Google says TCP congestion control algorithm BBRv3 performs well and will be submitted to the Linux kernel mainline this month

BBR (Bottleneck Bandwidth and Round-trip propagation time) is a set of congestion control algorithms released by Google in 2016. It is particularly suitable for use in weak network environments with a certain packet loss rate. For example, Google uses BBR to enable internal and external networks to operate more efficiently with higher throughput and lower latency. In such environments, the performance of BBR far exceeds that of traditional congestion control algorithms such as CUBIC.

BBR has been updated to version v3. Google said that BBRv3 has been widely used internally and has performed well. They are actively submitting BBRv3 to the upstream Linux kernel mainline.

Google engineers attended the IETF 117 event in San Francisco at the end of July. According to the engineers, BBRv3 includes various fixes and algorithm updates. In addition, BBRv3 reduces the packet retransmission rate by 12% and slightly improves latency.

According to the plan, Google engineers will submit BBRv3 to the upstream Linux kernel mainline TCP/networking module in August and upgrade the BBR module from v1 to v3 code. BBRv3 will adopt a dual license of GPL and BSD.

For more details, see:

• https://datatracker.ietf.org/meeting/117/materials/slides-117-ccwg-bbrv3-algorithm-bug-fixes-and-public-internet-deployment-00
• https://github.com/google/bbr/blob/v3/README.md