An Effective Available Bandwidth Estimation Algorithm for the Networks with Bursty Traffic

End-to-end available bandwidth estimation is important for several applications such as network management, traffic engineering, flow and congestion control, and rate adaptation in real-time multimedia streaming. Among the proposed available bandwidth estimation technologies, most of the estimation algorithms are based on the fluid cross-traffic model. Inevitably, the estimation accuracy is limited in the network environments with bursty cross traffic. In this paper, we develop an available bandwidth estimation algorithm which is based on the fluid traffic model but can effectively filter out the measurement noise incurred in network environments with bursty traffic. The proposed algorithm employs not only the gaps of two consecutive probing packets but also those of nonadjacent probing packets to evaluate the bottleneck link available bandwidth. In addition, a two-stage filtering scheme is proposed to enhance the estimation accuracy. Numerical results demonstrate that the presented estimation algorithm can effectively reduce the effects of bursty cross traffic on the estimation accuracy of the bottleneck link available bandwidth.