Over the signaling layer.

WebRTC needs to negotiate IP addresses with the other side of the call so they can communicate. For that purpose, it will send a slew of IP addresses (local, public, TURN routed) and then the remote side will try to connect to them. Once connected, media can be sent and received.

By “default”, WebRTC assumes it is fine to send any IP address if the user approved access to the camera or mic. This may give away its local IP address over signaling even if the call ends up over a VPN connection. There are mechanisms to overcome that for quite some time now, but “WebRTC IP leak” and the the concerns/FUD around it are still going strong.

I know this is an old thread, but I'm a little confused. The first step in the WebRTC connection is going to be a UDP connection to a STUN server to see if you're operating behind a proxy/nat/firewall/vpn. This will be done over UDP, which is routed over the VPN, so the answer is yes. The connection will then be directed to a TURN server so that communication is made using outgoing connections. If your VPN only proxies HTTP traffic, then yes this will leak your IP just as any other protocol you try to use would leak it. I would expect the correctly configured VPN to act as the sole, not preferred, gateway if you're trying to conceal your IP. The traffic to the TURN server would therefore also go through the VPN. Performance would suffer from the limitations of the VPN, but how does the leak occur?

Thanks Roland.

You can capture the computer audio by using GetDisplayMedia: https://developer.mozilla.org/en-US/docs/Web/API/Screen_Capture_API/Using_Screen_Capture

I only have one question, is possible to stream the audio of a device with WebRTC? I know is possible send the microphone audio, but what about the sound of a notification on my computer as a replacement of PC speakers?

Selçuk,

P2P doesn’t scale well beyond a small number of participants. In your case, you want sessions with thousands of participants. There are a team or two trying to achieve this without the use of media servers. I don’t think we’re there at this point due to many different reasons.

If you want to cater for such use cases, then WebRTC can be the solution for you but you will need to deploy media servers as part of the solution.

Steven, there is no extension needed in any of the browsers in order to use WebRTC. Up until recently, Chrome required an extension if you wanted to screen share in WebRTC (anything else worked without the extension) – and even that is going away now.