[rpd] [Last Call] Draft Policy Proposal - Soft Landing, Recovered Space and Priority (AFPUB-2026-IPv4-001-DRAFT02)

[Benson Muite]

Jordi,

A few translations below :)

On Sat, Aug 1, 2026, at 11:24 AM, jordi.palet--- via RPD wrote:
> Hi Noah,
>
> Unfortunately, googling I found too many definitions of what is 
> “perspective Kaka”, so not sure about your context. Anyway, my “short" 
> considerations follow.

Kaka is brother in Kiswahili.

>
> 1) The change to 8 months, was to resolve the discrepancy as in the CPM 
> you have 12 months, but then the softlanding change to 8 months. So the 
> point was the same as other changes in this proposal: to align the CPM 
> to avoid confusion to any reader. I could have done in the other way 
> around: changing the 8 months back, to 12 months, but my understanding 
> was “the community decided 8 months when working on the softlanding 
> proposal, that decision is chronologically later than the 12 months, so 
> why going back?”. And also, if anyone could have provided this input 
> when I submitted v1 of the proposal, it could have been discussed and 
> changed in v2.
>
> 2) I’ve already commented on this before. "Final and irreversible" is 
> from the standing point on who is reading the CPM. We *all* know very 
> well that any proposal can amend the CPM (in all the RIRs). The CPM is 
> a live document, we have fixed the language and CPM for a given time, 
> but we can always come back to revise it, so yes, it is actually "true 
> at once” it provides suficiente clarity to who is reading the text when 
> it is an active policy.
>
> 3) The refill of the /12 has a very important benefit for the 
> community. If we reach the point where recovered resources are in 
> quarantine, and there are pending requests, those pending requests can 
> be satisfied immediately, instead of waiting for the quarantine period. 
> There is no any cons on doing that, only benefit, despite that I 
> believe it is certainly possible that the increase on IPv6 deployment 
> will mean that we could not actually reach that point. It all depends 
> on us doing the right thing and be more active deploying IPv6.
>
> The PIER showed the problems and was presented every year. I’m not sure 
> right now (because the broken web site), when we had the last PIER, but 
> I can find a document from 2022 (after I recall a presentation from 
> December 2025). Since then, for months (or even years) nobody talked 
> about all that in the list, nobody presented any ideas. I asked the 
> staff to refresh it 
> (https://lists.afrinic.net/pipermail/rpd/2026/014716.html), and about a 
> month after, the staff created a summary 
> (https://lists.afrinic.net/pipermail/rpd/2026/014725.html), which 
> again, with the broken site, now is not available (could we restore it, 
> please?). We need to be proactive and as much up front as possible! 
> There is not excuse to not work in policy development to fix issues, 
> specially if they are clearly spelled out by the staff because they 
> find these problems in their daily activity, same as members, or 
> consultants working with members, may find the same or other issues 
> working in projects.
>
> Final comment. Yes, the proposal changes several things, but I think 
> those changes are sufficiently clear on their own that doesn’t make 
> sense to make one proposal for each individual change. If the 
> discussion from v1 had raised that, I will had not any issue to split 
> the proposal, but was not the case. Also, there was a clear problem 
> statement, spelled out by the staff, as said, was there for months. 
> Have you or anyone else even tried to discuss on the list?
>
> To be clear: don’t take me wrong. I’m not pretending to attack or blame 
> you or anyone else. It is just the way we work, or at least most of the 
> community work. I try to be as proactive as possible, considering my 
> own obligations, probably the difference is that I also believe that 
> broken protocols (in IETF) or broken policies (in RIRs), even if that 
> is not paid job, is also part of my obligation with the community, 
> despite others dislike that - I just care on the global good, not who 
> dislike me working on trying to fix problems, and yes, of course, I can 
> make mistakes, but at least I try and I try my best and as early as 
> possible so we have time for discussion, not just waiting for when the 
> time for new proposals is gone, or during the LC.
>
> You know very well, for example, in the PDP update proposals, how much 
> I tried in pvt emails to you and other co-authors, co-chairs in copy, 
> to follow the co-chairs recommendations to work together. Absolutely 
> ignored. It seems you don’t like working with others. Fine, I decided 
> not to submit my own proposal to avoid repeating the situation of 
> reaching consensus both proposals and not being implementable. Now, I’m 
> every day more convinced that I need to do that, I need to ignore 
> people that don’t like to work with others, and proceed myself, because 
> I don’t do this for me, but for the good of the community and the 
> community deserves whatever I can do to help, even if others don’t like 
> it because they may have their own personal priorities, not community 
> priorities.
>
> I think the lesson to learn is: the community must discuss the 
> proposals in the early stage, not in the LC.
>
> Regards,
> Jordi
>
> @jordipalet
>
>> El 31 jul 2026, a las 23:11, Noah <noah at neo.co.tz> escribió:
>> 
>> Salama Jordi,

Greetings Jordi,

>> 
>> Let me try to put things into perspective Kaka..
>> 
>> Basically, the question afrinic staff put to us the PDWG was whether recovered IPv4 space falls under Phase 2 or needs a new policy, plus tidying up the conflicts between Soft-Landing and other CPM sections. 
>> 
>> The parts of the proposal that actually answer that are fine by me and can go ahead putting recovered space into the available pool under Phase 2 after a 12-month quarantine, since some sections already cover that. No issue with any of that.
>> 
>> Kwaio, (shida yangu) as in my problem is with three things that came along for the ride and don't really belong to that question.
>>

Kwaio, (shida yangu) - Therefore (my problem is)
 
>> First, the change from twelve months to eight, tightens the utilisation and SAW obligations on every assignee and every LIR making sub-allocations. My brother...whatever the merits, it has nothing to do with recovered space and it is not mentioned anywhere in your problem statement. A change to member obligations of that kind should stand on its own bro and be discussed by the PDWG as such, not slipped in alongside the recovered-space fixing hahahaha..... you get it? If you dont get it, forget about it... :-)
>> 
>> Another thing bro the "final and irreversible" wording. I would point out that the proposal's own reasoning says it "doesn't prevent a better choice over the time." Both cannot be true at once. If the intention is simply to give us a predictable default for today and I would support that then the irreversibility language is doing more than the purpose needs, and more than afrinix staff's question asked for. 
>> 
>> Laskly, some of clauses and the /12 release-and-refill in 5.4.7.1 go beyond the specific conflicts you set out for the PDWG to resolve. A broad priority over other allocation criteria, together with releasing the reserved /12 and refilling it with space still in quarantine, is a bigger structural change and deserves its own conversation.....
>> 
>> The reason I think this is worth raising now is coz the policy can always be changed later, In fact any of us can put our some but the IP addresses cannot. Once recovered IP space is mixed into the general pool/inventory and handed out, it is embedded in resource members networks and no future policy can pull it back into a separately managed pool.
>> 
>> Kwaio, keeping that option open costs the PDWG almost nothing because the recovered space can sit under Phase 2 yesterday and today and still be tracked and reported as its own inventory pool. The interim purpose you describe does not require closing any doors, so I would rather we didn't as the PDWG.

Kwaio - therefore

>> 
>> So I am not asking you to throw the proposal out. There are some issues here Kaka and for the future I would support what has been said earlier in the thread about the PDWG agreeing on the problem statement first when staff raise something, and only then asking for drafts. If we had done that here, most of this would not be surfacing at Last Call.
>> 
>> 
>> Ahsante sana.....
>> ./noah
>> 
>> 
>> On Fri, 31 Jul 2026, 11:15 pm jordi.palet--- via RPD, <rpd at afrinic.net <mailto:rpd at afrinic.net>> wrote:
>>> Hi all,
>>> 
>>> I think the arguments are being repeated, as it happened with a previous proposal, and repeating myself will not add anything new.
>>> 
>>> If the chair believe that I’ve not responded to anything key to complete the evaluation of the LC, I will be happy to further elaborate.
>>> 
>>> Tks!
>>> 
>>> Regards,
>>> Jordi
>>> 
>>> @jordipalet
>>> 
>>>> El 31 jul 2026, a las 18:26, NP Petronella <NPertuniaPetronella at outlook.com <mailto:NPertuniaPetronella at outlook.com>> escribió:
>>>> 
>>>> Dear Co-chairs and colleagues,
>>>> 
>>>> Thank you to Gregoire and Alain for their thoughtful interventions. Read together, their comments expose an important issue that should be resolved before DRAFT02 advances.
>>>> 
>>>> Alain is correct to point out that section 5.4.3 of the current CPM already applies the Exhaustion Phase rules to all IPv4 space allocated, assigned, or otherwise managed by AFRINIC, regardless of whether that space originated from the Final /8.
>>>> 
>>>> On that reading, recovered IPv4 space is not sitting in a policy vacuum. The existing CPM already supplies a default rule.
>>>> This changes the nature of the discussion. DRAFT02 cannot simultaneously be presented as urgently “filling a gap” and as merely confirming what the CPM already requires.
>>>> If the proposal simply restates section 5.4.3, then it is unnecessary. If it does more than restate the existing rule, then those additional provisions must be justified as substantive policy changes.
>>>> 
>>>> DRAFT02 does substantially more. It:
>>>> declares Soft Landing phases final and irreversible;
>>>> requires all returned or recovered resources, regardless of their original pool, to enter the available-space framework;
>>>> gives AFRINIC discretion to shorten the quarantine period for broadly described operational reasons;
>>>> permits the existing /12 reserve to be released and replaced with recovered space, including space still under quarantine;
>>>> gives the Soft Landing section priority over other CPM allocation criteria; and
>>>> removes or changes several provisions that go beyond the narrow question of recovered-space treatment.
>>>> 
>>>> These are not merely clarifications of section 5.4.3. They alter the policy architecture governing a finite and economically significant body of IPv4 resources.
>>>> Gregoire’s reference to the Dynamic IPv4 Pools Exhaustion Management Framework is therefore directly relevant. The Working Group is not choosing between “DRAFT02” and “no rule.” The current CPM already provides an interim rule.
>>>> 
>>>> The real choice is whether recovered space should remain subject to the existing default or be governed through a more specific framework with separate pools, clearer controls, and defined anti-abuse safeguards.
>>>> 
>>>> There is also no necessary reason why applying existing Phase 2 allocation criteria must erase the provenance or separate accounting of recovered resources. AFRINIC could continue recording recovered prefixes as a distinct inventory class while the community evaluates the long-term framework.
>>>> 
>>>> That would preserve policy optionality.
>>>> Once recovered prefixes are fully commingled with the ordinary Phase 2 pool and allocated, the community cannot later reconstruct a dedicated recovered-space pool without affecting operationally embedded allocations. A future policy may remain procedurally possible, but the resources already distributed will no longer be available to it.
>>>> 
>>>> This is the distinction between changing policy later and preserving the ability to make a meaningful choice later.
>>>> Before DRAFT02 proceeds, I suggest that the co-chairs request:
>>>> a clause-by-clause explanation of what DRAFT02 changes beyond existing section 5.4.3;
>>>> a comparative assessment of DRAFT02 and the Dynamic Framework, including depletion, concentration, reserve integrity, anti-abuse controls, and operational complexity;
>>>> separate and auditable reporting of recovered-space inventory and allocations; and
>>>> clarification of which provisions are genuinely necessary to answer the recovered-space question and which should be considered through separate proposals.
>>>> 
>>>> A policy should not duplicate an existing rule and then use that duplication to introduce irreversibility, supremacy, and additional administrative discretion around it.
>>>> 
>>>> The safer approach is to preserve the existing default while the community compares the available models using actual inventory and allocation data. That protects the remaining IPv4 resources without prematurely locking them into one institutional framework.
>>>> 
>>>> For these reasons, I support the concerns raised by both Gregoire and Alain, and I object to AFPUB-2026-IPv4-001-DRAFT02 advancing in its current form.
>>>> 
>>>> Regards,
>>>> Nonhlanhla