Hi Greg, Thanks for taking the time to lay out this model carefully. I think it's good to expose it, but I think the exposure makes it clear it isn't worth pursuing in greater depth. More below. On Tue, Apr 14, 2015 at 01:41:41AM -0400, Greg Shatan wrote:
In this model, ICANN would still be the source of the right to perform the IANA Functions, as in the current internal model. However, ICANN would enter into an irrevocable agreement with the Affiliate for the IANA Functions.
The difficulty I see with the proposal lies in those two statements. My view is that ICANN is not the "source of the right" to perform the functions. The source of the right to perform the protocol parameters function is the policy community for those parameters (i.e. the IETF), and the source of the right to perform the number resources function is the policy community for those parameters (i.e. the RIRs). The basic problem with this model (and this interpretation of your description) is that it requires a fundamental assumption that the functions are welded together, and that there is a single body responsible for "stewardship" of them. This is something that the other communities simply don't believe. I therefore don't think there would be value in further elaboration of study of this model, because anything that proposes a single source of stewardship for all the functions won't fly. If, on the other hand, the model is supposed to be one in which the affiliate gets the right to contract for an entity to act as the naming functions operator only, then it isn't clear to me what work this wheel is supposed to do. In order for it to be an effective steward of only those functions, it basically needs somehow to perform the job of the community for names functions. But we already have a policy body for names functions, however imperfect: ICANN, or perhaps some subset of it. So, to perform this stewardship function for names, the affiliate would have to instantiate most (or all) of the accountability measures that ICANN needs anyway. This seems like an excellent way to re-open a number of prior decisions. Moreover, it strikes me that it's quite likely to deliver subtle differences with the existing mechanisms, which would offer a tempting opportunity to game the system, a potential for procedural deadlock, or both. Therefore, no matter which way I interpret this model, it seems to me that it can't possibly yield enough benefit to be worth exploring in greater detail. Best regards, A -- Andrew Sullivan ajs@anvilwalrusden.com