Re: [Geopriv] Geopriv L7 LCP: New Requirement

From: Henning Schulzrinne ^lt;>
Date: Wed Feb 14 2007 - 21:24:02 EST

Just to make this a bit more concrete: The protocol-level requirement
I see here is that the ISP needs to be able to issue a location query
with some non-IP address, such as the port number) to the ISP LIS.
(And there needs to be some kind of authentication between the two
providers, but that's a relatively easy one.)

I think the requirements already address the need to accommodate
identifiers other than IP addresses.

On Feb 14, 2007, at 9:11 AM, Stark, Barbara wrote:

> The IP address IS a good identifier. But it needs to be correlated
> back
> to the service location through information that only the physical
> access provider has. So the device queries the ISP's LIS (with IP
> address as ID), and the ISP queries the access provider's LIS (using
> L2TP tunnel, port and other IDs). The device cannot directly query the
> access provider LIS, because of several reasons: (1) the device
> doesn't
> have an ID that can be used to directly query the access provider; (2)
> the access provider frequently isn't legally allowed to communicate
> with
> the customer at anything above a transport layer; (3) the access
> provider's LIS can't be located by the device.
> Yes, the IP address does ultimately correlate to the DSLAM port, which
> correlates to a circuit ID, to a service address, through the
> aggregation point, through the ATM PVC, and through all the other
> points
> traversed in the access and ISP networks. Between the ISP and the
> access
> provider, our operation systems have all the information needed to do
> this correlation. We've already looked at that, and we know it can be
> done. But to do that correlation, the LISs must communicate.
> This is rather well described and documented in the NENA location TID,
> Section 8. That TID is at
> This is the same TID that has the NENA requirements for an acquisition
> protocol.
> Barbara
> -

Geopriv mailing list
Received on Wed, 14 Feb 2007 21:24:02 -0500

This archive was generated by hypermail 2.1.8 : Wed Feb 14 2007 - 21:24:21 EST