|
With the rapid growth of multimedia technology and internet applications, media streaming plays a very important role in our lives. Online Video on Demand, Internet TV Broadcasting, Distance Education, Video Conferencing and other media streaming applications, enriched people’s life in a more and more profound way.Media streaming services requires massive storage, bandwidth and computing power. In large-scale media streaming systems, it’s hard for operation and maintenance personnel to quickly find the fault, locate the fault, and address the fault reasons, which result in high cost of operation. Therefore, automatic monitoring the media streaming system is necessary.According to this situation, we study on the quality of service detection in the application layer and stream layer. In stream layer, we do a survey of quality monitoring on RTP stream and MPEG2-TS stream in the industry. In application layer, design and implement three algorithms of real-time online fault detection of black fault, freezing fault and mosaic fault.Compared with some known algorithms to detect mosaic faults, which shows that our algorithm can dramatically reduce the false detection rate, and guarantee the detection rate in the premise,.We also study several systems architectural, including P2P, CDN, which can enhance the quality of service in media streaming systems. And propose a system architectural, using router caching to improve the quality of media streaming system.This system architecture leverages the free performance of edge routers, caching the most popular video data. Router caching system can save server bandwidth, reduce the initial buffer delay in the client, and improve the quality of streaming media system. Our simulation results show the performance of the system.
|