Re: [rtcweb] [BEHAVE] New Version Notification for draft-chenxin-behave-turn-websocket-01.txt

[Sergio Garcia Murillo <[e-mail]>]

El 13/09/2013 11:23, Simon Perreault escribió:
> Le 2013-09-13 10:58, Chenxin (Xin) a écrit :
>>     We have been working on a new version of the TURN over Websocket 
>> draft, which is now available at:
>>
>>     http://www.ietf.org/id/draft-chenxin-behave-turn-websocket-01.txt
>
> There are two major things missing in this draft:
>
> - How are TURN UDP "channels" supported? (See RFC 5766 section 2.5.)
>
> - How are TURN TCP "data connections" supported? (See RFC 6062 section 
> 3.)
>
> I think the answers to these questions will greatly affect the overall 
> design of TURN-over-WebSocket.
>
>
Hi Simon,

As per RFC 5766:

    For some applications (e.g., Voice over IP), the 36 bytes of overhead
    that a Send indication or Data indication adds to the application
    data can substantially increase the bandwidth required between the
    client and the server.  To remedy this, TURN offers a second way for
    the client and server to associate data with a specific peer.

    This second way uses an alternate packet format known as the
    ChannelData message.  The ChannelData message does not use the STUN
    header used by other TURN messages, but instead has a 4-byte header
    that includes a number known as a channel number.



So, at the end, the ChannelData message is just another TURN message, so 
it will be encapsulated inside a websocket frame. If you feel it is not 
clear, we can add more details about it in next draft.

Regarding RFC 6062, we had some internal discussion regarding if we 
should add websocket support at all to it or not. Mainly, because AFAIK, 
webrtc will use RFC 5766 and not 6062. Anyone please correct me if I am 
wrong.

I agree with you that this has not been well covered in the draft. In 
order to make the minimun changes possible, and align it with the 
general view of changing just the turn client to turn server 
encapsulation protocol, I would propose to extend websockets usage only 
for the turn control connection in RFC 6062 but keep turn data 
connections on TCP. That, or drop RFC 6062 extension altogether and just 
focus on 5766. What do you feel it would be best?

Best regards
Sergio