Saahil Claypool, Mark Claypool, Jae Chung, and Feng Li
Loss-based congestion control protocols, such as TCP CUBIC, can unnecessarily fill router buffers, adding delays that degrade application performance, particularly streaming video. Newcomer TCP BBR uses estimates of the bottleneck bandwidth and round-trip time (RTT) to try to operate at the theoretical optimum - just enough data to fully utilize the network without excess queuing. We present detailed experimental results that show in practice, BBR can misestimate the bottleneck bandwidth and RTT, causing high packet loss for shallow buffer routers and massive throughput variations when competing with TCP CUBIC flows. We suggest methods for improving BBR's estimation mechanisms to provide more stability and fairness.
Mark Claypool, Feng Li and Jae Chung. BBR' - An Implementation of Bottleneck Bandwidth and Round-trip Time Congestion Control for ns-3, In Proceedings of the Workshop on ns-3 (WNS3), Mangalore, India, June 2018. Online at: http://www.cs.wpi.edu/~claypool/papers/bbr-prime/
Feng Li, Jae Won Chung, Xiaoxiao Jiang, and Mark Claypool. TCP CUBIC versus BBR on the Highway, In Proceedings of the Passive and Active Measurement Conference (PAM), Berlin, Germany, March 2018. Online at: http://www.cs.wpi.edu/~claypool/papers/driving-bbr/