Demonstrating the Impact of P2P Streaming on Video Quality

I. DESCRIPTION OF THE DEMO This demonstration will present a tool that allows the evaluation of the impact of P2P distribution on video quality, as a function of various parameters and design choices of the streaming system, such as: • media-aware chunkisation with and without substreaming vs. non-media-aware chunkisation; • different chunk trading protocols [2]; • different chunk/peer schedulers; • etc... Respect to the traditional evaluation methodologies, which are based on simulations and on the evaluation of some chunk loss and delay characteristics, the presented tool will show the impact on real videos, by dividing the video stream in chunks (according to different strategies, for example the one presented in [1]) and removing/corrupting the chunks lost during the P2P streaming process. The resulting (corrupted video) will be played simultaneously with the original one to show the artefacts introduced by the P2P streaming system. A graphical interface will allow to: • play the original video • configure and start the P2P diffusion process • play the received video stream (synchronised with the original video) • modify some of the streaming parameters, to evaluate their effects on the received video stream • compute and plot performance graphs on the fly, showing the impact of the changed parameters on various performance metrics The presented tool is highly modular, and can use different software components for computing the chunks lost during the P2P diffusion process. To keep the demo predictable (and to limit the requirements in terms of Internet connectivity and number of machines needed for the demo), a simulator will be used for computing the list of chunks lost by each peer. However, a real P2P streamer can also be used for this purpose, and will be demonstrated if possible.