Median filtering simulation of bursty traffic

Abstract

The estimation of Retransmission Timeout (RTO) in Transmission Control Protocol (TCP) affects the throughput of the transmission link. If the RTO is just a little larger than the Round Trip Time (RTT), retransmissions will occur too often, and this increases congestion in the transmission link [1]. If the RTO is much larger than the RTT, the response to retransmit when a packet is lost will be too slow, and this will decrease the throughput in the transmission link. Currently, Jacobson's Algorithm [2] for estimation of RTO is implemented in TCP. He uses an Exponential Weighted Moving Average (EWMA) filter to estimate RTT and then determines RTO. The EWMA filter is good if RTT follows a Gaussian distribution. In reality, traffic on the Internet is bursty and follows a heavy-tailed distribution. Median filter has been recognized as a useful filter due to its edge preserving characteristic for image processing applications, and it performs well for heavy-tailed distributions. Thus the median filter is efficient for removing impulsive noise [3], [4]. In this paper, we demonstrate the applicability of the median filter in our simulations for timeout estimation in the presence of bursty traffic flows.