Increasing the root zone file -- costs/benefits and future technology

[George Kirikos <gkirikos@xxxxxxxxx>]

Hello,

Just to add to my prior comments, and to followup on some comments from a blog 
post of a couple of years ago regarding the size of the root zone file and 
caching:

http://blog.icann.org/2007/03/factsheet-dns-attack/


1. Isn't it true that increasing the frequency of updates, and the size of the 
root zone file would make it much harder to have benefits from caching of the 
entire zone file locally? If the entire root zone is 68KB, it might in the 
future be treated like the hints file. Stability (through a small root zone 
file, rarely changing) would be enhanced if the number of TLDs is small, in 
other words.

2. Aren't there costs that the team is not anticipating, if the zone file is 
increased to huge levels? (e.g. 68 MB, instead of 68KB today, or even smaller 
with compression) For example, if the zone file was tiny and only changing a 
couple of times per day, it is conceivable that one day satellites that 
transmit signals for the Global Positioning System (GPS) worldwide could 
transmit root zone updates through a low bandwidth (in other words, cheap!) 
part of the spectrum on a continuous basis, worldwide to all devices. Systems 
such as this would be "out-of-band" and provide an extra level of resilience to 
the entire global internet that is increasingly important to world economies.

But, increasing the size of the zone file might make such a system infeasible, 
or too costly. 

I'm sure there are other creative ideas out there that folks haven't invented 
yet, so isn't it true that, all else being equal, a parsimonious design is best 
in the long-run? In other words, your "tests" need to be robust not just for 
today's problems, but for things you've not even thought about or imagined.

Sincerely,

George Kirikos
http://www.leap.com/