Bug 1120065 implements a one shot TURN allocation. To make sure the de-allocation request actually makes it to the TURN server a new STUN transaction should be used for the request. Problem is that this STUN transaction would need to outlive the TURN context and self-destroy on success or error.
Phillip do you see on your TURN servers already if the existing implementation in 38 helps savings resources? Or do we need to implement this more complex solution? Or should we wait with the judgement until 38 is release?
Lets wait for 38 to be released for some time before deciding.
Nils, let's reevaluate this after Whistler
backlog: --- → webRTC+
Priority: -- → P2
One feedback from a TURN server provider indicates that they see 91% de-allocation rates for Firefox. At this point I would argue that we don't need to implement the more complex solution, unless/until someone thinks differently (feel free to re-open if you do).
Status: NEW → RESOLVED
Last Resolved: 3 years ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.