The Record-Route cannot change between a provisional and a final response. Once a Record-Route is established, it cannot change. I don't believe the protocol allows proxies that have record-routed on a transaction to ever drop out of (subsequent transactions on) the call. The UAS Contact may change, which affects the Route.
However, if a pre-loaded Route is used on the INVITE, the Route will have to be updated when a response with a Record-Route is returned. -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Medhavi Bhatia Sent: Friday, August 02, 2002 11:50 AM To: Kross Joachim ICM N PG U ID A 1; [EMAIL PROTECTED] Subject: Re: [Sip-implementors] record-route in 183? The response which creates a dialog should determine the record routes. Also the whole thing must be compliant with 3261, so the record route in 200 OK should overwrite the earlier one. As far as I can see there are only two places where you need to check: First response which creates a dialog (ie., has remote tag), and the 200 OK. ----- Original Message ----- From: "Kross Joachim ICM N PG U ID A 1" <[EMAIL PROTECTED]> To: "'Medhavi Bhatia'" <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]> Sent: Friday, August 02, 2002 12:38 PM Subject: RE: [Sip-implementors] record-route in 183? > This is still not entirely clear to me. What I gathered so far is that a > Record-Route header received in a 200 OK updates any route set that might > have been established in the early dialog phase. However, I believe that the > Record-Route header in different 1xx responses may be different from each > other, as a proxy is allowed to re-write (side-note: I guess that excludes > removal?) its own entry in the response. Whether it is useful or not is > another matter, but I could not find anything in the spec that prevents a > proxy from rewriting its own entry in the Record-Route header differently > for different responses. And even so, one can never prevent a proxy from not > following the spec, so I think it would be useful to more exactly define the > behavior. > > I could imagine two kinds of behavior: > > * Build the route set from the Record-Route header received in the response > creating the early dialog, not change it until the 200 OK arrives, at which > time the route set for the remainder of the dialog is built. Or: > * Build the route set from the Record-Route header received in the response > creating the early dialog and and rebuild it for any subsequent provisional > responses, until the 200 OK is received, at which time again the route set > for the remainder of the dialog is built. > > Or shouldn't it be similar to the SDP in responses (where the initial offer > was in the INVITE)? In that case, if I understand it correctly, the SDP in > the 200 OK is authoritative, but the UAS MAY include the same SDP in any > prior responses. Since the UAC MUST treat the first SDP it receives as the > answer, the media is effectively fixed with the first response that contains > an SDP, even if the 200 OK (wrongly) contains a different SDP. > > For the Record-Route and the ensuing route set, this would mean that > Record-Route in the response creating the early dialog defines the route set > not only for the early dialog, but also for the dialog created by the 200 OK > (even though that may possibly contain a differen Record-Route header). This > would however deviate from the provision that any headers in a response > during an early dialog are only valid during the early dialog. > > Which one should it be? > > Regards, > Joachim > > > -----Original Message----- > > From: Medhavi Bhatia [mailto:[EMAIL PROTECTED]] > > Sent: Thursday, July 25, 2002 3:42 PM > > Cc: [EMAIL PROTECTED] > > Subject: Re: [Sip-implementors] record-route in 183? > > > > > > I think that argument applies to the non-reliably sent PRs also. > > > > ----- Original Message ----- > > From: <[EMAIL PROTECTED]> > > To: "Medhavi Bhatia" <[EMAIL PROTECTED]> > > Cc: "Aditya" <[EMAIL PROTECTED]>; "Baniel Uri-CUB001" > > <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]> > > Sent: Thursday, July 25, 2002 8:50 AM > > Subject: Re: [Sip-implementors] record-route in 183? > > > > > > > > > > > > > > > > Hi, > > > > > > The 200 OK response & any number of reliable provisional > > > responses (RPRs) that belong to a particular dialog > > > (identified by these particular From and To header tags) > > > cannot have different Record-Route headers. > > > > > > If at all you get a different record-route header set, > > > the response would also have a different tag in the To > > > header, thereby, creating another dialog. > > > > > > The UAC needs to keep track of all route-sets for each > > > new dialog that results for the initial INVITE and for > > > refresh requests for the existing sessions. > > > > > > Regards, > > > Siddharth. > > > > > > --------------- > > > Siddharth Toshniwal @ Hughes Software Systems, > > > Bangalore, India. > > > > > > > > > > > > > > > > > > > > > "Medhavi Bhatia" <[EMAIL PROTECTED]> on 07/25/2002 06:08:13 PM > > > > > > To: "Aditya" <[EMAIL PROTECTED]>, "Baniel Uri-CUB001" > > > <[EMAIL PROTECTED]>, [EMAIL PROTECTED] > > > cc: (bcc: Siddharth J Toshniwal/HSSBLR) > > > > > > Subject: Re: [Sip-implementors] record-route in 183? > > > > > > > > > > > > > > > Any headers in provisional responses are valid only for > > early dialogs, > > > so the 200 ok headers should replace. > > > > > > However, I dont see how they would be different. The spec does seem > > > to be fuzzy in this respect. > > > > > > ----- Original Message ----- > > > From: "Aditya" <[EMAIL PROTECTED]> > > > To: "Medhavi Bhatia" <[EMAIL PROTECTED]>; "Baniel Uri-CUB001" > > > <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]> > > > Sent: Thursday, July 25, 2002 12:48 AM > > > Subject: Re: [Sip-implementors] record-route in 183? > > > > > > > > > > > > > > Will the UAC override the route Set made with 183 > > > > (provisionals) Record Routes with one from 200 OK? > > > > > > > > If multiple provisionals come will a new route set be > > > > made for each provisional at UAC ? > > > > > > > > -AD > > > > > > > > --- Medhavi Bhatia <[EMAIL PROTECTED]> wrote: > > > > > If the INVITE had the record route, the 183 MUST > > > > > have it. > > > > > Consider a scenario, where the 183 reaches the UAC, > > > > > and > > > > > it had a contact in it. If the UAC now wants to send > > > > > a BYE, > > > > > and there was no record-route in the 183, the BYE > > > > > may > > > > > bypass any proxies originally intended to be in the > > > > > call path. > > > > > > > > > > ----- Original Message ----- > > > > > From: "Baniel Uri-CUB001" <[EMAIL PROTECTED]> > > > > > To: <[EMAIL PROTECTED]> > > > > > Sent: Wednesday, July 24, 2002 7:48 PM > > > > > Subject: [Sip-implementors] record-route in 183? > > > > > > > > > > > > > > > > Is record-route returned in 183? (or only in the > > > > > final response that > > > > > follows > > > > > > such as 200)? > > > > > > > > > > > > > > > > > > > > > > > ============================================================== > > ========= > > > > > > Uri Baniel - Distinguished Member of Technical > > > > > Staff - Motorola > > > > > > Tel: (847) 632 4616; Fax: (847) 632 3963; > > > > > > "Learning that does not daily increase will daily > > > > > decrease. - EWC > > > > > > > > > > > > > > > > > ============================================================== > > ========= > [...] > _______________________________________________ Sip-implementors mailing list [EMAIL PROTECTED] http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors _______________________________________________ Sip-implementors mailing list [EMAIL PROTECTED] http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
