In this paper, we have developed, based on empirical results from 3 systems (the VOSAIC system, the hierarchical VOD system and the remote VCR system) an integrated video QoS metric - the weighted cost-satisfaction ratio. We discussed in detail the economic relation between the user satisfaction and resource consumption as well as the advantages of providing unified metrics for multimedia services within an economic framework. Due to the close tie between the empirical results in Section 3 and the modeling of the QoS metric in Section 4, the design of our metric-based QoS architecture implies enhanced performance and cost-effectiveness for the user and SP.
We are exploring the use of the proposed metric in the design of admission control policies and resource management algorithms for multimedia systems. Specifically, we are interested in developing adaptive policies that use the metric to measure performance upgrades and degradations, determine the sources of bottlenecks, apply trade-off techniques and possibly adapt parameters of service. We also continue to investigate the formalization of the economic framework, the optimization relations and its implications in the design of effective metrics for multimedia systems. Our current work focuses on extending the proposed metric to more generalized Internet and Web server environments by exploring any additional parameters such as signal fidelity, locality, cache size and security that may be needed in these environments.