[AFRINIC-rpd] New Policy Proposal: Inter RIR IPv4 Address Transfers (AFPUB-2013-V4-001-DRAFT-01)

[David Conrad]

Nii,

On Jan 14, 2013, at 2:49 PM, Nii Narku Quaynor <quaynor at ghana.com> wrote:
> On Jan 14, 2013, at 20:52, David Conrad <drc at virtualized.org> wrote:
>> I also think that an extreme imbalance such as AfriNIC having multiple /8s when all the other RIRs free pools are exhausted is likely to lead to even more unpleasant politics and fracturing of relationships, but perhaps I'm too pessimistic.
> How come Africa having the smallest overall allocation % of v4, other numbers, not the concern instead?

I'm not sure % of IPv4 allocations is (or should be) an area of concern, rather it is the relative lack of Internet deployment that should be the issue. I believe the fact that the AfriNIC region has the smallest overall percent of the IPv4 address space amongst all the RIRs is a symptom, not the problem.

My impression is that the reason for the smallest % of IPv4 allocations being in the AfriNIC region is that it is a function of the relative lack of infrastructure and/or regulatory policies of some African countries that have limited and/or delayed Internet deployment. That is, the African region has fewer addresses because the addresses weren't requested and they weren't requested because AfriNIC's pools of addresses were sufficient to meet the artificially limited demand. While I can't speak to AfriNIC policies, I know that when I was at IANA, if AfriNIC requested additional address space, it was allocated within a day or two (just like the other RIRs). I doubt that policy changed, at least until the IANA free pool was exhausted.

However, we're now at a point where the AfriNIC free pool of IPv4 addresses is larger than the other RIRs with consumption projections (however flawed) that suggest AfriNIC will continue to have IPv4 addresses long after the other regions have consumed their IPv4 free pools. It isn't clear to me that having a large IPv4 free pool is a good thing as it means those addresses are not being used to provide Internet connectivity.

If the African region IPv4 consumption rate increases, it implies accelerated deployment of a dead end technology which may be good but only (IMHO) if it is part of a dual-stack deployment strategy. If the IPv4 consumption rate stays where it is or decreases and AfriNIC does not address its free pool, I perhaps pessimistically predict (more) political unpleasantness while an increasingly valuable resource is locked in a warehouse, slowly spoiling. The reason I crawled out from under my rock to comment on this thread is because I personally would like to avoid that latter outcome.

To borrow Maina Noah's analogy, I believe the question should be "what is the African region going to do with the leftovers knowing that over time, they're going to spoil and (possibly) that the knowledge that the leftovers exist is going to give folks a false sense of security that they don't need to go get new food?"

A long time ago at a different RIR, I proposed a policy that (in essence) required in addition to meeting increasingly stringent IPv4 justification requirements, requesters demonstrate their IPv6 deployment plans in order to obtain additional _IPv4_ address space. The proposal was shot down for various reasons but I think the concept remains valid: if the goal is increasing Internet deployment, the focus needs to shift from IPv4 to IPv6 and stretching out the IPv4 free pool isn't helping anyone.

Regards,
-drc