[90324] in North American Network Operators' Group
Re: Geo location to IP mapping
daemon@ATHENA.MIT.EDU (Roland Perry)
Tue May 16 07:03:11 2006
Date: Tue, 16 May 2006 12:01:19 +0100
To: nanog@merit.edu
From: Roland Perry <lists@internetpolicyagency.com>
In-Reply-To: <Pine.SOC.4.61.0605152255040.4154@paixhost.pch.net>
Errors-To: owner-nanog@merit.edu
In article <Pine.SOC.4.61.0605152255040.4154@paixhost.pch.net>, Bill
Woodcock <woody@pch.net> writes
> > I just tried that, says I'm 100 miles south of where I really am. That's
> > quite a long way out in a small country like England.
>
>1.3ms is longer in small countries like England?
I'm virtually certain it's not being done by propagation delay.
What they've apparently done is look up the RIPE database, and found
that my ISP has registered an address, with postcode, for the hostmaster
function.
They've reported the major town associated with the two most significant
(out of six) characters of that postcode (Hemel Hempstead), although the
address is actually in a smaller town twenty miles to the west (Stoke
Mandeville).
To complicate the issue, the ISP is formed by the acquisition of several
smaller ISPs, and it seems unlikely (from my knowledge of the local
topology) that the physical NOC is at the hostmaster's address.
Finally, the "tail" from the NOC to my house (which appears as one hop)
is over a connection into British Telecom's ADSL backbone, and then over
the BT internal network which supports their wholesale ADSL product, as
far as my local telephone exchange (which I can see out of my office
window) and a short length of local copper. There's nothing in either
the RIPE database, or timing of packets, which could say where in the
country that tail is delivered. The ISP has my billing address, of
course.
--
Roland Perry
