Hadriel Kaplan wrote:
-----Original Message-----
From: Dean Willis [mailto:[EMAIL PROTECTED]
Sent: Thursday, January 17, 2008 2:13 PM
Call for clarity on exactly which problems we're solving with UA-
Loose-Routing, target, CPID, etc. Proposed: target identity, end-to-
end parameters.
Not exactly. I believe the idea in the UALR draft is to get the
target identity/params created by the nearest re-targeting event
relative to the UAS, where those created by the UAC is the default
such event (and thus the logical "farthest"). In other words, what
you say is true, except if the request got retargeted, the identity
and params will be those of the new target, and thus not end-to-end
and not always what the UAC sent. Right?
AFAIK, the original goal of UALR was to get parameters to applications.
The last time I thought I understood the draft, these were preserved
back somewhere on teh route stack so that the application could get them.
The other thing UALR does, and this is kind of neat, is allow different
sets of parameters to be stuffed in at different places, and for the UAS
to be able to differentiate where those parameters were inserted (at
least a little bit -- "original", "middle", and "last hop" being easy to
parse.).
For example, given the following, where "RTRG" means Re-Targeting,
"RRT" means Re-Routing, "RT" means Routing, and the single letters
represent connections:
RRT
+---+
|R1 |
B /+---+\ C
RT / \ RT
+---+/ \+---+
|P1 | |P1 |
A /+---+ +---+\ D
/ \
+---+/ \+---+
|UAC| |UAS|
+---+ +---+
UA-Loose-routing wants the req-uri seen on connection "A" I think.
I think so too.
To header gives you A.
PCPID gives you B.
Hist-Info gives you A,B,C.
and D is directly visible to UAS, should RT have inserted any.
But in the re-targeting scenario such as:
RTRG RRT
+---+ +---+
|R1 | |R2 |
B /+---+\ C E /+---+\ F
RT / \ RT RT / \ RT
+---+/ \+---+ D +---+/ \+---+
|P1 | |P2 +---+P3 | |P4 |
A /+---+ +---+ +---+ +---+\ G
/ \
+---+/ \+---+
|UAC| |UAS|
+---+ +---+
UA-Loose-routing wants the req-uri seen on connection "C" I think.
To header gives you A.
PCPID gives you E.
Hist-Info gives you A,B,C,D,E,F.
As I thought I understood it, UALR gives us the stack of A and C.
But I may well not understand.
However, this does illustrate that there's another axis of proxy
operations that can be happening.
In addition to affecting the route, the target, and the identity of the
target, proxies may also affect parameters.
But "repararametering" is such an ugly word. And it's an ugly concept,
when you think about the consequences. I'm tempted to think of all
reparametering as retargeting.
--
Dean
_______________________________________________
Sip mailing list https://www1.ietf.org/mailman/listinfo/sip
This list is for NEW development of the core SIP Protocol
Use [EMAIL PROTECTED] for questions on current sip
Use [EMAIL PROTECTED] for new developments on the application of sip
