In order to forecast exhaustion rates, we needed something to measure against. It would be rather naive of us to assume that allocation policy would remain the same tomorrow as it was yesterday, if APNIC received a /8 from IANA. This is where we looked at pre-prop127 delegation sizes of up to a /22. If we were to allow applicants who have received either a /23 or /24 post-prop127 to apply for resources up to a maximum holding of /22 this would last (again, under current policy) 20+ years. These of course as mentioned are dependent on 3 x /8 prefixes.

The intent of this isn't just to drop more space into the wild to be snatched up by the highest bidder, it's supposed to afford new players an opportunity to connect without having to fork out a small fortune to do so. I can only hope that people understand and see this, and instead of selfishly saying no, see what it's trying to do, who it can impact and at least understand. I definitely understand that RIR policy can change in as little as 12 months and it very well could happen that policies will change that see the exhaustion policies implemented over the last 15 years all undone for the sake of being able to get a quick /20 and for space to disappear in a few years (again) which I don't really think is the right way to go. This is a second chance to purposefully ration out a finite resource.

Christopher,

On Feb 13, 2024, at 2:15 PM, Christopher Hawker <chris@thesysadmin.au> wrote:

Let's not think about ourselves for a moment, and think about the potential positive impact that this could bring.

Let’s assume that the class E checks in all IP stacks and application code that do or can connect to the Internet are magically removed (not going to argue feasibility of this) and control of 240/4 is put into the hands of IANA to allocate to the RIRs. Subsequent steps would be:

1. RIRs, following https://www.icann.org/resources/pages/allocation-ipv4-rirs-2012-02-25-en, would request new /8s, and receive those allocations.

2. Entities[*] with pent up demand would submit requests and have those requests filled by the RIRs

3. While more /8s in 240/4 remain, go to step 1

4. Return to status quo ante.

In other words, while the IANA free pool is not (again) empty, network operators would be able to get IPv4 address space at a fraction of the market price, and then we’d go back to the way things are now.

This suggests the length of time the primary benefit (cheap IPv4 addresses) would be enjoyed depends on RIR allocation policies. ISTR a comment from you earlier suggesting that based on current consumption rates, 240/4 would fulfill needs for 50 years. However, this appears to assume that current “soft landing” (etc) policies would remain in place. Why would you assume that? I would imagine there would be non-trivial pressure from the RIR memberships to return to the pre-runout policy regime which was burning through multiple /8s in months. In particular, I’d think the large scale buyers of address space (as well as IP market speculators) who tend to be the most active in RIR policy forums would jump at the opportunity to get “huge tracts of land” at bargain basement prices again.

This doesn’t seem all that positive to me, particularly because it’s temporary since the underlying problem (limited resource, unlimited demand) cannot be addressed. What positive impact do you predict?

Thanks,

-drc

* I’ve purposefully ignored the geopolitical aspect of this here. In reality, I suspect there would be pressure for ‘entities’ to include countries, etc.