Multi-server optimal bandwidth monitoring for QoS based multimedia delivery

Network aware multimedia delivery applications guarantee QoS parameters like time limit for media object transmission by actively monitoring the available bandwidth of the network and adapting the object to a target size that can be transmitted within the given time limit. A critical problem is how to obtain an accurate enough estimation of available bandwidth, possibly for distributed servers, while not wasting too much time in bandwidth testing. In this paper, we present an algorithm to determine optimal amount of bandwidth testing between one client and several distributed servers given a probabilistic confidence level for network-aware multimedia object retrieval applications. Our algorithm treats bandwidth estimation on each channel as sampling from an actual bandwidth population, and then uses a statistical method to estimate the combined bandwidth of all the channels. Simulation results show that the algorithm finds out the optimal amount of bandwidth testing and achieves the user specified QoS confidence level effectively.