[CWG-Stewardship] Initial Discussion Draft on Transition Models

[Andrew Sullivan]

Hi,

I apologise that I didn't come across as clearly as I should have.
Let me try again.

On Wed, Apr 08, 2015 at 09:38:07PM +0000, Milton L Mueller wrote:
> 
> 
> > -----Original Message-----
> > The IETF, however, is in a different boat, as it doesn't think it needs such an
> > agreement, so that is a cost of some forms of the legal separation that are
> 
> Really? It already has such an agreement, so it must think it needs one.

My point is that the IETF already has an agreement with which it is
satisfied.  It has identified a couple small changes -- as far as it
can tell entirely consistent with the spirit of its MoU and the annual
SLAs that have been negotiated -- that ought to be made as part of a
new SLA in order to tighten things up on the assumption that ICANN's
agreement with NTIA will go away.  But in effect, the IETF's position
has been that it needs no new agreement.  That is not what the RIRs
have said, so that was the difference I was trying to talk about.

> The only change would be that its IANA has a slightly different
> corporate form.

"Maybe," is my point.  Under some of the forms, IANA is still just a
department of ICANN.  ICANN could argue plausibly that it could
maintain its agreements with other bodies and then use an
administrative mechanism to separate the provision of those functions
by having IANA do it.  This is certainly the case when IANA remains a
department of ICANN with a bunch of new ICANN-internal rules around
how IANA is to be treated within ICANN the corporation.  It is
possibly the case under an arrangement in which PTI is an affiliate of
ICANN, and ICANN is the sole member of PTI.  It seems much less
plausible under the case where PTI is some more-independent
organization with a board composed of members other than just ICANN.

So it could be that an agreement between IETF and PTI would need to be
negotiated.  For the purposes of transition, it is hard for ICANN to
argue to the IETF that a significantly altered agreement between ICANN
and IETF is necessary for the transition to happen: the IETF has
available to it the argument from stability: "Here is the agreement
we're operating under, and the transition is supposed not to affect
any of the operations, so what is the reason to change the terms?"
But a completely new organization has its own changed circumstances
available as a premise in a counter-argument, which means that the
terms between the IETF and an independent PTI could well need to be
different.  For instance, perhaps PTI's funding model would depend on
a continued unity of IANA; and if that were to fail, PTI would become
unsustainable.  In such a case, PTI might not agree to the 6 month
termination that the IETF currently has, or might not agree that the
IAB has final say in dispute resolution, or whatever.

I do not know whether any such case is plausible or a real risk.  My
point is that, the further we move from "non-names communities'
arrangements remain with ICANN," the more we are changing as part of
the transition.  Each additional change incurs some amount of risk.
That risk might be worth trading against other gains we might think
we're making, but the engineer in me says that when you are changing
things you want to change as little as you can get away with in order
to encourage stability of the whole system.

Perhaps you're right in saying that that, if the IETF walks away from
the current arrangement, it's no big deal.  I think a unified IANA is
valuable, but I agree that it's not a must-have.  On the other hand,
losing one of the "three legs" of IANA as part of the transition also
seems like a fairly large change, and so one that seems to me
undesirable if we can avoid it.

> But here's the bottom line: there is no way the IETF can reasonably
> demand that the entire names community abstain from choosing its
> optimal solution because such a solution would create what is, after
> all, a very minor inconvenience for IETF.

I don't think that was what I was saying.  I was observing that adding
a new entity does not appear to offer a way to ensure that the
pre-existing arrangements will automatically carry through to new
organizations.  It's worth noting that the MoU text in RFC 2860 does
not have a successors and assigns clause.  So there is some risk in
trying to move the agreement to a new organization, particularly if
that move entails changes to the agreement or if one of the parties
thinks that it's an opportunity to try to renegotiate.  Given that
several people seem to have taken the transition itself as an
opportunity to forge new arrangements for ICANN, it doesn't seem
impossible that someone might attempt similar renegotiation of
agreements.

Best regards,

A (speaking as ever for myself)

-- 
Andrew Sullivan
ajs at anvilwalrusden.com