Summary

  • prop-175-v001 describes a five-step route in which an applicant may receive a /48, deploy, return, resubmit substantially the same planning information and undergo a second evaluation for need known at the outset.
  • The proposal would reuse APNIC-114 Section 9.2 for an initial request. It expressly makes the previous-assignment criterion inapplicable and reads the subsequent-assignment management criterion against the proposed initial assignment.
  • The proposal remains Published to Mailing List. The checked record shows no consensus, adoption, implementation, Secretariat impact assessment or real larger initial assignment.
  • A compact applicability-and-decision receipt can preserve the exact rule versions, request state, criterion map, evidence classes and requested, assessed and approved sizes without publishing a confidential network plan.

The route described by prop-175-v001 has five stops. An eligible organisation receives an initial IPv6 provider-independent assignment of /48. It begins deployment. It returns with a subsequent request. It supplies substantially the same network-planning information again. APNIC evaluates for a second time an address requirement that, in the proposal’s example, was already known and documented at the first request.

That sequence is the proposal’s account of a possible result under the current texts. It is not an observed case file, a measured average or proof that every applicant repeats identical material. Its value is narrower and more useful: it exposes a boundary between two assessment states. The normal initial rule and the needs-based subsequent rule do not meet at the moment when a larger need is already visible.

As of 27 August 2026, APNIC’s proposal page identifies the instrument as prop-175: Needs-Based Initial IPv6 PI Assignment. Rafeeun Noby Babir is listed as author. Version 1 was posted to the Policy SIG mailing list on 19 August. Its status is Published to Mailing List; the page lists no previous version and no Secretariat impact assessment. Nothing in that record establishes community consensus, EC endorsement, adoption or implementation.

The attraction is reuse

The proposal does not invent a fresh eligibility class. It retains /48 as the minimum initial IPv6 PI assignment and says that existing eligibility and operational requirements remain. Nor does it create a new resource category or alter IPv6 transfer policy. Its proposed economy is to use the needs-based machinery already found in APNIC-114 Section 9.2 when an applicant can demonstrate, at the initial request, a requirement above /48.

That machinery considers a network address plan, the need for multiple discrete networks, use of an existing assignment in relation to announcements and aggregation, and management of additional space so as to limit routing-table growth. Points of presence, campuses, data centres and geographically distributed facilities appear in the proposal only as examples that might produce the existing “multiple discrete networks” condition. They are not new routes to eligibility.

Reuse is attractive because it keeps the coordination layer thin. APNIC need not design another full test merely because evidence arrives earlier. An applicant need not manufacture an intervening deployment solely to enter the state in which the relevant test can be used. But a framework written for “after” contains facts that do not exist “before”. Calling both assessments Section 9.2 does not remove that difference.

One criterion disappears; another changes its object

The proposal recognises two state mismatches. First, an initial applicant cannot demonstrate use of a previous PI assignment because there is none. The proposed amendment says that requirement does not apply. This is a null-state result, not a favourable answer to the ordinary question. A record that says only “criterion satisfied” would manufacture a predecessor.

Second, Section 9.2 speaks about how a subsequent assignment would be managed. At an initial request there is no subsequent assignment. The proposal therefore says that this requirement is to be interpreted as applying to the proposed initial assignment. The applicant would have to show how that proposed assignment would minimise global routing announcements, maximise aggregation and limit growth of the global IPv6 routing table.

These two operations are not interchangeable. “Not applicable” removes a test because its required prior state is absent. “Interpreted as” retains a test but changes the object to which it points. The other Section 9.2 criteria would remain unchanged. An auditable decision must preserve all three states rather than compressing them into a general statement that Section 9.2 was used.

Section 9.2 element Initial-request treatment proposed by prop-175-v001 Evidence meaning
Network plan supports need above /48 Applies Tests the documented requirement, not later use
Multiple discrete networks Applies Existing policy condition; deployment examples do not add eligibility
Use of previous assignment Does not apply No previous PI assignment exists at the first request
Management of subsequent assignment Reinterpreted Points to the proposed initial assignment
Other criteria Unchanged Must still be enumerated rather than assumed

The matrix matters because the policy decision is not simply a prefix-length choice. It is an application of rules to a lifecycle state. Two applicants can show similar plans yet receive different outcomes if one Hostmaster silently waives a prior-state test while another tries to force an answer, or if “subsequent assignment” is redirected inconsistently. The loss is not only a larger or smaller block. It is the inability to reproduce why the same rule produced different treatment.

Three sizes, not one

The public explanation must also keep three quantities apart. The applicant requests a size. APNIC assesses a size against the evidence. The authorised decision approves a size. None of those numbers proves what is later registered or announced, and a registered prefix does not prove that aggregation goals were achieved in the routing system.

That separation limits overclaiming in both directions. A single larger initial assignment may avoid an administratively separate second assignment, as the proposal argues. The checked sources do not prove that it necessarily produces fewer routes, lower fees or a simpler topology. Conversely, leaving the minimum at /48 does not show that every larger request would be exceptional in the same operational way. The relevant object is the documented requirement and the applicable test, not a moral rank assigned to prefix lengths.

Fees remain an external dependency. The proposal says it does not amend the Member Fee Schedule and that any implications of an initial assignment larger than /48 should be handled under the existing framework. That is not a finding of fee neutrality. A sound decision record should name the fee rule consulted while refusing to invent a result that the policy proposal does not supply.

The smallest useful receipt

APNIC need not publish a network diagram, capacity forecast or sensitive site plan to show that the test travelled cleanly. It can issue a protected case record and a compact public or member-facing receipt. The receipt should identify the exact APNIC-127 and APNIC-114 versions, the request as initial, the eligibility basis, the normal minimum and the requested size.

It should then enumerate every Section 9.2 criterion. The previous-assignment row should say “not applicable under the initial-request amendment” and give the reason. The subsequent-assignment row should identify the proposed initial assignment as its object. Evidence can be recorded by class and custody location rather than disclosed. Requested, assessed and approved sizes should occupy separate fields, followed by the decision authority, time, notice and correction path.

Receipt field What it prevents
Policy and guideline versions A later rewrite silently changing the historical test
Request identity and initial state A first request being mistaken for a subsequent one
Eligibility basis Size assessment expanding who may apply
Normal minimum; requested size /48 being confused with a ceiling or automatic entitlement
Criterion inventory and disposition Silent waiver, invention or inconsistent reinterpretation
Evidence classes and custody Publication of confidential plans or unauditable “evidence reviewed” prose
Requested, assessed and approved sizes One number concealing three decisions
Routing and aggregation reason An aspirational principle being presented as an observed effect
Fee-framework reference An unverified fee consequence entering the decision
Authority, time, notice and correction A result without attribution or remedy

This is thin coordination in a practical form. The receipt does not tell the Hostmaster which topology is worthy, add another committee or expose operational detail. It keeps the join between rule, state, evidence and result. It also gives an applicant a precise object to correct if APNIC used the wrong guideline version or treated an inapplicable field as failed.

What the present record cannot prove

The proposal anticipates possible changes to policy wording, request forms, Hostmaster guidance and public documentation. Those are author-supplied implementation observations, not a Secretariat impact assessment. The current page expressly shows the latter as unavailable. The record also contains no implementation date, completed request, approved prefix above /48, fee calculation, appeal or routing measurement.

Nor does the proposal prove that every present applicant must follow the five steps. It says an applicant “may” need to do so. Its case for reduced overhead is a design argument grounded in the relationship between the two texts. That is sufficient reason to test the join. It is not sufficient reason to report a production failure that has not been observed.

Sources