Re: Tired of Waiting

jnc@ginger.lcs.mit.edu (Noel Chiappa) Wed, 15 July 1998 18:00 UTC

Delivery-Date: Wed, 15 Jul 1998 14:12:36 -0400
Return-Path: cclark
Received: (from adm@localhost) by ietf.org (8.8.5/8.8.7a) id OAA21352 for ietf-outbound.10@ietf.org; Wed, 15 Jul 1998 14:00:02 -0400 (EDT)
Received: from ginger.lcs.mit.edu (ginger.lcs.mit.edu [18.26.0.82]) by ietf.org (8.8.5/8.8.7a) with SMTP id NAA21325 for <ietf@ietf.org>; Wed, 15 Jul 1998 13:59:28 -0400 (EDT)
Received: by ginger.lcs.mit.edu id AA04431; Wed, 15 Jul 98 13:48:16 -0400
Date: Wed, 15 Jul 1998 13:48:16 -0400
From: jnc@ginger.lcs.mit.edu
Message-Id: <9807151748.AA04431@ginger.lcs.mit.edu>
To: ietf@ietf.org, patrick@namesecure.com
Subject: Re: Tired of Waiting
Cc: List@giaw.org, comments@iana.org, discussion-draft@giaw.org, domain-policy@open-rsc.org, jnc@ginger.lcs.mit.edu, perry@piermont.com

    From: Patrick Greenwell <patrick@namesecure.com>

    > Phone numbers aren't like routing information? Then why do we have area
    > codes and exchange prefixes?

You will note that the "local number portability" is just that - *local*
portability. Move from one state in the US to another, and get a new phone
number.

And even LNP is being done with Yet Another Giant Mapping Database, so that
when your incoming call arrives at a LATA, the IEC knows which LEC to give it
to. In other words, your "phone number" here in the New World Order *doesn't*
contain your exchange any more - rather, it's a portable "phone name" which
has to be mapped into a *real* phone number befire the phone switches can
route the call.

In other words, a phone number under LNP is just like a DNS name - it has to
be mapped into a real location before being useful. People don't think of
them that way because the *syntax* is identical - looks like a phone numbers,
works like a phone number. But they aren't really line numbers any more.


    > Further, I would love to have a truly permanent, portable IP
    > address(es).

Sure. Now, remembering the above details about how LNP really works, with
mapping on all inbound calls, please explain to us the details: are you going
to emit packets with these "portable" IP "addresses" in them?

If so, what happens: does each individual router along the path do a mapping
before routing the packet - or does the first hop router do the lookup once
and modify the packet? (If so, what about TCP checksums - not to mention
authentication...)


Gee, guys, where's Tim Bass? Maybe we could get him to explain all this to
Mr. Greenwell... :-) (Tim, if you're listening, my apologies in advance!)

	Noel