​

Using Echo Pipeline Modes with a Server-Server Architecture

With both Audio and Text Echo pipeline modes you can use the Daily/WebRTC room to provide a video call interface to your user while controlling the replica. However you can also directly establish the connection with your server and receive the replica video stream directly back to your server for further processing.

​

When should you do server to server?

You should do server to server if you need to forward the video stream and cannot directly connect the viewer of the stream to the Daily room.

Examples:

• If you are doing a one to many stream (live streaming) and need to pipe into RTMP
• If you are piping into a third party meeting platform like Zoom or Google Meets and need to create a virtual camera using the video feed

Note that going down the server to server route will add latency, as the video stream has to make it back to your server and then you have to transmit it to the end viewer.

​

When should you connect your viewer directly to the Daily WebRTC stream?

We recommend this for most use cases to minimize latency. You can directly connect your viewer to the Daily WebRTC stream minimizing any latency. You can choose whether to use the Daily WebRTC room we create for a conversation to receive your viewer’s video/audio stream, or create a direct client/server connection with your own server.