Implementation of a Smart Traffic Prediction and Flow Control Mechanism for Video Streaming

Recently, all-IP intelligent video surveillance/door system is one of the attractive and important applications in the world. However, because the current Internet architecture cannot provide sufficient QoS for such kind of real-time applications, the performances of them are often poor under congested or interfered Internet environment. When Internet traffics are congested or interfered, the frame loss and delay jitter of video streaming are severe. In this paper, we design and implement a smart traffic prediction and flow control mechanism on an industrial video streaming product to improve the above problems. Here, we adopt the principle of Kalman filter to design the mechanism. From the practical tests and evaluations, the proposed mechanism can indeed improve the transmission performance of video streaming and make the streaming playing more smoothly on receivers.