accepted, a NOTIFY is sent immediately (containing the current state
of what the subscribed-to entity knows about the subscription event).
If the subscription is to last "over time", then additional NOTIFYs
are sent when new changes to the "event" occur.
This, I don't believe, is something Conveyance has to account for,
but PhoneBCP certainly does. If can be as simple as having text
saying that subscriptions for dereferences need to last for 15 or
more seconds (i.e., enough to get that second location update from
"the system"), but it cannot be completely open ended. This should
cover the concept.
Do you agree?
James
At 05:20 AM 7/28/2009, Hannes Tschofenig wrote:
>I was more focused on the following scenario:
>
>The SIP proxy gets a location URI via SIP. Now, it wants todo location based
>routing for the emergency call. It will dereference the location URI as you
>mention in Section 4.5 of your document. The LIS does not know that this is
>request is only for location sufficient for routing and it tries to
>determine location at a best possible extend. This takes time. Then, when
>location is available the NOTIFY is sent. The SIP proxy can route the call
>but had to wait longer (potentially longer than Brian states in Phone BCP
>regarding the maximum call setup time).
>
>If I understood the HELD mechanism correctly then there is a way to indicate
>what type of reference the URI should point to (this relates a bit to the
>HELD context document). And the same stuff is available in HELD deref as
>well.
>
>Now, I was wondering whether this would be useful in
>draft-ietf-geopriv-loc-filters-05.txt as well. Namely, to let the SIP proxy
>tell the LIS when it sends a SUBSCRIBE that it wants location for routing
>(and that typically means faster>
>
>If it does not make sense to anyone then I remove it from
>draft-ietf-geopriv-loc-filters-05.txt.
>
>Ciao
>Hannes
>
>
> >-----Original Message-----
> >From: James M. Polk [mailto:jmpolk@cisco.com]
> >Sent: 28 July, 2009 13:02
> >To: Hannes Tschofenig; 'Thomson, Martin'; geopriv@ietf.org
> >Subject: RE: [Geopriv] Another Update to
> >draft-ietf-geopriv-loc-filters-05.txt
> >
> >At 11:54 AM 7/27/2009, Hannes Tschofenig wrote:
> >>Hi James,
> >>
> >> >> >More seriously: I'm not entirely sure how to interpret
> >> >> >responseTime in the context of a SIP PA. This might work as you
> >> >> >have described, but I would suggest that it complicates things
> >> >considerably. There's
> >> >> >discussion on this point in my draft here:
> >> >> >
> >> >>
> >><http://tools.ietf.org/html/draft-thomson-simple-cont-presence->va
> >> >> >l
> >> >> -req-02#section-4.6.1.1>
> >> >> >
> >> >> >I don't know how to solve this problem yet.
> >> >> >
> >> >>
> >> >>I am not sure I see the problems.
> >> >>
> >> >>Imagine a SIP proxy receives a SIP message with an LbyR
> >and wants to
> >> >>retrieve location information for routing. Depending on the
> >> >URI scheme
> >> >>(HTTP vs. SIP) he would either use
> >> >>http://tools.ietf.org/id/draft-winterbottom-geopriv-deref-prot
> >> >ocol-03.t
> >> >>xt
> >> >>Or
> >> >>http://www.ietf.org/internet-drafts/draft-ietf-geopriv-loc-fil
> >> >ters-04.t
> >> >>xt
> >> >
> >> >I'm confused.
> >> >
> >> >For SIP, why would the filters doc be used when Conveyance already
> >> >has how a server dereferences a location URI for routing?
> >> >
> >> >This isn't a soliciting action requiring filters.
> >>
> >>How does SIP Location Conveyance dereference location URIs?
> >>My reading of draft-ietf-sip-location-conveyance was that it conveys
> >>location by value and/or by reference.
> >
> >I have a section on this. Specifically it is section 4.5
> >
> > "Using sip/sips/pres as a Dereference Scheme"
> >
> >I even have a pretty diagram.
> >
> >
> >>What "indication" is used in a SIP SUBSCRIBE to indicate "Please give
> >>me location information suitable for routing."?
> >
> >err, a proxy routes based on the location information it is
> >given (if it initiates a dereference transaction in the first
> >place, there is no "suitable-ness" to it.
> >
> >The location-based routing process asks for, and receives (or
> >not) the PIDF-LO of the Target, then processes the message
> >(i.e., making routing decisions) as if it were sent the
> >PIDF-LO in the original SIP request. There's no magic to a
> >location URI to get a routable location vs. an unroutable
> >location -- especially since there is no difference in our
> >that same proxy receives location by value.
> >
> >If you believe there needs to be a suitable-ness to location
> >information, then we need a new error code for cases in which
> >the supplied location isn't suitable, with exactly what an
> >inserter is supposed to do with that error code (i.e., in that
> >error code, there needs to be an indication what was missing -
> >other than the whole location - that needs to be in the
> >subsequent SIP request to make is worthy of being routed.
> >
> >Which do you want -
> >
> >#1 - route based on the available location supplied (by-value
> >or by-reference)?
> >or
> >#2 - a new error code and inserter behavior to know what to do
> >when insufficient location was provided in the initial SIP request?
> >
> >#1 is easy, cuz that's the way Conveyance is written now.
> >
> >#2 is much more complicated - both to the routing proxy, and
> >the burden on the location inserter to understand what isn't
> >enough location, and to correct for just that amount - to
> >satisfy the proxy's determination of what it thinks it needs
> >to route the message propoerly.
> >
> >James
> >
> >
> >>Ciao
> >>Hannes
> >
_______________________________________________
Geopriv mailing list
Geopriv@ietf.org
https://www.ietf.org/mailman/listinfo/geopriv
