[discuss] shifts in IANA/accountability discussion: your thoughts?

[John Curran]

On Jun 21, 2014, at 4:00 PM, Milton L Mueller <mueller at syr.edu> wrote:

>> For instance, one could imagine an "IANA Stewardship Transition Plan" which
>> made plain the fact that the IANA registries are in fact "IETF IANA" registries,
>> with policy authority for the general-purpose registries (names, addresses)
>> having already been durably delegated to organizations which are
>> representative of the served communities (ICANN for DNS, RIRs for IP
>> address space); these registry organizations that agree to follow open and
>> transparent processes, and agree to have that confirmed periodically by well-
>> defined independent third-party multi-stakeholder review process (i.e. the
>> AoC review team model, as appropriately strengthened)
> 
> This implies a separation of the 3 aspects of IANA, not their consolidation under ICANN. Also note that within ICANN, the DNS policy development process is under the direct control of ICANN staff and board, rather than a separate entity. This is why structural separation of the IANA functions from ICANN's policy development process is necessary. Or, if you prefer, keep IANA in ICANN but separate out the DNS policy making process. 

Clarity of the various roles is very important, but that doesn't necessarily
mean that those roles cannot be performed by the same organization. For example,
ICANN performs administration of all of IANA registries per the ICANN/IETF MOU
(RFC 2860); this is distinct from the policy development roles for those registries 
(which is handled by IETF for the technical protocol parameters, ICANN for the 
general-purpose DNS registry policy, and the RIRs for the general-purpose IP 
address space policy)   It is not apparent (at least to me) whether a hypothetical 
ICANN or RIR failure in their policy development capacity would (or should) 
necessary result in any change to ICANN's IANA registry administration role.  
However, a failure of ICANN to follow the appropriate registry policy (whether 
developed by IETF, ICANN's DNS community, or the RIR community) would definitely 
represent a significant concern, one potentially with the consequences noted in
section 2 & 4 of ICANN/IETF MOU (RFC 2850)

>> It would leverage some existing but strengthened accountability review
>> processes, and would leave the IAB/IETF at a safe and healthy distance from
>> such organizations, except in the rare case of an actual organizational failure
>> so significant that the independent review resulted in a finding that one of
>> these organizations was no longer representative of the served community;
> 
> Would this finding result in a legally binding requirement of redelegation? If not, it isn't a real accountability mechanism.

It should - as noted above, we have a fairly clear agreement with respect to IANA
registry administration/operations; there is a lack of similar agreements covering 
the delegation of _policy development authority_ when it comes to the general-purpose 
IANA registries.  This has been addressed somewhat with respect to IP address space 
via RFC 7020 ("The Internet Numbers Registry System") and RFC 7249 ("Internet Numbers 
Registries") but those are not the same as contracts (i.e. with breach, notice, cure, 
arbitration provisions) and does not speak at all to the delegation of policy authority 
for the general-purpose portion of the DNS space.

FYI,
/John

Disclaimer: my views alone.