Media Streaming TCP-Friendly Congestion Control Using Multiple Time Scale Prediction

Multimedia demands consistent quality of service (QoS) transport for commercial applications while underlying network traffic affects its streaming performance. Finding effective traffic control schemes capable of perceiving underlying self-similar network traffic has become an important problem for media consistent QoS transport. This paper presents a TCP- Friendly congestion control mechanism based on multiple time scale prediction (streamTFRC) for improving media traffic performance by maintaining traffic rate as smooth as possible, including constructing a 2-level multiple time scale congestion control mechanisms. Simulation experiments illustrate, compared with TFRC, streamTFRC can enhance the tracking ability to the network traffic and achieves better performance in terms of friendliness, fairness, and smoothness.