Re: GPS info in IP addresses

Tony Li <> Fri, 16 August 1996 18:38 UTC

Received: from by id aa25915; 16 Aug 96 14:38 EDT
Received: from cnri by id aa25911; 16 Aug 96 14:38 EDT
Received: from by CNRI.Reston.VA.US id aa11946; 16 Aug 96 14:38 EDT
Received: from ( []) by (8.6.10/8.6.10) with SMTP id DAA19842 for <>; Sat, 17 Aug 1996 03:32:22 +1000
Received: from ( []) by (8.7.5/8.7.3) with ESMTP id KAA19900; Fri, 16 Aug 1996 10:32:16 -0700 (PDT)
Received: (from tli@localhost) by (8.7.5/8.7.3) id KAA06201; Fri, 16 Aug 1996 10:32:12 -0700 (PDT)
Date: Fri, 16 Aug 1996 10:32:12 -0700 (PDT)
Message-Id: <>
From: Tony Li <>
In-reply-to: <QQbdae01233.199608161409@rodan.UU.NET> (louie@UU.NET)
Subject: Re: GPS info in IP addresses

[inappropriate mailing lists elided...]

   > The benefit of placing geographic information in the IP address, despite
   > the fact that we cannot (yet) use this info for routing, is that various
   > network devices will be locatable within a physical space.  This will be
   > a real benefit for people trying to locate devices that need service,
   > etc. in a large network.  No protocol to date provides such information.

   Er, there are resource records defined in the DNS for exactly this
   purpose.  Having the geographical coordinates imbedded within the
   network-level address doesn't seem to be a particularlly good fit for
   this particular application.

In fact, the resource records are _inappropriate_ for this because of the
dynamic nature of the location information.  And using the IP address is
even less appropriate.  We need those bits, thank you very much.

However, location IS a very valuable bit of information, both for network
management purposes of static equipment, for general tracking of mobile
equipment, and for just being nosy about where your neighbors house
_really_ is.  It would make sense to be able to extract this information
from a system using a TCP/UDP "small service" which returns a GPS location
string.  I would suggest that someone who is interested in this (and I can
think of at least one major router vendor who has the necessary GPS
information available), I suggest that someone develop the protocol and
then write the RFC.


p.s. I prefer a binary representation of the information.  However, if you
decide to do an ASCII representation, please be rigourous about the syntax
as folks WILL write parsers to it...