DNS Scaling issues

[Jorge Amodio]

My pleasure Mary,

I agree with you that we as NCUC need to start covering the whole spectrum of
ICANN activities and try to take advantage of the vast expertise we all can
contribute to the process.

As a related comment to the root scaling studies, you also need to take in
account that so far all the studies are focusing on the root zone and not on
the DNS as an integral system.

There are major implications that all these changes are already manifesting
on the client side. I'm in permanent contact with a great friend and
colleague also member of NCUC from the University of La Plata in Argentina
(Eduardo Suarez) who has been performing many tests on a real and
running network at the Faculty of Astronomic Sciences.

Given that DNS is not only used for simple name resolution, it's also being used
to manage email routing, consult many different spam blocking services
that use DNS as the foundation for it, and other stuff like geo location, etc,
the increased load on a local client or small server grows dramatically when
not only the server has to dedicate more memory and CPU because of the
increased number of threads handling TCP connections, with DNSSEC you
also require many more CPU cycles to digest and process the DNSSEC
responses.

As far as I know, no studies have been included yet in the ICANN process
to have a better understanding of the entire system, there some inferences
but not actual work done on the field and particularly in places where
connectivity or computing resources are limited.

Regards
Jorge

On Sun, Oct 25, 2009 at 11:03 PM, Mary Wong <MWong at piercelaw.edu> wrote:
> Jorge, thanks so much for this summary and update. It's very useful, and as
> one of the non-technie lawyers here in Seoul (and periodically but regularly
> needing to deal with technical issues on the Council) I for one appreciate
> your sharing your expertise.