,live video has become the standard for information communication in all walks of life,especially due to the impact of the COVID-19,the traditional face-to-face information exchange has been impacted.For colleges and universities,teachers’ lectures,various cultural and sports activities on campus,knowledge competitions and expert lectures can be broadcast live to transmit live information to each student’s terminal device in real time,enabling teaching or entertainment anytime,anywhere.Activity.Nowadays,there are many types of live broadcast software in the market,but at present,almost all live broadcast software are commercial live broadcasts based on entertainment,games and shopping,and there are many problems in terms of security,flexibility and applicability.In response to the above problems,this paper implements a high-concurrency streaming media server cluster based on the open source SRS streaming media server combined with the Kubernetes clustering technology.At the same time,the cluster monitors the resource usage data in real time according to the Metrics monitoring component to realize the dynamic scaling of the SRS edge cluster.,to ensure the stability of a large number of users watching live broadcasts.At the same time,as a live broadcast system for campus users,this paper uses micro-service architecture,Dubbo and Zookeeper technologies to design and implement a campus live broadcast system.The live broadcast business system provides functions such as instant conversation,on-demand video viewing,data statistics,and authentication and authorization.First of all,this paper realizes the functions of SRS cluster deployment,automatic expansion and contraction based on Docker container and Kubernetes cluster management.The cluster is divided into source station and edge server cluster.The source station is responsible for video stream forwarding and HLS video stream transcoding and slicing functions.Transcoding uses H264 video encoding and AAC audio encoding technology to compress the original video stream,ensuring data quality while reducing The size of the transmission data volume,and finally complete the distribution of the HLS video stream through the Nginx server.The edge cluster completes the transcoding and distribution of RTMP and FLV video streams,which can effectively deal with high-concurrency scenarios in which a large number of users watch live broadcasts at the same time.At the same time,multiple streaming media protocols ensure the compatibility of different client devices for watching live broadcasts.The edge server cluster can quickly and easily adjust system resources according to the set elastic scaling strategy,allocate more system resources to new replica applications,and automatically reduce the number of Pod replicas when Pod usage is relatively idle,which is more conducive to SRS Cluster distributed deployment.Secondly,a live broadcast system with microservice architecture was designed and implemented,and the technologies such as Springboot framework,Dubbo,Zookeeper,and Redis used to realize the live broadcast platform,as well as the Shrio security framework to realize the authentication and authorization function were sorted out.The design and implementation process of the live broadcast system is described in detail.First,the system is divided into different microservices according to the implementation content and demand analysis of the live broadcast platform.The live broadcast service is responsible for live broadcast tasks and session management,and the on-demand service is responsible for live broadcast and review videos and organizer users.The management of uploaded videos,the data statistics service is responsible for the statistics and display of live broadcast data,and the authentication and authorization service is mainly for the administrator to assign user rights or roles to the organizer.The detailed design and database design are carried out for these modules,and then the system is implemented.Finally,perform performance test and function test on the system.The test results show that the live broadcast system can meet the design requirements in terms of streaming media cluster performance and system functionality.