(T) One of the main reasons why the Internet has been able to sustain ever-growing traffic in its infancy is because of its famous TCP network congestion-avoidance algorithm. The TCP sender starts sending data incrementally with a congestion window, that stops sending data if the link becomes congested. When the receiver is not receiving data, it assumes that the network is congested, and as a result, it decreases the rate at which it is sending data.
Google with Van Jacobson, one of the key contributors of the TCP/IP stack, has proposed to the IETF a new congestion control algorithm called BBR (Bottleneck Bandwidth and Round-Trip Propagation Time) that use two parameters: the estimated round-trip time for a packet and the maximum bandwidth available for the connection. Leveraging those two parameters, BBR seeks an optimal operating point for the connection with high throughput and low delay which is more adapted to the buffers used by the various types of network equipment in today’s access and core Internet networks.
- BBR Congestion Control Internet Draft
- BBR presentation at the IETF 97 meeting in Seoul in November 2016: BBR_IETF97
- BBR presentation at the IETF 98 meeting in Chicago in March 2017: BBR_IETF98
- Google blog posting on BBR
- Note: The picture above is the point (in red) where BBR is operating.
Copyright © 2005-2017 by Serge-Paul Carrasco. All rights reserved.
Contact Us: asvinsider at gmail dot com.