Re: [dispatch] Fwd: New Version Notification for draft-kyzivat-dispatch-trs-call-info-purpose-00.txt

[Henning Schulzrinne <[e-mail]>]

The IP address requirement, while crude, addresses the need to ascertain whether the calling party is currently located in the United States since only such calls are reimbursable. (VPNs are a well-known issue, but the concern is not so much that an there will be an occasional VPN user but wholesale rule bypass. Almost all legitimate calls will originate from consumer or business end user addresses, not, say, cloud servers or other likely VPN entities. There are well-known and pretty accurate IP address-to-country databases, as well as databases of which ISP has the address. The number of VRS users is sufficiently small that any odd-looking patterns can be investigated manually.)

Apparently, since other countries, including other countries in North America, do not offer the same level of VRS service, there's an incentive to "loan" accounts to friends and family elsewhere, or to use it by citizens living abroad.

The idea is not "to identify the caller" as an individual; each caller is already authenticating to the service.

There could well be other uses for this, particularly as other proprietary video communication systems are bridged into VRS.

________________________________________
From: dispatch [[e-mail]] on behalf of Paul Kyzivat [[e-mail]]
Sent: Thursday, January 22, 2015 11:29 PM
To: Dale R. Worley
Cc: [e-mail]
Subject: Re: [dispatch] Fwd: New Version Notification for draft-kyzivat-dispatch-trs-call-info-purpose-00.txt


> The use of "original IP address" as the means to identify the caller
> seems rather odd to me.  But I assume that this is what the FCC
> specifies, and they have good reason to do this.  Given that:

It is what it is, whether it makes sense or not.

> Should the 'purpose' value be "original-ip-address"?


> Should the draft specify that the URI must have the syntax "sip:" +
> IPv4address?  It doesn't specify that now, although the usage has that
> restriction.  What about IPv6 addresses?  Or should we leave no
> restrictions on the URI to allow future improvements, and just note that
> the current practice is to use a very limited syntax?

I was trying to make the mechanism at least a little less special
purpose. In practice, when used for TRS, it will contain an IP address.
But the mechanism doesn't need to be that specialized.

Currently an IPv6 would be acceptable. And an h323 address might also
make sense. (I included use cases for that.) And in some other context
(without the FCC requirement) maybe a non-ip address would be ok.




        Thanks,
        Paul