[CWG-Stewardship] A few additional comments for … Two additional webinars on 6-7 May

[Martin Boyle]

In UK English, instructions tell someone what to do.  So alternative wording is needed, if there is a difference between US and UK usage.

Currently the ICANN Board plays a role of (what I would see as) due diligence for its operational team.  It is not included in the SOW, but, on a critical function where its reputation is on the line, I could imagine it would want to check that staff had done their job and there would be no come-back on ICANN.  But I do not see it as an instruction, so different wording is needed.

As I said last night, I do not understand why, in introducing legal separation, we would split the role between subsidiary and parent company.  I’ve not heard any clear reasoning as to why it would be appropriate and I would see a clear issue of accountability associated with this split function.  It is not a failing of the PTI if the ICANN Board screws up.  Moving the PTI to a new operator doesn’t solve the problem.  People can only be held accountable for things they can do something about.

On separability, I said in an e-mail early in this thread that “Our objective in oversight of the operation of the IANA functions should be to ensure as near a self-healing process as we can get…  If … we end up with a failing IANA functions operator…, we need to be able to find a solution and implement it quickly.”  By this I mean, we should try to make the system work (and take some time to do this to avoid disruption), but if it is failing and won’t reform we need to respond and separate quickly.

For reasons that we’ve now discussed at length, I feel nervous about turning over the knowledge of ccTLD issues without first taking the obvious action of seeking to remedy defects.  I am sure my threshold is higher than yours, but from a TLD point, stability is important.  However, once a consensus decision has been made, I oppose making separation inordinately difficult.

Hope this helps clarify my position.

I have a question back to you on gTLDs:  doesn’t the ICANN Board have a different (operational) role given that gTLDs are in a contractual relationship with ICANN?  No contract, no IANA entry?  So it would be quite appropriate for the ICANN Board to instruct the PTI to act, wouldn’t it?  And the PTI might (as I believe is currently the case) expect to see a similar checklist of meeting terms and conditions.  In my mind, accountability for a decision badly made would be with the ICANN Board (especially if it lied in the checklist).  The counterpart to that is that the PTI Board could refuse an instruction on the grounds that it was not sufficiently documented.


MB



From: Milton L Mueller [mailto:mueller at syr.edu]
Sent: 03 May 2015 22:33
To: Martin Boyle; 'Gomes, Chuck'
Cc: 'cwg-stewardship at icann.org'
Subject: RE: [CWG-Stewardship] A few additional comments for … Two additional webinars on 6-7 May

OK. I think I understand better now.

What set you off, apparently, was the word “instructions” which you mistakenly interpreted to mean: the ICANN board unilaterally makes delegation decisions for ccTLDs. This is not what I meant at all, and I am pretty sure it’s not what Chuck is looking for either.

I note that you had to admit that the ICANN board _does_ play a role in  “ensur[ing] that due process has been followed and documented.” Some of us would call that an approval role, something akin to the NTIA’s authorization role. At the end of that process, I assume, it issues some kind of decision, approval, or statement, which may be no more than “go ahead and do what you proposed to do.” It would be perfectly reasonable to use the word “instruction” for that, but if the word scares you I’ll use another. “Due process check”? “approval”?

(As an aside, some of us fervently wish that the board had the same limited role with respect to GNSO policies.)
I understand better now your point about how the institutional environment around ICANN had to be “educated” to accept a method of making delegation decisions with a very limited role for the ICANN board (a view with which I am in violent agreement). And I think we are all sensitive to the need to do this in ways that avoid allowing governments to interfere with the ccTLD associated with another government. You have admitted that ICANN (not the board, but the community in the broader sense) “provides the framework for policy development.” But you have clarified the way in which the IANA staff _apply_ that framework, subject to a board-level “traffic light metric,” and it is clear that this is more than clerical. I have no problem with PTI continuing in that role. I am not sure why these very subtle distinctions would create such indignation and fear on your part; I think it’s constructive for these things to be teased out more carefully and taken into consideration in designing PTI.

I don’t agree with you, however, that this means that separability should be discouraged or made inordinately difficult. And if you’ll recall, that issue (ease of changing operators) was what initially provoked the controversy.

The fact that ICANN and the community around cc’s has reached an acceptable and somewhat intricate balance around how to do delegations does not mean that the right to change IANA function operator should be mitigated or made extremely difficult. This is a non-sequitur.

To begin with, the knowledge is there now and is accepted politically. While it may have been difficult to establish the right way to do redelegations over the course of the past 15 years, those methods are accepted and applied broadly now. Hence, it would not be that difficult to recreate them in a new operator and formalize them in a SLA. Indeed, it might be a good idea to formalize the current procedure more, in the form of a contract or SLA with PTI, precisely so that maintenance of the methods that are so important to you is not completely reliant on the institutional memory of a single entity (the current IANA department). Perhaps the ccNSO should have its own SLA with PTI, just as the address and protocols do. Making these things more explicit would reduce the community’s reliance on a single operator. Second, no one would want to change IFOs unless something were going seriously wrong. If things are going wrong, it’s a bad idea to make it highly difficult to effectuate a change. Indeed, one of the things that might go wrong is that the IFO might start deviating from what you and other ccTLD operators consider proper procedure.

To summarize, I don’t think your view of the role of PTI is inconsistent with mine; I do think that your conservatism regarding a change in the IFO is unwarranted by any logic or argument you’ve made, and could easily backfire on you/other ccTLDs.

--MM


From: Martin Boyle [mailto:Martin.Boyle at nominet.org.uk]
For ccTLDs, ICANN’s sole role is to provide the framework for policy development.  (Caveat:  as Paul Kane would note, actually only for those ccTLDs that were members of the ccNSO.  And actually the policy document did not come from ICANN anyway.)  It is the IANA functions operator’s obligation to base its decisions on the policy in place – as identified below.

In making its decisions:


•         Neither the ccNSO nor the GAC has any say over the decisions except in that they might (I suppose – to my knowledge this has never happened) raise questions with the Board about conformity to policy).  In particular, I would hope that GAC members (in line with paragraph 63 of the Tunis Agenda) would not see it as their role to comment – it would be interfering with a country’s national sovereignty.


•         The current Board view (and one that I feel greatly comfortable with) is that its role is simply to ensure that due process has been followed and documented.  It uses a traffic-light metric and recognises that there might be different interpretations (for example on how multi-stakeholder engagement takes place).  In essence and for national sovereignty reasons, any role for the Board is simply to assure itself that the decision has been properly made.  It has taken us a long time to get there, and I do not want to see us return to ICANN having a decision making role for ccTLD delegations:  been there and it doesn’t work.



•         All the analysis, document checking, evaluation, discussion with the interested parties in a ccTLD amendment is currently carried out by the IANA functions operator.  I failed to notice where we decided that the PTI as new IANA functions operator would only do the clerical and technical bit and I would love to know who does the detailed work in the new model.  It seems to leave behind in ICANN a key IANA functions operator role in the existing model.


You can claim indignation all you like, but you still do not address my concerns about the appropriateness of your model of ICANN Board instructions for ccTLDs.  I stand to be corrected, but I’m not convinced yet.

As for your interpretation of the new model:  if this is right, then Nominet for one will not be able to accept the current proposal.  I suspect other ccTLDs will have similar problems.  Incidentally, I seriously wonder why we have put all this effort in on separability if the only thing we are able to separate is the mechanical bit (which could all be done automatically).

The learning process has taken place on the ICANN side of the divide over the years.  ICANN is currently the IANA contractor and the learning is also in place in the IANA team.  I’ve given no thought to date about who validates changes – up to now I’ve assumed it would be the PTI Board as responsible for the IANA functions operator’s work.  Placing it elsewhere gives me two concerns:


1.       The validation body (remember that we’ve agreed there should not be an authorisation function in the new model) is the responsible agent:  if it is in ICANN we cannot hold PTI responsible for any failure.


2.       The validation body operates a gatekeeper function outside the direct surveillance of the mechanisms we have been discussing for so long.  We will have additional work to make pretty clear what the limitations of that function actually are:  I think it should only be able to refer decisions back to the IANA functions operator (the operational team) on the basis that the process has not been followed or that the necessary documentation is not present or correct.


So I’m still unconvinced by your arguments.  They might work for gTLDs – and I’m happy for existing arrangements for gTLDs to stay in place.  But I don’t think what you assert works for ccTLDs,

Martin


From: Milton L Mueller [mailto:mueller at syr.edu]
Sent: 01 May 2015 14:53
To: Martin Boyle; 'Gomes, Chuck'; 'CW Lists'; 'Grace Abuhamad'
Cc: 'cwg-stewardship at icann.org'
Subject: RE: [CWG-Stewardship] A few additional comments for … Two additional webinars on 6-7 May

From: Martin Boyle [mailto:Martin.Boyle at nominet.org.uk]
Boyle: Obviously then I’m misunderstanding something Milton.

MM: You are.

> The IANA Functions operator (IFO) should simply take ICANN’s instructions.

Boyle: Err, no!  The IANA functions operator should base its decisions on agreed policy.  In ccTLD cases, rfc1591 is now supplemented (informed) by the work of the Framework of Interpretation Working Group.

MM: which is part of the ccNSO, which is part of….ICANN. Or are you asserting it is part of IANA?

Boyle: It also has advice from the GAC 2005 Principles.

MM: GAC is part of….ICANN. Or are you asserting it is part of IANA?

Boyle: ICANN (as the Board) does not have a role in telling the IANA functions operator what to do in the case of ccTLD delegations or revocations.

MM: This is where the misunderstanding resides. When I said “ICANN” I meant the policy-making side of the ICANN/IANA split, which includes not just the board, but the policy making entities such as ccNSO, GNSO, the ACs. When I say IANA “takes instructions” there could still be ambiguities about the exact way in which those instructions are issued, and by whom, and when, but clarifying that is one of the virtues of separation. I think we both agree that once a policy is set, IFO should implement it, and that IFO does not make the policy or have discretion as to whether to implement it.

Boyle: This is fundamental – clear separation of policy and the IANA functions operator should not be undermined by ICANN giving instructions.

MM: Agreed. My goal is precisely the clear separation of policy and the IFO. This should be clear.

Boyle: The issue – and hence the complexity – is about who actually has the right to make the decisions and direct the IANA functions operator.  The role of the IANA functions operator is to identify that a decision is appropriate – and the SOW gives some guidance to that, as does the work of the FOIWG.  But it is absolutely not ICANN’s decision.

Boyle: Your failure to understand this is perhaps why you feel quite casual about re-bidding the IANA functions operation and I do not.  The ccTLD community has seen the friction caused by ICANN imposing arbitrary conditions or questioning legitimate national decisions.  (This history of this high-handed approach was one of the motivations for the 2005 GAC Principles.)  I do not think that I’m alone in wanting to avoid as much as possible having to start the learning process again with a new contractor.

MM: I think it’s clear from this exchange that you are failing to understand my point, and you are overreacting - it’s not me failing to understand the risk of arbitrary and high-handed actions by ICANN. Really, that’s a pretty insulting and uninformed assertion for you to make, given my role in analysing and leading attempts to reform the accountability problems at ICANN.

MM: As for “starting this learning process over with a new contractor,” the learning process has to take place on the ICANN side of the divide, not at the IANA contractor, and be reflected in the ICANN-IFO contract. The problems you mention have occurred precisely because IANA has been a wholly controlled department of the ICANN board. Any contract between ICANN and an IFO could and should clearly account for the concerns you have and make sure that ICANN’s board cannot issue “instructions” that are not authorized or approved by the proper parties on the policy making side of the divide.

--MM

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mm.icann.org/pipermail/cwg-stewardship/attachments/20150503/8b723e31/attachment-0001.html>