[90312] in North American Network Operators' Group
Re: Geo location to IP mapping
daemon@ATHENA.MIT.EDU (Robert Bonomi)
Mon May 15 19:06:47 2006
Date: Mon, 15 May 2006 18:04:53 -0500 (CDT)
From: Robert Bonomi <bonomi@mail.r-bonomi.com>
To: nanog@nanog.org
Errors-To: owner-nanog@merit.edu
> From owner-nanog@merit.edu Mon May 15 17:42:13 2006
> From: Kevin Day <toasty@dragondata.com>
> Subject: Re: Geo location to IP mapping
> Date: Mon, 15 May 2006 17:40:23 -0500
>
> We use a Geo/IP location database. It's surprisingly accurate, with a
> few exceptions.
>
[[ sneck ]]
>
> Comparing the database to the IP that our customers used to make
> purchases we exceed 95% accuracy in identifying the country, and
> 75-85% in city/state. The big exception is AOL, since their IP
> assignments are pretty well randomized with respect to geography.
*dynamically* randomized, no less, or so I've been told. As in: _all_ the
customer session addresses are assigned out of *one* DHCP pool.
I thought that 'unlikely', considering the mayhem on internal routing tables,
but the AOL rep was rather insistant that that _was_ the case.
They also have 'virtual' POPs, where they just backhaul ('tens if not hundreds'
of miles, in same cases) voice, rather than having any physical equipment
present.
