I think there are multiple issues here:
1) what the sender places in the To and R-URI of the PUBLISH
2) Whether the (proxy) server that services the AoR for a
user should be presence-aware and be able to route PUBLISH
requests for presence to the presence server for the AoR
3) What the presence server looks at in a PUBLISH to decide
what to update.
When most requests are originated, a single address is used in both
To-URI and R-URI. (The exception is REGISTER, which can perhaps be
considered a historic mistake.) Making PUBLISH different complicates
things. If you expect the originator to insert the address of the
presence server in the R-URI, then it must know that. How would it? That
would be yet one more thing that needs to be configured. The point of an
AoR is that it is the Address of Record for a person. Its the place you
address things to reach devices that act on behalf of that person.
The proxy in general needs to be capable of routing requests to
different destinations based on attributes of the request. For instance,
routing to a VM server. The presence server is just another example of
this. Its how the home proxy earns its keep. If it can't handle this for
PUBLISH, then I suppose it can't handle it for SUBSCRIBE either. Then
you would have to configure all your buddies about your presence server too.
If you happen to end up with a proxy that doesn't have this capability,
all is not lost. Some UA registered for the AOR could be configured to
redirect PUBLISH requests to the desired presence server.
Once the request reaches the presence server, its actually PIDF in the
body that determines what will be updated - none of these choices affect
that.
Thanks,
Paul
Iñaki Baz Castillo wrote:
> 2009/4/14 Iñaki Baz Castillo <i...@aliax.net>:
>> El Martes 14 Abril 2009, Paul Kyzivat escribió:
>>> I think you are missing that there can be multiple publishers of
>>> presence *about* Alice that are themselves *not* Alice. So:
>>> 1) they can't assert that the publish is from Alice, because it isn't
>>> 2) they can't address Alice's presence server because they may not
>>> know it. They just know presence info about alice.
>> Ok, good points. But then, wouldn't make more sense to use the To URI?:
>>
>> Alice PUA Proxy Presence-Server
>> PUBLISH (1) ------------->
>> PUBLISH (2) --->
>>
>> (1)
>> PUBLISH sip:presence-ser...@domain SIP/2.0
>> From: sip:al...@domain
>> To: sip:al...@domain
>>
>> (2)
>> PUBLISH sip:presence-ser...@domain SIP/2.0
>> From: sip:p...@domain
>> To: sip:al...@domain
>>
>>
>> How can be logical to use a RURI that doesn't point to the real destination?
>> as I explained before, if the target is the RURI then theproxy routing that
>> request mustbe presence aware (it must not route the PUBLISH based on the
>> RURI
>> as a normal request).
>
>
> Hi, more comments on it please? I think that my suggestion makes sense
> while the real specification seems strange according to normal SIP
> rules.
>
_______________________________________________
Sip-implementors mailing list
Sip-implementors@lists.cs.columbia.edu
https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
