Research On The Architecture Of Video Conference System And MCU Scheduling Algorithm Based On Cloud Platform

With the increasing development of information and communication technologies,video conference system are being applied and promoted in more and more enterprises.Constrained by the current network bandwidth,server performance,and software technology,video conference system still face many problems that we urgently need to improve.For example: high hardware costs,low equipment utilization,poor system reliability,cumbersome management and maintenance,and long communication delays.Therefore,it is of great significance to deeply study the problems encountered in the application of the current video conference system and provide high-quality conference services for users.The core of the video conference system is Multi-point Control Unit(MCU),whose performance and number will affect the service quality of the entire video conference system.This topic uses "cloud computing","virtualization","resource pool" and other ideas,and proposes put video conference deployment on the cloud platform,and builds a private cloud video conference system based on the MCU resource pool to realize the video conference cloud.The platform SaaS service provides enterprises with a cloud video conferencing solution that unified resource allocation management,high equipment utilization,good system reliability,convenient management and maintenance,low communication delay,and flexible scale of conferences.This Scheme is divided into hardware architecture design and software architecture design.The hardware architecture is the design of the underlying infrastructure of the entire video conference cloud platform.This topic is mainly focused on the research and implementation of MCU deployment methods,using the distribution MCU resource pool solution.Compared with traditional video conferencing MCU deployment,it has many advantages.The software architecture is mainly related to the research of the resource layer,management layer,and service layer of the cloud platform to realize the automatic deployment of resources and the efficient control and management of the conference.In a dynamic,autonomous,and heterogeneous cloud environment,server resource scheduling is related to the performance of the entire system.Therefore,in the cloud video conference system,reasonable scheduling of the MCU in the resource pool has become another research focus of this topic.In order to reasonably distribute the load of MCU and reduce the forwarding delay of video conference audio and video,this topic discusses the load balancing scheduling algorithm under various cloud platforms,and combines the characteristics of video conferencing to propose the load balancing of MCU under cloud video conference system.In terms of load balancing,the algorithm considers the intrinsic and external factors that affect the MCU load,and it designs and implements the traffic adaptation factor.Not only the multiple parameters that affect the MCU load are equivalently converted to load weights,but also the impact of traffic on the load is taken into account.In terms of reducing audio and video forwarding delays,the algorithm uses the conference ID number to assign terminals under the same video conference to the same MCU,and uses the IDMaps algorithm to measure the distance between the MCUs to provide the MCU with the closest cascading distance.Through performance testing,the algorithm is used for load balancing scheduling of MCU under cloud video conference,which can make the load of each MCU more balanced and the forwarding delay of audio and video is shorter.