Round Trip Time Estimation Based on Adaptive Filtering

To improve the performance of TCP transmission, round-trip time (RTT) estimation is often adopted as an effective approach. There exist many positive or passive methods for RTT estimation, but rarely of them take account of RTT's self-similarity which should not be ignored. In this paper, a new RTT estimation algorithm employing adaptive filtering is proposed. The algorithm is based on RTT's self-similarity. It uses the preceding values of RTT for its prediction. It is illustrated by simulation on NS-2 that our algorithm can predict well, especially in scenarios where the link state changes suddenly. Comparing with other RTT methods, this method is more simple and easier to be transplanted to online systems.