Re: DNS: Draft selection criteria for new DNAs and 2LDs

Re: DNS: Draft selection criteria for new DNAs and 2LDs

From: Simon Hackett <simon§internode.com.au>
Date: Tue, 08 Jul 1997 22:48:27 +0900
>
>Please feel free to comment copiously. Submissions for the final
>document will close at the endof July.

Great to see, Kevin. Here are my comments/constructive criticism - what do
you think? 

I'll preface this by strongly suggesting that you review the draft DNA
selection criteria at http://www.gtld-mou.org - they've thought this over
for longer than we have. 

I'm playing devil's advocate in what follows. Hope you can think of some
answers to some of these, cos they are (in some cases) quite non-trivial :)

>Selection criteria for DNAs (Minimum requirements for DNA status):
>
>1. Must be an incorporated body,

Sounds fine. 

>
>2. Must have at least 5 full time employees working for the company or
>intending to work for the company with at least 1 full time employee
>to devote to DN applications,
>

Why those sizings? i.e. what lead to "5" and "1" - just seemed like good
numbers at the time? What if a new company fires itself up and plans to
hire people as needed, and doesn't have 5 initially - do they have to hire
5 people, then apply, then sack them if their application is un-successful? 

If there is a need for at least one full time "DN" person, what are the
other 4 people for - window dressing? 

Could a potential DNA simply nominate a company that they would contract
staff in from if successful? 

Personally, I'd suggest that this clause is un-necessary. If you want to
define "serious" in some useful way, define it with the application fee and
the size of the network link (2mb/sec min), and forget about associating
company size with capacity to operate as a DNA - those who use creative
automation will need a lot less staff than those who decide to make paper
mountains.

>3. A charge of $5000 to get setup - refundable minus any costs 
>incurred in researching application if unsuccessful,
>

probably fair - high enough to deter silly applications, but not so high as
to deter genuine competition. What, exactly, happens to the $5000 - does it
just float into the bank account of ADNA as general revenue? Does any of it
go towards software development for the benefit of all DNA's, or is this
just money towards ADNA's legal bills in practical terms? 

>4. Must have permanent connection to the Internet of at least 64K.
>This connection should be maintained to be unsaturated at all times,
>

Where is the 64K determined from, i.e.:

	- will the DNA be required to run a primary and multiple secondary DNS's
that are listed as official secondaries for the domains they operate in (in
which case I submit that 64K isn't enough - for com.au at least) OR
	- will the DNA be required to just hand synchronization/database change
requests to some other entity running the public, official primary and
secondary servers for each relevant domain (in which case 64K would be
enough, because only the database deltas would be transmitted over the
link, no user requests for DNS name lookups).

Either way, I personally feel that number ought to be 2 megabits per
second;  The $5000 application fee is presumably an idiot-deterrant. Why
not specify a link size which also is a deterrant to non-serious entrants
who might not understand the potential traffic loads of operating an
official primary/secondary for com.au. 

>5. Must be covered by at least $500,000 professional indemnity 
>insurance,
>

Sounds ok.

>6. Must submit business plan for company as a whole,
>

Where the DNA is a subsidiary of a larger organization, or a business unit
of a larger organization, what is the relevance of requiring the business
plan for the company as a whole?

For instance, let's take an operative example - Melbourne IT; What is the
acceptance committee (?) of ADNA going to do with the business plan for
Melbourne IT's other business interests?

more specifically, this is useless unless you plan to use it as part of a
pass/fail determination for accepting a company to become a DNA; So what,
exactly, would the criteria be for a business plan in order to have it
"pass" - e.g. is it:

	- must be longer than xx pages
	- must not be a company that also does... (what?)
	- must be a company that also does... (what?)
	- must be a mate of (please specify)

or whatever? i.e. why bother to ask for this at all - isn't the business
plan implicit ("operate a commercial sustainable business providing
services as a DNA in nominated namespaces") - what else will you require,
and exactly why do you need to know (or care)?

>7. For each 2LD that they want to deal with, they should submit 
>business plan for what they intend to do with it,
>

See above - don't you expect them to want to just sell registrations in it?
Can you provide an example of what more you might expect to see from
someone's business plan, and why you'd care? What, specifically, is the aim
here? 

>8. Must submit contingency plan for support of its domains if 
>business fails or decides to stop being DNA,
>

Wasn't this discussed in pre-ADNA meetings - why should this even be a
choice the DNA has - why not just define some rules for inter-DNA
cooperation (i.e. DNAs agree to mutually back up all public data in domain
name spaces with multiple operators, and agree to provide continued domain
name resolution at no extra cost if a DN dies, but may elect to charge a
re-registration fee if a customer requests changes to information
originally registered by a defunct DNA). That guarantees continuity of
service to clients. Just draw it up, and make all prospective DNA's sign up
to it. Why give them the lattitude to make their own (inconsistent) plans
in this regard? make them all agree to certain things "for the common good"
or they can't play!

>9. DNAs for commercial domain names should be required to operate on a
>proper commercial basis, charging on the basis of total service costs
>plus margin for the DNA activity, to avoid predatory behaviour,
>

Define "proper commercial basis". For instance, why would you block:

	- a DNA who decided to charge $0 for services for some reason (either
they're mad, or they think they can fund it some other way) OR
	- a DNA who wants to charge $5000 per name because they feel like it. 

... and why?

Isn't the whole aim to have multiple DNA's so that commercial market forces
lead to a common/reasonable price (as happens in the ISP industry as a
whole)? ISP's aren't required to sign up to charging reasonable prices, the
market forces them to - why is this different - as long as there are
multiple DNA's in a domain, market forces act as the great leveller here. 

	What do you do about a DNA who starts up at whatever you define as
"reasonable", and then change their pricing a week later - do you revoke
their status? 

Under what conditions can you revoke the status of an operational DNA in
general?

If you don't define these things damn carefully, be prepared to be sued by
the first DNA that gets their status revoked just because they decide to
use a pricing policy that the ADNA directors might feel is a bit radical
(especially if an office bearer of ADNA happens to be another DNA in the
same namespace who is being disadvantaged by that pricing policy -
hypothetically of course! - in reality you'd never be mad enough to have
operational DNA's who are also directors of ADNA - or would you? Can you
say "conflict of interest"?

i.e. unless your criteria are really clear, and non-emotional and
non-subjective, you'll take a lot of flak the first time you try to bounce
someone who isn't playing by your 'rules'. You cannot afford to allow any
potential for DNA rejection (OR initial rejection of the application of a
prospective DNA) on anything other than failure to meet specific, clean,
criteria. "I don't like the sound of this other division of their company"
just isn't going to cut it.

>10. All DNAs operating within an SLD are subject to the same policy as
>specified by the original applicant,
>

makes sense. What does that mean, specifically, in com.au, however, i.e.:

Does Melbourne IT have to reapply to be in com.au (surely the answer is
yes, as ADNA would presumably replace the existing delegation of authority
by Robert Elz to Melbourne IT - otherwise Melbourne IT could operate
"outside of the rules" if it suited them with respect to other entrants
into the com.au space.

>11. DNA licences should be reviewed and renewed annually.
>

The review criteria (in particular, the criteria for revoking the status of
a DNA) need to be specified, really clearly, to remove the potential for
other DNA"s who might be members of ADNA to be subject to potential
conflicts of interest when reviewing competitor DNA's. 


>----------------------
>
>Selection criteria for new 2LDs
>
>1. Nomination must come from DNA
>

ok I guess, but why can't anyone else propose a new 2LD and see if anyone
of the existing DNA"s wants to offer to handle it? 

>2. Name should be no less than 2 letters + .au
>

Why? Is the reason for stopping someone proposing "1.au" (for instance)
just because you think it'd be too much fun? Or what? [just curious, I
don't personally care either way]

>3. What is domain of new 2LD (eg. commercial enterprises for .com)
>

Can we, please, please, pretty please, start saying com.AU, not ".com".
That loose phrasing will come back to bite you. 

>4. What restrictions will apply to applicants for domains (eg. has to
>be significant part of registered business name for .com)
>

See above; Also I'd suggest this is encompassed in the general requirement
for a prospective DNA to indicate their naming policy in general - but
under what conditions can they change it in the future - e.g. in the com.au
case, what happens when Melbourne IT and (say) 4 other operators are
running in com.au and one of them wants to change the policy statement -
can any of them do it ? Who arbitrates? ADNA? 

>5. Justification for need for new 2LD (ie. Prove that the added 
>complexity of a new 2LD is justified by the appeal/demand for new
>domain and inapplicability of existing domains. Committee should be
>relatively flexible about this and let most things through unless they
>are frivolous or offensive. Let market forces determine demand for new
>2LD.)
>

It's not so much compexity as confusion from additional domains. There is
certainly zero technical reasons to go beyond com.au - until com.au is at
least 10 times larger than the US ".COM" domain is now, there is no
rational technical argument for adding more commercial domains. So we come
down to people wanting a whole new spot in which to argue over their name.
How, exactly, does a prospective DNA prove demand for something they can't
necessary supply? Do they guess? Do they lie? Do they put out a petition
(where?) for (who?) to agree that it's needed? 

Can a DNA just propose the creation of all 7 of the new domain names
proposed by the IAHC, under .AU? 

>6. Must allow 60 day public notice period for comment before final
>approval by board. In 60 day period objections to the new 2LD should
>be received and considered by ADNA.
>

Sounds ok to me. Clearly not applicable for com.au, where there is no
reason why new DNA"s can't enter the market 24 hours after they get the
go-ahead on their application. 

>7. DNA that proposed the 2LD to be required to guarantee to 
>maintain processing of the domain for at least 2 years, and then
>continuing until there are at least 2 other DNA's operating in that
>domain, and one of those has agreed to continue handling the domain
>until they can pass it on to someone else, always keeping at least 2
>DNA's (other than immediately after the domain starts, when there can
>be just one, or if the last competitor gives up, leaving only the
>original DNA or someone they transferred the responsibility to).
>

I don't think that's entirely fair. Insisting on a minimum of two years'
operation sounds great. Insisting that they operate forever if noone else
happens to want to also "do" that domain name space is silly, however -
what's wrong with saying "minimum 2 years" and just *leaving it alone* at
that - but why make it that DNA's "fault" if noone else wants to operate in
that namespace inside that two year period and force them to agree to exist
forever if noone else plays? All that happens if the first DNA dies or
looses interest is that this domain becomes dormant in terms of changes,
reverts to the 'owner' of .AU until someone else wants to claim/apply for it. 

Anyway, this won't happen in practice - Think it over - any entity who is
able to operate as a commercial DNA is likely to apply for *all* commercial
name spaces that ADNA manages to gain approval to say "yes" for - and will
apply for all new namespaces as they (also) become available - because if
their processes work, why shouldn't they? 

Anyway - additional things not specified are:

- time between application and return of a verdict by ADNA to the
prospective DNA

- timetable specifically for the addition of new DNA's into com.au (give
them all an application period, a closing time, a review time, and a start
time in formal operation so they all start at the same time, I'd suggest -
this includes reviewing and re-granting Melbourne IT)

- You might want to ask that prospective DNA's agree not to sue ADNA or its
officers

- You might want to specific that DNA's agree to submit to and/or
participate in a dispute resolution process if disputes occur either in
terms of inter-DNA cooperation (e.g. database updates, server operations
etc) or if their customers dispute their decisions.

- What will you do when, after 6 months, someone in com.au says "stuff it,
we'll register any name that's not already taken". Right now, you have to
leave them to do it for six months before their next 12 monthly review
comes up - and then what do you do ? You can't revoke the 10,000 cute names
registered in the meantime.

Cheers,
  Simon Hackett
---
Simon Hackett, Technical Director, Internode Systems Pty Ltd
31 York St [PO Box 284, Rundle Mall], Adelaide, SA 5000 Australia
Email: simon&#167;internode.com.au  Web: http://www.on.net
Phone: +61-8-8223-2999          Fax: +61-8-8223-1777
Received on Wed Jul 09 1997 - 00:51:56 UTC

This archive was generated by hypermail 2.3.0 : Sat Sep 09 2017 - 22:00:02 UTC