Traffic Shaping based Queue Management for Delay Sensitive Multimedia Applications

Abstract

Active queue management, as a promising approach to solve network congestion and control delay, has caught the attention of researchers in the past decades. However, most existing algorithms are sensitive to parameter settings or lack robustness in non-TCP flow scenario. This paper propose a traffic shaping method based on queue management. In order to shape the distribution of received traffic in queuing system into any desired distribution, a traffic shaping method by selectively dropping the arrival packets is then proposed. On the basis of this theoretical analysis the proposed queue management algorithm ITS (Input Traffic Shaping) turns the parameters adaptively for optimal system performance. The input traffic shaper serves as a guideline for packet dropping strategy and Poisson distribution is selected as the desired distribution. Simulation results show that compared with existing algorithms, ITS can control delay more effectively, reduce jitter and enhance system throughput in non-TCP flow scenarios.