Re: Issue 171: Failure of ICE consent

We don’t know if this will be a temporary situation that are specific to certain interfaces or a global issue that cannot be resolved. Going to disconnected may allow for an ICE restart with new interfaces that might resolve the issue… or continue to remain disconnected in which cause the transport can be fully failed when all restarted candidate tests have completed.

-Robin

On January 20, 2015 at 6:24:03 PM, Iñaki Baz Castillo (ibc@aliax.net) wrote:

2015-01-20 23:48 GMT+01:00 Bernard Aboba <Bernard.Aboba@microsoft.com>:  
> Proposal is to treat a failure of ICE consent as a transition of the ICE Transport state to 'disconnected':  
>  
> disconnected  
> Liveness checks have failed (such as those in [CONSENT]). This may trigger intermittently (and resolve itself without action).  

It sounds correct. Just a question:  

What about if the STUN refresh request is rejected due to bad  
credentials? Which is the point of keeping the ICE state as  
"disconnected" when it is mostly clear that it will never "reconnect"?  



--  
Iñaki Baz Castillo  
<ibc@aliax.net>  

Received on Wednesday, 21 January 2015 00:25:40 UTC