ICANN/DNSO
DNSO Mailling lists archives

[registrars]


<<< Chronological Index >>>    <<< Thread Index >>>

RE: [registrars] FW: [nc-transfer] Drafting Team Status Update


On 2002-08-30 at 13:42 -0400, Ross Wm. Rader wrote:

> The draft policy generally contemplates the following;
> 
> 1. That the default rule on a transfer request from the registry to the
> losing registrar should be an "ack" in all cases unless the losing
> registrar has explicit knowledge that the registrant does not wish to
> undertake the transfer.

What about the case where a determined hijacker repeatedly puts in
transfer requests for a domain name?  The registrant would be expected to
affirm repeatedly that they disapprove each transfer.  One could argue
that in such a case the current registrar has explicit knowledge, but
that's not the kind of thing that could easily be automated.  Locking the
domain at the registry would also help in such cases, but this still
places the burden on the legitimate registrant, and that is unfair: if the
legitimate registrant messes up even once, or they have a problem with
their e-mail, or someone takes a vacation, or the contact for the domain
is naive and unsophisticated, the domain might inappropriately transfer.

Even saying that the burden rests with the requesting registrar is no
solution, since presumably a hijacker would give whatever false assurances
were requested and could move from one registrar to another, creating fake
accounts and doing all sort of other underhanded things.  In the face of
this, it really seems inappropriate to burden the legitimate registrant.

> 2. That the gaining registrar must only initiate the transfer process
> with the explicit consent of the registrant or an entity that the
> registrar reasonably believes has the authority to act on the
> registrants behalf.

This is the core of the problem: the gaining registrar has no real way to
determine this.  On the one hand, the registrar can tell the customer that
initiating a request to transfer a domain is a claim of apparent
authority, and can ask the customer to affirm such authority.  Our
procedure is to make the customer check a box on a web form which makes
this claim under penalty of perjury.  Obviously, someone could lie, but it
gives us a little more leverage in undoing an improper transfer should we
decide that our own customer wrongly requested it.

On the other hand, the majority of transfer requests are legitimate, and
putting a lot of obstacles in the way is unfair as well.

What I am particularly leery about is the possibility that two competing
claimants for apparent authority will use registrars as proxies to fight
their dispute.  If this kind of thing happens, the gaining registrar is
likely to end up one of the defendants.

> 3. (This one is perhaps the most important) That the processes employed
> by registrars to undertake these types of transactions are registrant
> friendly and do not require the implementation of bureaucratic artifice
> such as double acknowledgements, artificial barriers to portability etc.
> In other words, the processes might be complex for registrars to carry
> out, but simple for registrants to deal with - "designed for the
> consumer" in other words - simple, efficient and safe.

Where we draw the line is between those cases which can be processed
automatically and those which cannot.  For the tiny minority of cases
which cannot, our approach is to involve a real human who can apply
reasonable common sense and decision making skills.  Trying to oversimplify
this into a set of rigid rules is really impossible: the losing registrar
has, I think, a clear duty to confirm the intent of the registrant before
allowing the transfer.  We do not request a notarized affidavit and a DNA
sample, but we apply whatever methods are appropriate to resolve the
uncertainty we believe is present in a particular case.

I concede that this duty of the losing registrar is in addition to the
duty of the gaining registrar to confirm apparent authority before
initiating the request, but such duty of the losing registrar seems to
exist nonetheless.  Trying to constrain the losing registrar into refusing
a transfer only on the basis of "explicit knowledge" of the registrant's
contrary intent would introduce very serious complexities and subtleties.

-- Mike




<<< Chronological Index >>>    <<< Thread Index >>>