Contrast Beckstrom's interview with DelBianco's Argument Against New TLDs

[Jorge Amodio]

Good observations Alex. IMHO, I believe that there are confusing statements
regarding the implementation of new gTLDs given the fact that nobody can
provide a concrete and concise answer to some questions.

There are still many open issues and a lot of pressure from folks promoting
many of the new gTLDs since the window of opportunity and speculation
about the potential ROI may be closing, and some folks eager to put money
may be rethinking if its worth the investment or not.

About competition and the real need for new gTLDs there are valid arguments
from both sides, but they are not 100% convincing.

Many of the studies associated with this program should have been done
before we started writing the DAG.

>From the technical side, besides taking a closer look at the current
implementation of the root zone, how the changes being introduced to the
DNS (such as DNSSEC, IPv6, IDN) will impact the overall system and
not just the root zone or primary servers.

In my opinion (I'm still reviewing and analyzing them) both scalability
studies of the root zone, while well done, are not 100% conclusive but
they (including the L-root server scalability test bed) all call for a more
quantitative analysis of the entire system.

During the "proof of concept" phase when new TLDs where introduced
we missed the opportunity to establish a common set of metrics and
start collecting this data as a baseline before the introduction of the
TLDs. All reports about expanding the root zone call again for a
system to collect this data and review the effects of the introduction
of any changes to the DNS.

Scaling hardware and software on the server side is not a big issue
(the L-root server exercise has very good numbers regarding
performance impact and among other things server memory
requirements), scaling processes and people its more complex.

Connectivity may become something to keep an eye on, due the
increased sizes of query responses, DNS clients where latency is
high and/or the path can't accommodate the new size of the UDP
datagrams carrying the response, will revert (after a certain delay)
to TCP, this has some side effects like adding the TCP handshake
for each single transaction, increased time for a name to be
resolved, extra traffic and increased memory usage both on the
client and on the server side.

Then, as the AoC says, "if and when" new gTLDs are introduced
we need to keep an eye on the performance and behavior of the
entire system, but also review if the the introduction of a new
gTLD has promoted competence, consumer choice, etc, as
stated on 9.3 of the AoC.

Other open issues, as you know the IRT has been dumped and
reaching consensus on that front will take a big effort and extra
time.

Personally, I'm not a big fan of the gTLD program, but I'm not
opposed to the creation of new TLDs if it's done gradually, and
without the pressure from speculators that only see how much
money they will potentially make out of vaporware.

My .02
Jorge


On Fri, Oct 2, 2009 at 3:08 AM, Alex Gakuru <gakuru at gmail.com> wrote:
> ...
> As for the generic top level domains, there's still a range of
> opinion. They are a little more complex, because there will never be a
> solution that suits everyone - we're doing it to create competition,
> and the people who don't want more competition don't like it. But we
> are moving forward with an aggressive time frame.
>
> Nevali: What's the benefit of gTLDs? Why doesn't Icann forget about
> expanding the number of top-level domains and just oversee what the
> country code administrators are doing?
>
> RB: The owners of those country codes already have the power to do
> pretty much anything they want. But it's not an answer because if you
> open up the root, you open up new gTLDs anyway. It's in our mandate to
> create new top level domains, so we're doing the right thing - but
> it's a very difficult problem.
>
> EllyD: How is Icann trying to improve the arbitration process for
> victims of domain theft?
>
> RB: We are proposing a new mechanism - uniform rapid suspension. There
> is a proposal and we're deliberating it.
>
> <http://www.guardian.co.uk/technology/blog/2009/oct/02/rod-beckstrom-questions>
>
> and
>
> 2. 'In Congress, A Confusing Argument Against New TLDs'
>
> <http://www.circleid.com/posts/in_congress_a_confusing_argument_against_new_tlds/>
>