Objective End-to-End QoS Gain from Packet Prioritization and Layering in MPEG-2 Streaming

Layered video coding as well as prioritized packet scheduling are two well-known methods that may improve the quality of service level in real-time applications with high bandwidth requirements, and are used over packet switched networks. However, it is often difficult to get an idea of, and to quantify, the actual gains that may be achievable, especially from an end-to-end perspective. In this paper, we present some experimental results from using temporally layered MPEG-2 video combined with basic per-layer IP packet prioritization. The goal has been to find out if a basic scheme is useful at all in combination with this particular source coding method, and if so, how much the objective video quality can be increased during bandwidthconstrained periods. The quality is measured in terms of PSNR and the results are compared to the case of equal packet priority. Also, different packet sizes as well as packet queuing disciplines are used. We conclude that using even a relatively simple temporal layering strategy in combination with packet prioritization can quite significantly improve the end-to-end quality of MPEG-2 video, especially in moderately bandwidth constrained situations. Furthermore, packet size and queuing discipline is found to have a impact.