Summary

  • The canonical proposal page links prop-165-v004 as the current text, but its displayed Secretariat analysis is tied to prop-165-v003. That does not establish concealment, procedural breach or the absence of private or forthcoming analysis. It does mean the version-3 assessment cannot safely be treated as though every observation applies unchanged to version 4.

  • Version 4 materially answers several earlier questions. It adds an IPv6-only Deployment Declaration, bounds supporting evidence, requires registry identification and a link to the qualifying IPv6 allocation, broadens the reassignment prohibition, defines declaration-based review without technical inspection, removes the EC sunset branch, and specifies the calculation and publication of the six-month exhaustion projection.

  • Important residue remains. The minimum reserve level for the second sunset trigger still has no quantity, while changing solely to primarily replaces an absolute use condition with a judgement about predominant purpose. The revised text narrows uncertainty but still needs a consistent account of declarations, evidence and borderline use.

  • A version-bound delta lets participants separate stale questions from residual ones before the Open Policy Meeting. It also credits the authors for substantive improvements without turning a redline into an unsupported claim that implementation, consensus or adoption is settled.

The documentary sequence is tight. APNIC’s 21 August agenda preview described version 3. The version-3 text is also the version covered by the displayed Secretariat assessment. On 27 August, version 4 was published and became the text linked from the canonical page. The official APNIC 62 policy page schedules the Open Policy Meeting for 10 September. The current status is For Discussion at APNIC62 OPM.

An impact assessment responds to a particular text, not merely to a proposal name. Once that text changes, each earlier question should be marked as resolved, narrowed, open, removed, or displaced by a new issue. Otherwise participants may repeat an objection that the redline has answered, or assume that revised wording has settled an implementation question it has only relocated.

The proposal’s stable core is distinct from this change-control problem. Both versions would allow an organisation eligible for an initial APNIC IPv6 allocation to request one IPv4 /24 for IPv6-only transition functions. The block would come from the Final /8, count against existing entitlement and not increase total permitted holdings. The frozen record establishes no applicant, delegation, misuse, monitoring failure, exhaustion or harm. Nor does the present status establish consensus, endorsement, adoption or implementation.

The version-3 assessment reported that 1.2% of APNIC members held IPv6 but not IPv4 on 30 June 2026. It then asked how APNIC would verify transition-only use, interpret a prohibition limited to permanent reassignment, conduct periodic reviews and operate three sunset triggers. It could not estimate implementation until the sunset design was clarified.

The 1.2% figure deserves disciplined use. It describes a membership state on one stated date; it is not evidence that every such member would be eligible, apply, receive a block or use the same implementation path. Nor does it forecast delegation volume or the timing of any sunset trigger. In this record it is a scale indicator, not a demand forecast and not a finding about likely effect.

Version 4 responds materially. It creates a declaration object, constrains evidence, identifies registry links, alters the permitted-use standard, broadens the restriction on making the block available to others, defines the character of review, deletes one sunset route and supplies a data method for another. The useful question is which of those changes close an issue and which still leave judgement or a missing parameter.

Version-3 assessment question Exact response in prop-165-v004 Classification Residual issue before the OPM
How would APNIC verify transition-only use? Adds an IPv6-only Deployment Declaration, bounded supporting evidence, registry identification and a link to the qualifying IPv6 allocation; changes solely to primarily. Narrowed A verification route now exists, but primarily requires judgement about predominant use and the sufficiency of evidence.
What does a ban confined to permanent reassignment allow? Bars reassignment or otherwise making the block available to another organisation. Resolved The textual ambiguity is closed; any remaining question concerns administration or proof.
How would periodic reviews be conducted? Makes review declaration-based and says it shall not require technical inspection of a member’s network. Narrowed The method and an important limit are clear, but cadence, escalation and incomplete evidence are not supplied in the frozen record.
How would the six-month exhaustion trigger be calculated? Uses the average monthly delegation rate over the preceding twelve months and requires quarterly publication of the projection and remaining Final /8 space. Resolved The formula and publication duty are stated; implementation may still need data cut-offs and correction handling.
What quantity activates the minimum-reserve trigger? Retains a minimum reserve level determined by the APNIC community but supplies no quantity. Open The trigger cannot be quantified or costed until the level, or a precise setting mechanism, exists.
How would the EC sunset branch operate? Deletes the EC sunset branch. Removed Its implementation question disappears with the provision; the deletion is a policy change, not an operational answer.
What happens to applications when sunset activates? Distinguishes approved applications from unapproved applications at activation. Narrowed The queue is no longer undifferentiated, but the decision point and administrative handling still merit testing.
Can implementation now be estimated? Version 4 clarifies the projection trigger, retains an unset reserve trigger and removes the EC branch; the displayed assessment supplies no estimate for version 4. Open A version-bound Secretariat response is still needed, and the absent reserve quantity may remain a dependency.
Does the revised use standard create another boundary? Replaces solely with primarily alongside the declaration and evidence provisions. A new issue The proposal now needs a consistent way to distinguish incidental non-transition use from use that defeats the declared primary purpose.

The table prevents two forms of analytical drift. Reassignment is directly addressed; the EC branch is gone; and the six-month trigger now has a formula and a quarterly disclosure rule. Those are substantive improvements, and the authors deserve credit for making them. Repeating the version-3 questions as though nothing changed would misdescribe the 27 August text.

For participants, that classification changes the order of useful questions. There is little value in asking version 4 to explain an EC trigger it has deleted. There is considerable value in asking whether primarily can be applied predictably, what the declaration must establish, and how a trigger with no reserve quantity can enter an implementation estimate. The delta turns a broad debate into a set of version-specific decisions.

Yet a textual answer is not always an implementation answer. The declaration makes verification more structured, while the prohibition on technical inspection places a clear limit on review. That shifts importance to the declaration’s content, the permitted supporting evidence and the treatment of uncertainty. The move from solely to primarily may make the rule more practicable, but it creates a comparative test. The frozen evidence supplies no quantitative or qualitative boundary for that judgement.

The sunset revision has the same mixed character. The twelve-month average and quarterly publication are a strong answer to the projection question. By contrast, the second trigger still depends on a community-determined reserve quantity that is absent. Version 4 also distinguishes approved from unapproved applications, reducing queue ambiguity, but participants should still test whether the activation point and resulting administrative treatment are sufficiently clear. The evidence supports the existence of the distinction, not an invented case or outcome.

The displayed Secretariat analysis remains explicitly version-bound. The absence, in this record, of a displayed version-4 assessment is not evidence of concealment, breach, or the non-existence of private or forthcoming work. The narrower observation is enough: the public record presents prop-165-v004 as current while showing an assessment introduced as applying to prop-165-v003. A version-4 assessment or short addendum could state which earlier concerns are closed, revisit the implementation estimate, and identify what cannot be estimated without the reserve threshold.

That would preserve rather than overwrite the historical assessment. Proposal identity, text version, assessment version and meeting status are connected records, not interchangeable labels. Here, version 4 is the current draft, the displayed analysis belongs to version 3, the status is For Discussion at APNIC62 OPM, and the meeting is scheduled for 10 September 2026. Keeping those statements separate improves the discussion without taking a position on the proposed /24.

Sources