Re: [DNSOP] [[e-mail]: I-D Action: draft-grothoff-iesg-special-use-p2p-names-00.txt]

[David Conrad <[e-mail]>]

Ted,

On Dec 10, 2013, at 12:26 PM, Ted Lemon <[e-mail]> wrote:
> (1) defining sTLDs produces a small (relatively) amount of useless traffic at the root

The amount of useless traffic is unknown/unknowable. All we have by way of example is ".local" which accounts for something like 5-10% of the traffic to the root servers steady state (based on stats available from "L" and extrapolating). I don't think it would have been possible to estimate that number prior to Apple deciding to deploy technology that uses ".local".

However, with that said, it is very likely for the traffic any sTLD or even combination of sTLDs generates to be far less than the root servers must be able to handle to deal with DoS attacks. Probably.

> (2) defining sTLDs may have trademark implications that the IETF is not competent to address

I think a better way of describing the concern is that 6761 preempts all the efforts ICANN has undertaken to protect intellectual property right holders and thus, puts the IETF/IESG in the crosshairs of IPR lawsuits. This (probably) is not directly relevant to the draft in question (although I suppose it is possible that the folks behind gnu.com could want to get a ".brand" TLD).

> (3) supporting sTLDs in stub resolvers requires changes to stub resolvers

Or, more relevantly, lack of support in stub resolvers most likely implies applications/systems using the DNS protocols to do the lookup, implying (a) noise sent to various parts of the infrastructure (resolvers, root servers, etc) and (b) confusion on the part of end users when the resolution fails on one system but works on another.

That is, pretty much the exact same arguments people used in the past to argue against alternative roots.

> (4) it would be nice to have stable specifications for the proposed sTLDs

Sure, but I imagine a sufficiently large deployed base could be a substitute.  Which leads to:

(5) it would be nice to have more objective metrics by which one can measure the potential impact of the sTLD, e.g., it probably does not make sense to reserve a sTLD for a protocol with exactly one user whereas it probably does make sense for a protocol with 1,000,000+ users.

With respect to the draft in question:

(6) bundling the various sTLD requests into a single draft makes discussing the pros/cons of a specific sTLDs more complicated.

> Please do not respond to this message by re-asserting your strong opinion.   Please do not respond to this message by asserting that there's a better way than what's been proposed in the draft we are discussing, in your opinion.

I hope I met your requirements.

Regards,
-drc

P.S. Shouldn't it be 'TLPDs' (using the terminology from 6761: "top-level pseudo-domains")? To me, "sTLDs" is confusingly similar to "sponsored TLDs", but that might be because of my ICANN infection.