[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Sip] 199 Open Issue: UAS sending 199

Title: 199 Open Issue: UAS sending 199
Christer,
 
Yes, that's what I meant.  Maybe this is a usage more common in IMS, but (as you know) in IMS it is possible for an S-CSCF to invoke application servers on behalf of the party (calling or called) it is serving.  The AS's may in turn establish early media dialogs.  It would be useful in terms of being economical with resources (particularly in the RAN) to release the associated resources as soon as possible.  Sometimes it's more than "useful", sometimes it's necessary, because of limitations e.g., in a "middlebox" (SBC or IMS equivalent thereof).  Today we redesign call flows to avoid such issues, but it would be nice not to have to.
 
I'm not sure this fits the proposed UAS rule.  My concern isn't with a single UAS that establishes multiple dialogs, it's with a series of UAS's each of which will generally only establish one dialog (and each of which is generally unaware of the others).
 
If I made S-CSCF's I'd probably feel that the AS's should "clean up their own mess" (issue a 199 prior to releasing control back to the S-CSCF); even if the AS had only established one dialog.  I suppose it's also possible for the S-CSCF to clean up after the AS, and arguably the S-CSCF is in a better position to know whether a 199 is necessary.  But it seems complicated to ask this of the S-CSCF since it would have to monitor the SDP of all messages to and from AS's it invokes, to know which ones established early media dialogs and failed to "close them out".
 
Is it problematic to allow a UAS to issue a 199 even if it has established only a single dialog?
 
tim
 

From: Christer Holmberg [mailto:christer.holmberg at ericsson.com]
Sent: Tuesday, November 25, 2008 1:31 AM
To: Dwight, Timothy M (Tim); IETF SIP List
Subject: RE: [Sip] 199 Open Issue: UAS sending 199

Hi Tim,
 
So, I guess the use-case you are thinking about is when an AS establishes an early dialog with the UAC, in order to play some early media stream, and then sends 199 afterwards to terminate that dialog?
 
I think that would fit with the proposed UAS rule, saying that a UAS which establishes more than one dialog would be allowed to send 199.
 
Or, did I missunderstand your use-case? :)
 
Regards,
 
Christer

From: Dwight, Timothy M (Tim) [mailto:timothy.dwight at verizonbusiness.com]
Sent: 19. marraskuuta 2008 8:45
To: Christer Holmberg; IETF SIP List
Subject: RE: [Sip] 199 Open Issue: UAS sending 199

Christer,
 
Sorry, when I said "mis-interpret what they see as forking" I meant to refer to serial forking, which is of course a form of forking.  So there's no mis-interpretation (except mine!). 
 
But anyway the thought was that by providing a way to explicitly release an early media stream the 199 could enable a "middle box" (SBC) that today blocks serial forking, to support it.  This is one of the use cases Hadriel mentioned.
 
tim
 
 
 

From: Christer Holmberg [mailto:christer.holmberg at ericsson.com]
Sent: Tuesday, November 18, 2008 4:14 PM
To: Dwight, Timothy M (Tim); IETF SIP List
Subject: RE: [Sip] 199 Open Issue: UAS sending 199

Hi Tim,
 
>My opinion is that 199 shouldn't be restricted to cases involving forking, since some of its utility lies in  
>mitigating the behavior of middle boxes that mis-interpret what they see as forking. 
 
I am not sure I understand. Could you please clarify?
 
>Is it possible to define a simpler version of alternative #4, giving general guidance as to when a 199 may be issued  
>and being prescriptive about what the recipient is to do upon its receipt? 
 
Personally I think it would be very difficult to come up with something general and useful.
 
Regards,
 
Christer
 
 
 

From: sip-bounces at ietf.org [mailto:sip-bounces at ietf.org] On Behalf Of Christer Holmberg
Sent: Tuesday, November 18, 2008 1:17 PM
To: IETF SIP List
Subject: [Sip] 199 Open Issue: UAS sending 199


Hi,

The main open issue in the 199 draft at the moment is when a UAS sends 199 - IF a UAS sends 199.

The alternative proposals I have at the moment are described below (can also be found in the slides I was supposed to present yesterday).


1.      UAS never sends 199:

In this case only forking proxies/B2BUA would send 199.



2.      UAS always sends 199:

The issue with this alternative is that 199 would be sent even if no forking has occurred - which can be assumed to be the case in a large percentage of all calls.

3.      UAS sends 199 if forking proxy does not support 199:

With this alternative the forking proxy would have to insert an indicator that it supports 199.

Also, the UAS may not know whether it is the forking proxy closest to the UAS that has inserted the indicator. This may not be a big issues, since I assume in most forking cases there will only be one forking proxy in the signalling path.




4.      UAS sends 199 once procedures have reached a certain state

With this alternative 199 would not be sent until certain actions have taken place on an early dialog

Example: preconditions have been indicated as met
Example: SDP answer has been sent

The issue is that one would always have to specify at what point of different procedures 199 would be sengt.



5.      UAC and UAS negotiate sending of 199 once the early dialog has been established:

With this alternative that UAC would tell the UAS that forking has occurred (could this be useful information also for non-199 procedures?), and that it wants 199 to be sent.

The issues with this alternative is that it may require additional signalling (unless PRACK/UPDATE won't be sent for other reasons) to inform the UAS that forking has occurred.

Other alternatives?

NOTE: Robert S also raised an issue on what Require: 199 means. But, I think that outcome of that issue may depend on what way forward we choose for the issue in this mail.

Regards,

Christer 

_______________________________________________
Sip mailing list  https://www.ietf.org/mailman/listinfo/sip
This list is for NEW development of the core SIP Protocol
Use sip-implementors at cs.columbia.edu for questions on current sip
Use sipping at ietf.org for new developments on the application of sip