Traffic-Smoothing for Delivery of Online VBR Media Streams by a Dynamic Window-Based Approach

Ray-I Chang, Meng-Chang Chen, Jan-Ming Ho and Ming-Tat Ko

psfileTR-IIS-98-013


Abstract

Traffic smoothing for delivery of online VBR media streams is one of the most important problems in designing multimedia systems. Given the available client buffer and the sliding smooth window, conventional approaches try to minimize bandwidth allocated in each window. However, they can not lead to the minimization of bandwidth allocated for transmitting the entire stream. Although a sliding-window approach is introduced by Rexford et al. in 1997 to further reduce the bandwidth allocated, it is time-consuming. To resolve these drawbacks, in this paper, an effective and efficient online traffic-smoothing scheme is proposed. Different from conventional constant-sized approaches, our approach can automatically decide the suitable sliding sizes to online smooth the burst VBR traffic. By examining various media streams, our approach is shown to have higher bandwidth utilization (or called bandwidth-occupancy in conventional approaches) and requires smaller bandwidth than conventional approaches. Considering the online transmission of a Star War movie, our obtained result is over 13% smaller in the required network bandwidth and over 4% smaller in the obtained network idle rate than conventional approaches. In this paper, a feedback control method is introduced to resolve the latency- and quality-tolerance applications. Besides, the relations between the characteristic of input traffic and the behavior of obtained scheduling results are also discussed.