BBRv2+: Towards balancing aggressiveness and fairness with delay-based bandwidth probing

Abstract

BBRv2, proposed by Google, aims at addressing BBR’s shortcomings of unfairness against loss-based congestion control algorithms (CCAs) and excessive retransmissions in shallow-buffered networks. In this paper, we first comprehensively study BBRv2’s performance under various network conditions and show that BBRv2 mitigates the shortcomings of BBR. Nevertheless, BBRv2’s benefits come with several costs, including slow responsiveness to bandwidth dynamics and low resilience to random losses. We then propose BBRv2+ to address BBRv2’s performance issues without sacrificing its advantages over BBR. To this end, BBRv2+ uses delay information to cautiously guide the aggressiveness of its bandwidth probing. In doing so, it achieves fast responsiveness to bandwidth dynamics and fairness against loss-based CCAs at the same time. BBRv2+ also integrates mechanisms for improved resilience to random losses and network jitter. Extensive experiments demonstrate the effectiveness of BBRv2+. Specifically, it achieves 25% higher throughput and comparable queuing delay in comparison with BBRv2 in high-mobility network scenarios.