Hello
Has anyone ever attempted to create a geographical database of networks?
In other words, zone the world into pertinent, well-known blocks and do a
network->zone key->value pair.
--matt hempel
Has anyone ever attempted to create a geographical database of networks?
In other words, zone the world into pertinent, well-known blocks and do a
network->zone key->value pair.
i just finished turning most of the us's zipcodes into lat/long pairs.
but that's not quite the same, is it?
Two words: IP Mobility. I always thought it was more marketing than
content but that doesn't stop some applications from being developed to
use it.
My guess would be that you'd have more trouble getting people to agree on
how the 'carve' would happen than the implementation. Then of course
there are the hackers who would love to mess with it.
To answer your question, all I ever heard about this was (the rumor of) a
draft RFC. I couldn't find it though so wasn't able to confirm. (That
was about 5 months ago).
Jack
I'll read into that ...
My idea is that if you know the geographic source of a route, you can make
cold-potato routing decisions while still receiving full advertisements
from your peers at each peering point. Every advertised route could be
matched against a ruleset which attaches preference to it based on
distance.
--matt
Many groups have done this to various levels of granularity. Generally,
this database is proprietary although I do recall one fellow who was
selling his database at ISPCON in 1997. Uses include, maintaining
compliance with US laws regarding export of encryption software and
analyzing Squid logs in Australia to determine where an additional
expensive trans-oceanic circuit might be more cost effective than paying
Telstra 19 cents per megabyte for incoming traffic.
Since this database will be changing all the time, I would suggest that
you either build the tools to do this yourself so that you can maintain it
up to date, or find someone selling such data who will provide regular
updates.