Note that "grid" is gone from the latest gruu draft. It was deemed that
the feature provided by grid was orthogonal to gruu and should be dealt
with separately.
Paul
Frank Shearar wrote:
> <[EMAIL PROTECTED]> wrote:
>
>> Frank Shearar wrote:
>>> draft-ietf-sipping-cc-transfer section 5 tells us that we can use GRUUs
> to
>>> associate an out-of-dialog REFER to an INVITE dialog usage.
>> That's not correct, actually. A GRUU is used to ensure that an
>> out-of-dialog REFER reaches the UA which is the endpoint of the INVITE
>> dialog usage (because the AOR used in the INVITE may not route to the
>> same UA now). The Target-Dialog header in the REFER is used by the
>> transferee UA to associate the REFER with the correct dialog usage.
>>
>> (See items 1-3 in section 5, and the paragraph following.)
>
> It was those items that I was referring to. But I see I missed the crucial
> bit:
>
> Problem 2 can be
> addressed using the Target-Dialog header field defined in [6].
>
> Here I had been thinking that one used the "grid" parameter (*) to find the
> INVITE. Where on earth did I get that idea from? *sigh*
>
>> Using a GRUU alone to determine the target dialog usage is impossible,
>> as that GRUU could be participating in many dialogs.
>
> Yes; I had thought that each dialog would be established with "grid" using
> GRUUs, resulting in every dialog having a unique GRUU. That works; it's just
> not what the drafts/RFCs define.
>
> frank
>
> (*) I think I mentioned in my original question that my stack implements
> gruu-10, not gruu-11 (at least at the moment).
>
> _______________________________________________
> Sip-implementors mailing list
> [email protected]
> https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
>
_______________________________________________
Sip-implementors mailing list
[email protected]
https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
