[Ecrit] RE: LoST returning location used??

["Brian Rosen" <br at brianrosen.net>]

I can't follow this email.

If the UA (foolishly in my opinion) puts more than one location in an
INVITE, and does not do the routing, then the proxy is going to pick one,
and which one it picks will, indeed, not be known to the UA.  

Actually, if there is an error, the "loop back" mechanism we are discussing
for -conveyance would return the location that was found to be in error, and
it might be marked "used for routing".

I fail to see a problem.  If the UA wishes to not be confused, it should not
put more than one location on the INVITE.

Now, the UA may put a location in the INVITE, and then the network may put
another on.  The UA may not know that one.  I don't see a problem with that
either, and the loopback is designed to make sure that the entity putting
the location on knows which one the error is referring to.

Brian

> -----Original Message-----
> From: Shida Schubert [mailto:shida at ntt-at.com]
> Sent: Wednesday, May 09, 2007 2:52 AM
> To: ECRIT
> Cc: Henning Schulzrinne; Brian Rosen; James M. Polk
> Subject: LoST returning location used??
> 
> 
>  I might have overlooked something completely but while I was
> reviewing the framework document the following question came up.
> 
>  The framework document expects some sort of flagging on
> the location to indicate which location was used to do location
> based routing (LoST & non-LoST) when there are multiple
> location information in the message.
> 
>  According to the LoST-05, the client can insert multiple
> location information but which location was used for the actual
> resolution is not returned to the client. One may guess or sometime
> clearly find out which location was used, simply analyzing the
> "ServiceBoundary" but according to the debate going on right
> now it doesn't seem so simple after all.
> 
>  If the client can't tell which location was actually used for routing,
> how would the client, let's say a proxy using the "used-for-routing"
> parameter indicate which location was actually used for routing the
> call.
> 
>  The framework also expects the entity on the signaling path
> acting as a LoST client to use the same location information flagged
> with "used-for-routing", would it restrict the LoST client to include
> only the flagged location information for LoST query once it's
> flagged using "used-for-routing"??
> 
>  Regards
>   Shida


_______________________________________________
Ecrit mailing list
Ecrit at ietf.org
https://www1.ietf.org/mailman/listinfo/ecrit