Videobridge

Videobridge is a component in video conferencing systems that routes media between participants in multiparty sessions. It typically functions as a selective forwarding unit (SFU). Unlike a traditional MCU, which decodes and re-encodes streams, a videobridge forwards selected streams without transcoding, reducing central processing and enabling scalable conferences.

How it works: Clients establish WebRTC connections to the bridge, sending audio and video streams. The bridge receives these streams and, based on signaling information, forwards only the streams each participant has subscribed to. It supports features such as simulcast, where a sender provides multiple encodings at different qualities, allowing the bridge to deliver an appropriate layer to each receiver. The bridge also handles networking tasks such as NAT traversal through ICE, STUN, and TURN, and may participate in RTCP feedback to assist adaptive streaming.

Advantages and limitations: Videobridges enable scalable multiparty conferences by avoiding full mesh topologies and extensive transcoding. They reduce server-side compute and bandwidth requirements for the bridge itself but can increase the number of incoming streams each client must process. End-to-end encryption models can be more complex to manage in SFU architectures, and security depends on the overall signaling and transport protections.

Examples and implementations: Jitsi Videobridge is a well-known open-source SFU. Other SFU implementations include mediasoup, Janus in SFU mode, and various commercial or cloud-based solutions integrated into WebRTC platforms.