[rtcweb] TCP vs UDP for media

["Young, Milan" <[e-mail]>]

Most of the use cases within WebRTC center around human to human communication, and for such cases RTP transport makes sense.  But there are at least a few use cases where one of the endpoints could be a machine.  Since machines have the ability to buffer and take a quick nap on the job, the "realtime" requirement isn't as critical.  Often, in such cases, the more important consideration is whether all of the media was transmitted.

I'm relatively new to WebRTC, so I apologize if I'm rehashing settled issues.  But I'm wondering if any thought has been given to TCP as a media transport.  If framing and timeouts could be worked out, would other folks besides myself find this useful?
 
Thanks