Machine-Learning based Loss Discrimination Algorithm for Wireless TCP Congestion Control
暂无分享,去创建一个
Loss-based congestion control of TCP, which was designed for the wired environment in the past, has a performance degradation in wireless environments where channel errors occur more frequently. In this paper, we propose a machine learning based loss discrimination algorithm (ML-LDA) for wireless TCP congestion control. ML-LDA learns how to distinguish packet losses due to congestion and wireless channel environment using multi-layer perceptron (MLP). Based on the learning results, the congestion control classifies the cause of losses and does not reduce congestion window in case of random losses. In order to verify the performance of the proposed congestion control, we implemented the algorithm in the Linux kernel and configured a testbed where packet loss occurs randomly. We compared the experimental results with TCP RENO, TCP WESTWOOD+, and TCP VENO, and showed that the proposed ML-LDA has 98% packet loss classification accuracy in wireless channel environments, and average throughput is greatly improved compared to the existing congestion controls.
[1] Soung Chang Liew,et al. TCP Veno: TCP enhancement for transmission over wireless access networks , 2003, IEEE J. Sel. Areas Commun..
[2] V. Jacobson,et al. Congestion avoidance and control , 1988, CCRV.
[3] Ren Wang,et al. TCP Westwood: End-to-End Congestion Control for Wired/Wireless Networks , 2002, Wirel. Networks.
[4] Yoav Freund,et al. Large Margin Classification Using the Perceptron Algorithm , 1998, COLT.