Re: [MMUSIC] Liaison from 3GPP on how to encode the invalidconnection address in IPv4

[Jonathan Rosenberg <jdrosen at cisco.com>]

I think its dangerous to assume .invalid will work with IPv4 hosts. Most 
will not know that it means something special, attempt to look it up, 
fail the DNS query, and probably do something like terminate the call. 
We were able to make this work for v6 because there was no backwards 
compatibility issue.

So my response would be that they should not use .invalid for v4.

-Jonathan R.

Brett Tate wrote:
> > The liaison statement referenced below by Gonzalo indicates 
> > that the IPv4 0.0.0.0 is currently used as a means to signal 
> > an invalid connection address in IPv4.  The liaison statement 
> > asks  |  "whether the .INVALID address may be used to 
> >  |   indicate an unspecified IPv4 connection 
> >  |   address and does IETF offer any 
> >  |   recommendation in this respect ?".
> >
> > --- specification wise
> > We had a thread on mmusic related to something close to this 
> > back in 2004:
> > http://www.ietf.org/mail-archive/web/mmusic/current/msg02532.html
> > due to the requirement of rfc 3264 Offer/Answer to require an 
> > agent to accept 0.0.0.0 (" An agent MUST be capable of 
> > receiving SDP with a connection address of 0.0.0.0, in which 
> > case it means that neither RTP nor RTCP should be sent to the 
> > peer.").  The thread seemed to conclude that 0.0.0.0 should 
> > be used for IPv4 and .invalid for IPv6 but the question of 
> > .invalid for IPv4 was not debated directly.
> >
> > After a quick check, the SDP BNF notations from RFC 4566 and 
> > its predecessors RFC 2326 & 3266 seem to allow the use of 
> > .invalid for IP4 connections (a connection address may be 
> > represented as an FQDN and .invalid is a valid TLD).
>
> AFAIK, no rfc has defined TLD ".invalid" to have the same meaning as
> "0.0.0.0" within rfc3264.  RFC 3725 does indicate to look for future
> specification.  However AFAIK, only draft-ietf-sipping-v6-transition has
> updated rfc3264 with such usage but only discusses IP6.
>
>
> > --- implementation wise
> > 0.0.0.0 is typically handled by implementers given the 2543 
> > history and the 3264 MUST requirement.  How would today's 
> > implementations react to "c=IN IP4 .invalid"?  can folks 
> > provide some feedback?
>
> Although it appears valid per rfc4566 ABNF, isn't something needed prior
> to ".invalid" to be structured like a valid FQDN?
>
> Until more devices support IPv6, I assume many will not treat IP4 TLD
> ".invalid" like "0.0.0.0".  However once IPv6 supported, I assume some
> will interpret receiving IP4 TLD ".invalid" similar to "0.0.0.0";
> however they likely would not send it unless interoperability concerns
> have been alleviated.
> _______________________________________________
> mmusic mailing list
> mmusic at ietf.org
> https://www.ietf.org/mailman/listinfo/mmusic

--
Jonathan D. Rosenberg, Ph.D.                   499 Thornall St.
Cisco Fellow                                   Edison, NJ 08837
Cisco, Voice Technology Group
jdrosen at cisco.com
http://www.jdrosen.net                         PHONE: (408) 902-3084
http://www.cisco.com
_______________________________________________
mmusic mailing list
mmusic at ietf.org
https://www.ietf.org/mailman/listinfo/mmusic