Re: [Sip] Fwd: I-D Action:draft-lendl-enum-sip-lr-00.txt

[Otmar Lendl <ol at bofh.priv.at>]

Hi,

On 2008/05/06 09:05, Hans Erik van Elburg <hanserik.van.elburg at ericsson.com> wrote:
> 
> Interesting, but this has nothing to do with "UA loose-routing". Everything
> with the retarget or reroute problem. 

I took the "UA loose route" from the name of Jonathan's draft.

> It seems an interesting way of configuring loose routes, but it also still
> requires configuration of normal DNS entries as the resolver may not support
> this sip:lr NAPTR application.

Yes, for backwards compatibility normal E2U+SIP NAPTRs might be
provisioned, too.  I expect this to be used mainly in controlled
environements like private SIP peering platforms (Infrastructure ENUM
or private ENUM) where one has a cleared picture on who will query such
records.

> The requirement to know if the next hop supports this yes or no, seems to be
> less of an issue as the configurator of the DNS should be expected to be
> able know this at configuration time.

Correct, according to the original ENUM role model, the destination network
provisiones the ENUM NAPTRs pointing to its ingress elements.

> This mechanism can be combined with the Target header proposal. 

Could you give an example? I have difficulties imaging the message flows
and routing logic when doing tel:-URI based calling. This paragraph is
unclear:

   The Target header, if present, represents the current target.  In the
   absence of the Target header, a receiving entity must assume that the
   Request-URI contains the current target.  Note that the Target header
   is not used for routing, it is just a means of presenting the current
   target to the receiving entity, information that otherwise would have
   been lost.

So which is it? Does the receiving entity evaluate the target: header or