Re: STUN/TURN server for WebRTC testing

On 2/12/2014 8:23 AM, Dominique Hazael-Massieux wrote:
> Hi,
>
> To properly the firewall-traversal part of WebRTC, I'll need to have a
> STUN and likely a TURN server that can be contacted:
> http://dev.w3.org/2011/webrtc/editor/webrtc.html#rtcpeerconnection-interface
>
> There is a publicly available STUN server made available by Google:
> http://io13webrtc.appspot.com/#52 (stun.l.google.com:19302) but I
> understand we don't want to reply on external services in our test
> cases.
>
> I guess that integrating at least a STUN server in WPT (e.G. with
> https://github.com/jtriley/pystun) should be reasonably doable, but it
> might be expanding the capabilities of WPT beyond what was envisioned
> for the project.
This is a bit of a non-sequitur. If you're testing firewall traversal,
then by definition you're relying on resources outside of the current
machine. If you're doing so then setting up a STUN server as part of
your infrastructure should not be terribly imposing. There's no way to
actually test anything that would involve STUN on a single machine, so
it's not clear why you'd be worried about this particular case.

-Ted

Received on Wednesday, 12 February 2014 17:26:48 UTC