Design Optimization And Implementation Of Voice Architecture Over BigBlueButton-based Video Conferencing System

After years of development, video conferencing has become an indispensable part of daily life of various sectors and even in the general public. Compared with based on software video conferencing based on C / S architecture and hardware video conferencing, Web conferencing has advantages of lower price, more convenient installation, simple operation and user-friendly experience,it has attracted more and more people and become a hot research hotspot.The front end of the modern popular Web video conferencing system is mainly developed by Rich Internet Technology(RIA), which runs on the browser with the Adobe Flash Player plug-in, while the back end uses the streaming media to make the voice and video processed. The research object of the paper is the open source Web video conferencing system- BigBlueButton. The system makes full use of Flex technology to develop the front-end, through the RTMP protocol to interact open-source streaming media server, through SIP protocol and RTP protocol Red5 and open-source softphone exchange platform FreeSWITCH interaction, voice mixing and other processing. In the mechanism of the system which processes voice, the server creates two voice channels connected to FreeSWITCH for each client joining the voice conference at the same time, for listening and releasing voice data respectively. This means that,for any user, even if you only need to listen to the voice(that is, only need to use the data channel to listen, do not need to use the voice data channel to release), will also occupy two voice channels. When there are a large number of participants, at the same time, only a small number of people speaking, and most people are in the listening state, there will undoubtedly bring additional burden to the FreeSWITCH server, resulting in a waste of resources, but also affect the concurrent degree of the system.In view of the above problems, this paper bases on analyzing the whole architecture of the system in detail, makes deep research on the related problems of the voice architecture and the bottleneck of the system performance, and proposes two new speech architectures Program, and makes a detailed description about the two models and the realization of these two models of online switching design principles and the implementation. The actual test and the result of comparative analysis show that the optimized design scheme can improve the speech effect of the system in the multi-person conference and improve the resource utilization and the concurrency degree of the system to a great extent.