Re: [sipcore] I-D Action: draft-sparks-sipcore-refer-clarifications-04.txt

[Adam Roach <[e-mail]>]

On 10/21/14 13:08, Paul Kyzivat wrote:
> I'd like some clarity on what "must use a gruu" means.
> Does this mean the contact must have a "gr" parameter, or only that it 
> must be globally routable? (Note that 3261 has always required contact 
> URIs to be globally routable.) The key difference is that with a "gr" 
> parameter the other end can verify that the uri has been asserted to 
> be globally routable.

I'm using GRUU as a shorthand for "the mechanism described in RFC5627" 
because it's easier to type. I think applying the term "GRUU" to other 
kinds of URIs, regardless of their routing properties, is going to muddy 
the waters, and make it nearly impossible to talk about the RFC5627 
mechanism.

In any case, the global routability of Contacts hasn't been true in a 
formal sense since the introduction of RFC3327, and hasn't been true in 
a practical sense since the introduction of NATs. We can't rely on it 
being true; and, in fact, it's going to be false in the vast majority of 
cases unless RFC5627 is employed.

/a