Re: Mucking with SDP

On 04/07/2015 09:16 PM, Roman Shpount wrote:
>
> On Tue, Apr 7, 2015 at 4:00 PM, Harald Alvestrand
> <harald@alvestrand.no <mailto:harald@alvestrand.no>> wrote:
>
>
>     From Byron's comment it seems that there are cases where IPv6 will
>     be produced by conforming implementations.
>
>     What's the situation under which IPv6 in the c= line will cause
>     problems?
>
>
> This can cause problems when communicating with IPv4 only end points
> which do support ICE and DTLS-SRTP. In such cases you would want to
> change the "c=" line to have IPv4 address and remove all the IPv6
> candidates. I believe Firefox is (or at least was the last time we
> tested it) one of the end points that had issues with IPv6 c= line and
> candidates.
>
> I think a more generic requirement would be a constraint that tells
> the browser to collect only IPv4 candidates similar to googIPv6 in Chrome.

As a long time advocate of the need to make IPv6 universally deployed,
the idea that one should address interoperability issues by asking apps
to turn off IPv6 support makes me seriously sad.

I'd rather change the default in JSEP (but that's an IETF matter).

Roman, can you raise that issue there?

> _____________
> Roman Shpount
>
>  


-- 
Surveillance is pervasive. Go Dark.