Re: [IANAxfer@apnic] Key elements of the transition of IANA stewardship
On Sep 12, 2014, at 4:03 AM, Richard Hill <rhill@hill-a.ch> wrote:
> What I was trying to say above is that, within the current IP address
> allocation scheme, the ICANN Board has the ultimate authority.
Incorrect.
The ICANN Board (per agreement with the RIRs) has the authority to ratify
global address policies that will be used by the IANA. That's not the same
as "ultimate authority" over the current IP address allocation scheme.
Note - it is quite possible for the IETF and RIRs to decide to go elsewhere
and contract for a different IANA operator for these spaces, if the present
operator (ICANN) were to become dysfunctional. I doubt either the IETF
(as the technical authority) nor the RIRs (representing those using the
addresses) could or would wish to do so alone, but together you might
actually be surprised how straightforward such a change could be if it
proved necessary.
> Correct. But the point is that the RIRs have agreed to be subject to the
> ultimate ratification by the ICANN Board.
Yes - the RIRs have agreed to allow the ICANN Board to perform an appropriate
check and balance, via a very specific process, in the ratification of developed
global address policies. I would not call it "ultimate"; in fact, the ICANN
Board is rather constrained in its ability _not_ to ratify proposed global policy,
with a burden of delivering statement of its concerns with the proposed policy,
including in particular an explanation of the significant viewpoints that were
not adequately considered during the regular RIR process. Furthermore, there is
mediation if the ICANN Board should reject a global policy proposal after it is
submitted for a second time.
> What I meant to say is that, in the current system, absent the IANA
> functions contract with NTIA, the ultimate authority, within the system,
> would be the ICANN Board.
The ICANN Board's authority with respect to number resources is specified by
an agreement with the IETF and agreements with the RIRs.
> And of course ISPs etc. could in theory ignore that and use other addresses
> or establish another mechanism.
Not just the ISPs... The RIRs have always been prepared since ICANN's formation
to handle the central registry operations in case of ICANN's untimely demise;
this requirement even affected the structuring of the agreements (ICANN, as
you might remember, was just a little wobbly its early years...) The tasks
performed by the IANA in maintenance of the Internet number registries is
actually rather mundane.
> Yes, but as I recall the events, there wasn't really much choice, especially
> not after the White Hose allegedly threatned legal action against an alleged
> plan to move the root server.
You are incorrect; both the RIRs and the IETF had full choice in whether to
participate or go a different path, as both communities were already well-
organized... It was seen that RIR (via ASO) and IETF (via PSO) participation
would bring stability to the situation, but either community could have gone
its own route if that was the consensus (and one can argue that the IETF did
with respect to policy coordination when the PSO was later phased out...)
> Correct. And that is why I think that an equivalent provision has to be
> added somewhere, in the ICANN Bylaws and in a new MoU between the NRO and
> ICANN.
Indeterminate; presumes first that the IETF's existing RFC 2860 isn't already
sufficient for the purpose, and second that there is some need to constrain
ICANN further than provided for in the existing ASO/NRO agreements. It is a
question worth discussing, but so far there does not seem to be a compelling
argument either way.
> Otherwise, when the IANA functions contract lapses, there would be no
> written clause anywhere that says that the ICANN Board cannot make material
> changes in the policies.
The existing global policy development process (which cannot be changed
except by agreement with the RIRs) only allows ICANN Board the ability
to _request_ changes to the proposed policies (per MOU and attachment A)
<http://archive.icann.org/en/aso/aso-mou-29oct04.htm>
<http://archive.icann.org/en/aso/aso-mou-attachmentA-29oct04.htm>
Richard - please take an opportunity to review these agreements when making
statements such as "there would be no written clause anywhere that says that
the ICANN Board cannot make material changes in the policies."; I believe
doing so would save us all some significant time (and the accuracy of your
responses to both myself and David might benefit substantially as a result.)
>> I can't actually imagine any of the involved organizations (NTIA, ICANN,
>> IETF, or the RIRs) taking a position to the contrary simply to align with
>> your theory (i.e. that ICANN is the "authority" for Internet
>> number resource
>> policy) but I would welcome you to directly inquire with any of
>> them if you
>> have any doubt.
>
> I thought that the purpose of these discussions, and of this mailing list,
> was precisely to explore this sort of question.
Indeed! I was offering you some additional options for your exploration.
> ...
>> Upon further review, you'll see that "ICANN" does not appear at all in
>> the NTIA statement of IP numbering principles; this shouldn't be all
>> that surprising because ICANN's very important role in providing overall
>> coordination for the Internet's identifier systems does not equate with
>> having "authority" over Internet number resource policy.
>
> So, in your view, the RIRs' agreeing that it is the ICANN Board that
> ratifies their policies does not have any legal meaning or effect. You are
> of course entitled to your views, as is everybody, but I don't agree with
> that view.
Interestingly enough, I actually have to worry full-time about questions such
as the above, in insuring that ARIN is able to perform its mission, and it is
my view that the present agreement between the RIRs and ICANN is sufficient
even absent the NTIA IANA functions contract. That said, it is still a change
that is worthy of community consideration, since NTIA does today have extant
control over the IANA function operator and that would change. You appear to
be concerned that such a change will enable ICANN to exercise some form of
latent "ultimate authority", but have yet to point out where such authority
will originate from...
I have also clearly stated that reviewing the existing agreements is quite
worthwhile at this point in time, as there may be some additional clarity
desired. I frankly would be hesitant to put a statement in any agreement
that says "the ICANN Board cannot make material changes in the policies"
(nor a statement that reads "Richard Hill cannot make material changes in
the policies") out of concern that someone might later misinterpret these
as the sole constraints governing the situation. I know risks of such are
rather remote, but your recent conjectures have definitely opened my eyes
in this respect.
Thanks!
/John
John Curran
President and CEO
ARIN