Transport-CC (defined in draft-holmer-rmcat-transport-wide-cc-extensions, originally a Google contribution, widely deployed since 2016) is the per-packet feedback mechanism that powers modern WebRTC bandwidth estimation. Where classical RTCP receiver reports give coarse loss/jitter statistics every few seconds, Transport-CC sends a compact RTCP message reporting the precise arrival time of every recently-received RTP packet, every few hundred milliseconds.
The sender uses this stream of timing data to estimate bottleneck bandwidth and queueing delay. Google's GCC (Google Congestion Control) algorithm — the WebRTC default in Chrome and Firefox — uses Transport-CC feedback to adjust send rate continuously. The result is much more responsive bandwidth adaptation than classical RTCP allows, which is essential for sub-second-latency WebRTC where there's no buffer to absorb mistakes.
Transport-CC is asymmetric — the receiver sends rich feedback, the sender uses it to adjust. Both ends need to support it (negotiated via SDP transport-cc extension and the goog-remb / goog-cc lines). Every modern WebRTC stack supports Transport-CC by default; older or specialised stacks may fall back to receiver-side bandwidth estimation, which works but with less precision. Custom WebRTC pipelines for low-latency live streaming often write their own bandwidth estimators that consume Transport-CC feedback directly.
The book · Volume 3 of 7
Video Streaming: A Complete Guide to Video Delivery: From ABR and CDNs to Players, WebRTC, DRM, and the Economics of Streaming
Transport-CC feedback is what lets a WebRTC sender estimate bandwidth accurately, and missing or delayed feedback shows up as frozen or blurry video. This volume covers Transport-CC in detail as part of WebRTC, with bandwidth estimation, congestion control in real-time calls, simulcast and SFU behaviour. Written by Nikolay Sapunov, CEO at Fora Soft.
WebRTC video freezing or dropping quality because bandwidth estimation goes wrong?
Share call stats from affected sessions and we will return a Transport-CC and bandwidth estimation review with the SFU and client fixes for each failure pattern. 250+ video products shipped since 2005.
