Packet Media Streaming with Imprecise Rate Estimation

We address the problem of delay-constrained streaming of multimediapackets over dynamic bandwidth channels. Efficient streamingsolutions generally rely on the knowledge of the channel bandwidth,in order to select the media packets to be transmitted, according totheir sending time. However, the streaming server usually cannothave a perfect knowledge of the channel bandwidth, and importantpackets may be lost due to late arrival, if the scheduling is basedon an over-estimated bandwidth. Robust media streaming techniquesshould take into account the mismatch between the values of theactual channel bandwidth and its estimation at the server. Weaddress this rate prediction mismatch by media scheduling with aconservative delay, which provides a safety margin for the packetdelivery, even in the presence of unpredicted bandwidth variations.We formulate an optimization problem whose goal is to obtain theoptimal value for the conservative delay to be used in thescheduling process, given the network model and the actual playbackdelay imposed by the client. We eventually propose a simplealternative to the computation of the scheduling delay, which iseffective in real-time streaming scenarios. Our streaming methodproves to be robust against channel prediction errors, and performsbetter than other robustness mechanisms based on frame reorderingstrategies