| Efficient adaptive Internet video streaming of scalable video requires suitable transport control algorithms to adapt the transmission rate to the fluctuating network conditions. The algorithms implemented in TCP, the most prevalent Internet transport control protocol, are not suitable for video streaming as they create undesirable rate and hence viewer quality fluctuations. In this thesis we propose a streaming congestion control mechanism that uses the receiver buffer to monitor the occurrence of congestion and to determine the degree of its effect on the video connection. This information is used in determining how much the connection should adjust its rate to adapt to the network condition, resulting in a more robust technique of predicting available bandwidth for the video session. The simulation results show that the scheme performs well under congestion and bandwidth probing, and is fair to TCP and other streams while maintaining a smooth throughput profile. |
