Location informations and the DNS (GL, LOC, ...) (was:draft-costanzo-dns-gl-00.txt)

[Alfred Novacek <Novacek@pop.idv.uni-linz.ac.at>]

Hello!

I could not help (inspired by the discussions on the draft) to keep the
matter in the back of my mind during the past days, and I've come to the
conclusion that there are better places to put location informations than
the DNS.

Al Constanco, the major proponent of a new GL RR, cites as the major
application that needs such an RR a PING or TRACEROUTE utility that
displays its results geographically. Unless RFCs are rewritten to allow GL,
LOC, ... RRs in the additional data section of responses to PTR, ...
queries (which I doubt that it will pass IETF), a second DNS query is
needed to obtain the location information; but instead of asking the DNS
again, one could ask the device itself (using some new protocol) "Hello,
what's Your geographic location?".

Al Constanco blames the difficulty of obtaining geographic coordinates to
cause that LOC records are not widely deployed, and proposes a new GL RR
that is more simple to create; however, I seriously doubt that GL will help.

First, the major DNS servers (BIND, MS, ...) will have to be adapted to
support the new RR; I estimate it will take approximately one year *before*
DNS admins *may* take action.

Then, DNS administrators (and no one else!) will have to define the GL
records for all existing systems. That may be feasible for small sites; but
for DNS admins in charge of a company with many sites scattered around the
globe, it may be much work, that will not be reduced essentially by the
fact that the new RR is much more intellegible that another. (I think there
are some RRs around that failed to be deployed widely because they created
work in the first place - with no clear benefit perceived justifying the
effort).

BTW, I even don't know the geographic location of some web servers with
entries in our DNS servers. We outsourced them to internationally operating
web space providerse, and the internet ensures that their physical location
really does not matter to us.

A new protocol for this purpose may be very simple (a requester sends a
small UDP packet to the system; a special server application answers with
another UDP packet containing the requested information).

Given the simplicity of the protocol, it would take only a few days (or
perhaps weeks) to create servers for the most common platforms and place
them on the apropriate download sites. Maybe that many users will download
and install it on the computers they use because it is really cool to let
others know via the internet where they are.

Creating a separate protocol will also allow to use not only preconfigured
location informations, but also to obtain them dynamically via GPS (or some
other appropriate way). Supporting mobile devices this way is something
that is very hard to achieve with semi-static DNS, but something where a
separate protocol really shines: Think of a shipping agency that keeps
track of the location of its trucks; or imagine Your laptop has been stolen
and You told the police its (mobile) IP address - the thief will be very
surprised when he first turns on Your computer.

Al,
don't waste Your time and energies forcing Your proposal through the WG
(given the reactions of some members to Your draft, it won't be easy to
convince them). An (experimantal) protocol of Your own will get much faster
deployed, and - when realizing the benefits - potential users will be glad
to discuss their ideas, requirements, ... with You.

HTH - Alfred Novacek

------------------------------------------------------------------------
Dipl.-Ing. Alfred Novacek
Institute for Data Processing in Business Administrations,
     Economics and Social Sciences
Johannes Kepler University Linz / Austria
E-Mail: Novacek@idv.uni-linz.ac.at
        Novacek@pop.idv.uni-linz.ac.at
WWW: http://www.idv.uni-linz.ac.at