On Jul 17, 2008, at 7:39 AM, Geoff Huston wrote:
Jonny Martin wrote:Smoothing this uneven distribution requires that this last /8 is held in reserve with vastly different allocation policies to that in place now.Such as assertion prompts the obvious question: Why is this _required_?If the current address distribution policy framework is "unfair" then why wait until the current allocation framework is effectively exhausted before attempting to change it? And given that the current allocation framework allocates addresses to any applicant on the basis of demonstrated need, then its difficult to understand precisely what "unfairness" means within the current policy framework.Changing the allocation policies of this "last" address block is something that deserves some thought. It seems that the proposals all advocate some form of rationing of the resource among the applicants.Beneficiaries of any last block rationing scheme obviously may regard the outcome as "fair" (although if they get allocated less than they believe they need I'm not sure that even they would agree that it was "fair"), while those who are not beneficiaries would obviously feel that such measures are "unfair". So this does not really look like an exercise in "fairness" does it? This appears to me to be more like an exercise in implementing particular industry policies that make certain distinctions between industry actors.
Hi Geoff,You are of course correct; this proposal envisions a kind of industry policy, just as the status quo envisions a kind of industry policy. So your assertion of that fact doesn't really amount to much of a counterargument, unless you want to claim that the exhaustion of the unallocated IPv4 pool is not really a significant change.
Now, in general, I'd make the observation that its not easy to engineer particular industry outcomes using only address allocation policies as the mechanism to achieve them.
Engineering specific outcomes is not a goal. IP address policy can only affect IP address distribution outcomes. IP address distribution outcomes *do* matter, I think you would agree... Isn't that why there are five RIRs today rather than one global source?
If your concern is barriers to entry in this market and the issues related to economies of scale, the discipline of competition and the impetus of innovation, it seems like a tough ask that you use address distribution policies as the means of implementation of industry policy. There is
This is a red herring. To repeat, IP address policy can only affect IP address distribution outcomes. There seems to be a pretty well established global consensus that terms of access to "public", globally useful IP address resources can influence outcomes, even if they do not *determine* them exclusively. I'd even venture that that influence is why people care about IP address policies.
also a considerable risk that there may be a collection of unanticipated outcomes. I would not normally advocate the wisdom of putting the address community in the hot seat of determination of mainstream telecommunication policies, particularly so given that a more conventional approach is to use public policies and the associated legislative, regulatory and taxation framework at national levels to achieve desired industry and social outcomes, if thats the underlying agenda here.
Another red herring. Most previous and current existing *regional* address policies do not have any special or direct effect or conflict on national policy prerogatives or industry outcomes; there's no reason to assume that a few more/similar changes in regional policies would be any different.
For what it's worth, I support this proposal :-) TV