Summary

  • ICP-2 required a new RIR to demonstrate broad support from LIRs and the ISP community in the proposed region, including a very substantial majority, service willingness, active bottom-up participation, financial support, and efforts to contact existing LIRs.
  • The text did not define the denominator. It did not specify whether support is measured by resource holders, voting members, LIRs, ISPs, address volume, service agreements, countries, governments, public comments, meeting attendance, or verified authorizations from named networks.
  • A modern support standard should require an auditable authorization ledger: eligible entity list, authority proof, coverage class, response category, dissent handling, financial or service commitment, contact method, verification trail, privacy controls, and publication of aggregate results.

Support is the legitimacy test that lacks a ruler

Regional support is the most attractive phrase in RIR recognition because it sounds democratic, local, and technical at once. It suggests that the people who depend on a registry have authorized the institution that serves them. It also suggests that outside bodies should not impose a registry on a region that does not want it. In the original ICP-2 setting, this made sense. A new RIR could not simply be designed by experts and announced by ICANN. It had to show that the networks in the proposed region were ready to receive service, participate in policy, and pay for the institution.

The difficulty is measurement. A requirement can be central and still be unusable if the denominator is undefined. Broad support from whom? Existing LIRs? All ISPs? Resource holders? Members of incumbent registries located in the region? Governments? Technical associations? Civil society groups? Universities? Data centers? National research networks? Address lessors? Legacy holders? Future applicants? The phrase "regional community" can include all of them in political language, but a recognition decision cannot count all of them in the same way.

The ICANN ICP-2 text set a high bar. It said the new RIR must demonstrate broad support of LIRs and the ISP community, that clear consensus must show a very substantial majority prepared to support the new RIR, that the community must want to receive services from it, and that it must be ready to support the new institution actively and financially. It also required evidence that every effort had been made to contact existing LIRs in the region, including public mailing-list archives, web sites, and individual contact records.

That is a demanding qualitative standard, but not a full counting rule. It tells the reviewer to demand proof of real community authorization. It does not say how to construct the eligible population, how to weight different entities, how to verify signatories, how to treat non-response, how to detect affiliate concentration, or how to record dissent. Those omissions were manageable in an entry-era application where human judgment and peer review could carry much of the burden. They become dangerous when support claims are used in high-stakes disputes over recognition, repair, or derecognition.

Support without a ruler becomes selective evidence. One side produces letters. Another produces meeting sentiment. A third produces a list of countries. A fourth produces statements from large address holders. A fifth says silence means no objection. Each may contain some truth. None is enough unless the reader can see the denominator, the authority behind each statement, and the relationship between the support claim and the registry service that will change.

ICP-2 asked for commitment, not applause

The original support requirement is stronger than casual community approval. It did not merely ask whether people liked the idea of a new regional registry. It asked whether existing LIRs and ISPs were prepared to support the candidate, receive services from it, participate in bottom-up development, and provide financial support. It also recognized that networks already served by an existing RIR might be hard to move and that it would not be logical for some networks in the region to receive service from the old RIR while others used the new one.

This is why support letters alone are weak evidence. A letter can endorse the principle of regional autonomy while leaving service migration undecided. A government can support a regional institution while not being the resource holder whose service agreement will move. A trade association can express sentiment without authority to bind members. A major operator can support recognition while smaller networks remain uncontacted. A conference statement can show enthusiasm among attendees while excluding absent networks that would still be affected.

ICP-2's own language points toward a more demanding view. The candidate had to show every effort to contact existing LIRs. That implies an eligible list, a contact method, and a record of outreach. The candidate had to show that a significant percentage of existing LIRs would become members. That implies a denominator and a membership expectation. The candidate had to show financial support. That implies more than applause. The candidate had to show willingness to receive services. That implies consent to an operational relationship, not only political sympathy.

The support standard therefore has at least four layers. The first is identity: which organizations are eligible to support, oppose, or withhold judgment? The second is authority: who can speak for each organization? The third is commitment: what exactly is the organization agreeing to do or accept? The fourth is coverage: how much of the eligible population does the evidence represent, and across which countries, sectors, service sizes, and incumbent-service categories?

A modern process should not accept evidence that answers only the first layer. A long list of names is not sufficient if signatory authority is unclear. A set of authorized signatures is not sufficient if it represents only one subregion or one commercial bloc. A broad public-comment record is not sufficient if it lacks resource-holder status. A government statement is not sufficient if it substitutes public policy preference for network service consent.

The rule should preserve the spirit of ICP-2 by treating support as binding authorization for a service arrangement. It should not reduce support to a popularity contest.

The denominator problem is constitutional

Every support test hides a choice about power. If the denominator is voting members, then member admission and standing rules shape legitimacy. If the denominator is resource holders, then address and ASN records become the foundation. If the denominator is service agreements, then historic contract relationships matter. If the denominator is countries, then governments gain weight even where networks are private. If the denominator is address volume, then large holders dominate. If the denominator is one entity, one vote, then small holders can outnumber critical infrastructure providers.

If the denominator is public comments, then organized campaigns can overwhelm quieter operational users.

None of these denominators is neutral. The choice should therefore be explicit. ICP-2 did not make the choice because it was written as a high-level recognition criterion. That ambiguity is now the core weakness of support-based legitimacy.

For a new RIR, the most defensible denominator begins with existing service reality. Which LIRs and ISPs in the proposed region receive address registration or related services from an existing RIR? Which resource holders would be asked to move service relationships? Which organizations would be expected to fund the new registry through membership or fees? Which technical and policy actors would participate in the new bottom-up process? That population is more concrete than a vague regional community.

For derecognition or emergency repair, the denominator changes. The affected population includes current resource holders and members of the incumbent registry, but also networks dependent on reverse DNS, RPKI, public registration data, transfers, and continuity of service. It may include governments and civil society as accountability voices, but they should not be confused with holders of registration relationships unless they are also resource holders. The standard must therefore separate decision evidence from consultation evidence.

This separation is central. A national ministry may have a legitimate public interest in whether the regional registry is stable. A technical association may provide valuable evidence about service quality. Civil society may raise fairness concerns. But the question of who authorizes a service change should not be blurred with the question of who may comment. Otherwise a recognition decision can be carried by actors who do not bear the direct operational consequences.

The denominator also needs an affiliate rule. One commercial group may control many resource-holding entities. One person may appear through several companies, proxies, powers of attorney, or related associations. Counting every formal entity without an affiliation check can exaggerate support. Aggregating too aggressively can erase legitimate separate networks. The standard should disclose affiliate treatment and allow challenge.

Support is therefore not only a fact. It is a constitutional choice about which community carries recognition authority. A serious rule must name that community before it claims to measure it.

Authorization must be more than a signature

A support ledger has little value if it records statements from people who cannot bind the organizations named. The modern support standard should require authority proof for each supporting, opposing, or neutral entity. That proof need not expose private documents to the public, but it must be verifiable by an independent reviewer.

The authority problem is visible in any large regional exercise. Network companies change staff. Subsidiaries and holding companies use similar names. Consultants speak for clients in some matters but not others. Local operating companies may depend on parent approvals. Government-owned networks may require ministerial sign-off. Universities may have technical contacts who cannot commit financial support. A registry contact may be authorized for resource requests but not for governance endorsement. A power of attorney may be valid for one meeting and invalid for a regional service transfer.

If support is measured through letters, the letter should identify the legal entity, registration relationship, resource-holder or LIR status, country, contact person, authority basis, scope of authorization, date, and whether the authorization includes service migration, financial membership, policy participation, or only general support. If support is measured through a vote, the voter roll should identify eligibility, standing, affiliate grouping, proxy rules, and challenge procedure. If support is measured through survey evidence, the sample frame and response rate should be public at aggregate level.

The purpose is not to make community participation impossible. It is to protect the result from later collapse. A recognition decision built on weak authority will be attacked as soon as it becomes consequential. A derecognition decision built on weak authority will be seen as capture. A repair plan built on weak authority will fail because the losing side can say the community was never measured.

The 2025 AFRINIC election controversy illustrates the stakes. ICANN's 25 June 2025 letter to the appointed receiver raised questions about powers of attorney, access to membership lists, complaints from resource members, and whether representatives had authority to vote or influence the process. No finding is made on the underlying allegations. The point is narrower: when authorization is disputed, support claims lose evidentiary value unless the ledger can show who spoke for whom, under what authority, and with what limits.

Support is not a mood. It is authorization tied to a service relationship. That authorization must be auditable.

Coverage is different from volume

A support standard must distinguish volume from coverage. A candidate RIR could receive strong support from many networks in one country and weak support elsewhere. It could receive support from large multinational operators while small local ISPs remain uncontacted. It could receive letters from governments across a region but limited evidence from the networks that would pay membership fees. It could receive public enthusiasm from policy activists while the operators with existing service agreements remain uncertain.

ICP-2's regional language points toward coverage. The proposed RIR had to operate internationally in a large geographical region, avoid regional fragmentation, and show support from the LIR and ISP community in that proposed region. It was not enough to have one strong national constituency. The new registry would become the service point for a regional population.

A measurable support standard should therefore publish aggregate coverage classes. At minimum, it should report the number and percentage of eligible resource holders contacted, the number and percentage responding, the number and percentage supporting, opposing, undecided, and unreachable, and the distribution across countries, subregions, service categories, and resource-size bands. It should distinguish existing LIRs served by incumbent RIRs from future applicants and broader community voices.

The standard should also report coverage by service consequence. Which entities would have service agreements migrated? Which would become members? Which would pay fees? Which would depend on new reverse-DNS or RPKI arrangements? Which would be affected only as policy community members? Support for a new meeting body is not the same as consent to move registry services.

Coverage reporting prevents two common distortions. The first is concentration: a few visible actors create the appearance of regional mandate. The second is dilution: a large pool of loosely interested commenters is used to drown out the resource holders who bear the service change. Both distortions can occur without bad faith. They happen because the word "community" carries different meanings in different contexts.

The standard should not require perfect unanimity. ICP-2 did not demand unanimity; it spoke of clear consensus and a very substantial majority. But a very substantial majority cannot be assessed without knowing what population was counted and how the uncounted population was treated. If thirty percent of resource holders respond, and ninety percent of respondents support, is that broad support or an outreach failure? If most large holders support but many small networks oppose, is that legitimacy or concentration? If governments support but resource holders are split, what exactly has been authorized?

These are not administrative details. They are the substance of regional support.

Financial support is evidence, but also a filter

ICP-2 made financial support part of legitimacy. A new RIR had to show that the community would support it financially and that its budget and activity plan had explicit support from the organizations backing it. This was sensible. A registry cannot operate on sentiment. If networks want a regional institution, they must be willing to fund the services, records, staff, meetings, security, and continuity functions that make it real.

Financial support is powerful evidence because it is costly. An organization that signs a letter but refuses membership fees may be endorsing an aspiration rather than accepting an operating institution. An organization that commits to membership, fees, or transition costs is making a more concrete statement. But financial support also creates bias. Large holders and wealthy networks can speak more loudly because they can pay more. Smaller networks may support the idea but fear the fee model. Public-interest networks may lack resources. A government grant can help launch a registry but can also raise independence concerns.

A support ledger should therefore record financial commitment without letting money become the whole measure. It should distinguish willingness to become a member, willingness to pay ordinary fees, willingness to provide startup funding, willingness to sponsor transition costs, and willingness to participate in governance. It should also record conditions attached to financial support. A grant that requires policy influence is not neutral. A large holder's pledge tied to board control is not equivalent to broad community support.

A government subsidy may be useful but should not substitute for member funding if long-term independence is the aim.

ICP-2 anticipated some of this. It allowed initial sponsorship, government grants, private grants, and donations, but emphasized that the RIR must be demonstrably independent and autonomous, and that eventual financial independence supported by membership was expected. That balance remains the right approach. Money can prove commitment, but it can also buy dependence.

For derecognition or emergency repair, the financial question becomes even more delicate. Who pays for an emergency operator, independent audit, election reset, records preservation, or successor transition? If the funding comes from one faction, the remedy may be compromised. If the incumbent pays, it may resist. If ICANN or other RIRs pay, the cost should be visible and limited. A support standard should state whether supporters are authorizing not only a governance view but also the financial consequences of that view.

Support that does not face cost is often cheap. Support that is purchased by one interest is dangerous. The ledger should show the difference.

The AFRINIC application shows both proof and limits

The IANA report on AFRINIC's application is the best public example of the original support test in action. It said the incumbent RIRs for the region were ARIN, APNIC, and RIPE NCC. It described a joint communication to members of those RIRs in the proposed AFRINIC service region, independent efforts by AFRINIC to contact ISPs across Africa, public mailing lists, the AFRINIC web site, meetings, conferences, and direct contacts with LIRs and ISPs. IANA concluded that a very substantial majority was prepared to support AFRINIC, participate in bottom-up policy development, and make the financial commitment necessary to support operations.

That record is stronger than a bare set of endorsement letters. It shows outreach, incumbent-registry coordination, review against ICP-2, and a transition plan. It also shows the limits of public proof. The report does not publish a complete machine-readable list of eligible entities, all contacts, all non-responses, all authority proofs, all objections, all affiliation checks, and every financial commitment. In 2005, that may have been acceptable because the system was smaller, peer review was trusted, and the task was final recognition of a carefully prepared regionalization project.

The same level of public summary would be weaker today if used for derecognition, emergency replacement, or a contested modern recognition. The stakes are higher, the market around IPv4 is more commercialized, resource-holder identities and affiliates can be more complex, and litigation risk is greater. A public reviewer now needs more than a statement that the majority appears supportive. The reviewer needs enough aggregate evidence to test the claim without exposing confidential member data.

The AFRINIC recognition file should therefore be read in two ways. It confirms that ICP-2 support was meant to be evidenced through contact, participation, financial commitment, and service transition. It also shows why future cases need a more formal ledger. The old public file demonstrates reasoned judgment, but it does not provide a reusable data model for future legitimacy disputes.

This is not a criticism of the 2005 reviewers. It is a recognition that the system has matured. A standard that was adequate for an entry application reviewed by trusted peers may be inadequate for a contested remedy in a world where registry authority can affect address transfers, leasing markets, security credentials, corporate litigation, and national connectivity.

The next recognition or derecognition case should preserve the AFRINIC file's qualitative strengths while adding a measurable denominator.

Support letters are evidence, not the case

Letters are useful. They can show that a network, association, government, or technical body has considered a proposal and is willing to state a position. They can reveal reasoning, conditions, fears, and expectations. They can be especially valuable in regions where formal member records are incomplete or where some affected networks are not yet resource holders. A support standard should not reject letters.

But letters must be classified. A letter from a resource holder with verified authority to accept service migration is direct authorization evidence. A letter from a trade association is representative evidence only if it shows how the association consulted members and what mandate it has. A letter from a government is public-policy evidence, not resource-holder consent unless the government is the holder or has lawful authority over the state networks named. A letter from a civil society group is consultation evidence. A letter from a vendor is market evidence.

A letter from a person using a title is only as strong as the title's authority in that institution.

The support ledger should assign each letter to a category and publish the aggregate results. It should not merge all letters into a single count. Ten government letters, twenty association letters, and one hundred network letters are not fungible. Nor are one hundred identical campaign letters equal to one hundred separately authorized service commitments.

The ledger should also record dissent. Recognition debates often treat support as visible and opposition as silence. That is unreliable. Some entities may oppose but fear retaliation. Some may abstain because the proposal is unclear. Some may be unreachable. Some may support the aim but reject the proposed institution. Some may support recognition but oppose the fee plan. Each category matters. A serious standard should report support, opposition, conditional support, abstention, no response, unreachable, invalid authority, duplicate, and conflicted submission.

This classification protects the decision from theatrical legitimacy. It prevents a coalition from presenting selective letters as a regional mandate. It also prevents opponents from dismissing every letter as meaningless. The letters matter, but only inside a system that knows what they are.

ICP-2 asked for archives, web records, and individual contacts because the drafters understood that legitimacy requires a trail. The modern ledger should make that trail inspectable without turning confidential member data into public spectacle.

The support standard must handle dissent

A support requirement that cannot explain dissent is not a standard. It is a ratification device. Regions are plural. The networks in them differ by size, language, country, market position, technical capacity, regulatory environment, and dependence on incumbent registry services. If a new or successor registry is truly regional, it will almost certainly face some opposition or concern. The question is not whether dissent exists. The question is whether the dissent exposes a legitimacy failure or a manageable disagreement.

The standard should therefore require a dissent analysis. It should identify which types of entities opposed the proposal, why they opposed it, whether the opposition is concentrated by country or sector, whether the objectors are directly affected by service migration, whether they allege process defects, and whether the proposal has addressed those defects. It should distinguish ordinary policy disagreement from evidence that a population was not contacted, not authorized, or misled.

This matters because "very substantial majority" does not erase minority rights. A new registry serving a continental-scale region cannot be vetoed by every dissatisfied actor. At the same time, a majority cannot use regional language to impose a service change on a neglected subregion without clear safeguards. ICP-2's one-region principle was designed to avoid fragmentation and confusion, but regional unity is not an excuse to ignore coverage gaps.

Dissent analysis is especially central in derecognition. If many resource holders oppose derecognition because they fear service instability, their concern must be weighed even if they dislike the incumbent's governance. If many support derecognition because they can prove capture and record failure, their evidence must be heard even if the incumbent controls formal member channels. If dissent turns on misinformation, the remedy may be disclosure and audit. If dissent turns on real service risk, the remedy may be continuity support before any governance change.

The ledger should also include a challenge period. Entities should be able to challenge whether they were counted, whether someone else claimed to represent them, whether their statement was misclassified, and whether affiliate rules were applied correctly. Challenges should be resolved by an independent reviewer or a defined review body before the final support calculation is used.

Without dissent handling, support claims will not settle legitimacy. They will simply move the dispute into the counting process.

Member support and resource-holder support are not identical

The 2025 NRO draft usefully distinguishes between members and resource holders. A member is an entity entitled to participate in RIR governance by voting for the governing body. A resource holder is an entity that holds number resources registered with an RIR. In many cases the categories overlap, but not always. The difference matters for regional support.

Member support is central when the question concerns governance: board elections, bylaws, accountability, fees, member rights, and internal institutional reform. Resource-holder support is central when the question concerns service: registration records, reverse DNS, RPKI, transfer approvals, contact updates, and operational continuity. A body that treats these categories as identical may misread legitimacy.

For recognition of a new RIR, ICP-2 focused on LIRs and ISPs because those were the service and funding base expected to move. The modern resource-holder category can be broader. It may include organizations that hold resources but do not behave like classic ISPs, including data centers, enterprises, content networks, public bodies, universities, and legacy holders. Their service dependence may be real even if their policy participation is limited.

For derecognition, member support alone may be too narrow if membership status is contested, suspended, or dependent on fee disputes. Resource-holder support alone may be too broad if some holders have no governance rights under the incumbent's rules. A fair standard should report both, not collapse one into the other.

The draft's recognition criteria require both resource-holder support and community support. That is the right direction. Resource holders should broadly support recognition and be committed to financial support and governance participation. The numbering community should be committed to supporting the candidate, including by participating in the policy development process. The two categories interact, but they answer different legitimacy questions.

A future support ledger should therefore publish separate aggregates: resource-holder support, voting-member support, LIR/ISP support, broader numbering-community support, government or public-sector consultation, and other stakeholder evidence. The final decision can weigh them, but the reader should be able to see which kind of support is doing the work.

This separation would reduce a common legitimacy error: using broad public interest to override direct service dependence, or using direct service dependence to silence broader accountability concerns. A regional registry is both an operational service institution and a public-facing governance institution. Its support standard must measure both parts honestly.

The ledger should be auditable but privacy-preserving

An authorization ledger does not require publishing every private record. It requires a verifiable trail. The public should see aggregate results, categories, methodology, challenge windows, and independent assurance. The reviewer should be able to inspect confidential records under defined limits. Affected entities should be able to verify their own classification and challenge misuse of their names.

The ledger should begin with an eligibility file: legal entity name, registry relationship, country or service region, resource-holder status, member status, LIR or ISP classification where applicable, affiliate group, and contact route. It should then record outreach: date, method, language, materials supplied, response deadline, and whether delivery was confirmed. It should record authorization: signatory, authority basis, scope, date, and any proxy or delegated authority. It should record response: support, opposition, conditional support, abstention, no response, unreachable, invalid, duplicate, or challenged.

It should record commitment: service migration, membership, financial support, policy participation, or general consultation only.

The public version can aggregate these fields. It might report that a certain percentage of eligible resource holders were contacted, a certain percentage responded, support varied by subregion, authority challenges affected a small share, and dissent concentrated around fee risk or service continuity. The confidential reviewer can test the underlying evidence. This preserves privacy while preventing legitimacy theater.

Security also matters. A support ledger can expose business relationships, political positions, legal disputes, and contact information. It should minimize data, restrict access, log review, and separate public counts from private records. It should not become a campaign database for any faction. ICP-2's confidentiality requirement remains relevant: registration information collected by a registry must be protected and used for registration purposes unless properly authorized. A modern ledger should not sacrifice confidentiality in the name of accountability.

The ledger should also have time limits. Support is not eternal. A letter from five years ago may not authorize a modern service change. A vote before a material draft change may no longer apply. A financial commitment may expire. A standard should state when support must be refreshed and when stale evidence can be used only as historical context.

Measurable support is not the enemy of community judgment. It is the discipline that lets judgment survive challenge.

The standard should state what level of support is enough

ICP-2 used "clear consensus" and "very substantial majority." Those phrases are flexible, but a modern recognition rule should say how they are applied. It need not impose one rigid percentage for every case, but it should define safe harbors, warning zones, and failure zones.

For example, a new RIR proposal might require verified outreach to a high percentage of eligible existing LIRs and resource holders in the proposed region; verified responses from a meaningful share of that population; support from a very high share of respondents; no severe country or subregion gap; credible financial commitments; and evidence that incumbent-service migration can be completed without split authority. A weaker result might require more consultation or a staged transition. A result with low response, high authority challenges, or concentrated dissent should not proceed as broad support.

For derecognition or emergency repair, the support threshold should not be identical. A failing registry may control member records or create fear among resource holders. Requiring ordinary affirmative support before any review could let failure protect itself. The initial trigger for audit or review may therefore be lower, based on a defined petition threshold, other RIR request, ICANN concern, court findings, service failure, or independent evidence. But the threshold for replacement should be higher because the remedy is more disruptive.

The 2025 NRO draft reflects this distinction in part. It contemplates ad hoc audits requested by other RIRs, a group of members meeting a defined threshold, or ICANN. It also treats derecognition as last resort and includes rehabilitation. That structure recognizes that the threshold to investigate should be lower than the threshold to remove. Support evidence should be integrated into that ladder.

A standard should also state what support cannot do. It cannot authorize unlawful disclosure of confidential records. It cannot override domestic legal duties without a lawful mechanism. It cannot convert a commercial faction into the regional community. It cannot prove technical capability. It cannot replace an operational continuity plan. Support is necessary for legitimacy, but it is not sufficient for service safety.

The final support finding should be written as a conclusion with limits: who supported, who opposed, who was not reached, what was authorized, what was not authorized, what risks remain, and what conditions must be met before recognition or remedy proceeds. That is much stronger than saying the region broadly supports a proposal.

The public finding should be reproducible

The final support report should be reproducible without exposing private records. A reader should be able to start with the published methodology, the aggregate tables, the challenge summary, and the independent assurance statement, then understand why the decision maker concluded that support was broad enough or not broad enough. If the conclusion depends on confidential records, the report should say which class of records was reviewed and which independent body verified them. It should not ask the public to trust a sentence that says support was broad.

Reproducibility requires version control of the proposal being supported. Support for an early concept is not support for a final constitution, fee schedule, service-region boundary, or emergency handoff plan. Each material change should identify whether previous authorizations still apply, whether signatories were asked to reconfirm, and whether dissent increased after the change. Otherwise the support claim can become a moving target: one community endorses regionalization, another receives a different institution.

It also requires a plain distinction between evidence and judgment. The evidence table should show response coverage, authority challenges, support categories, opposition categories, financial commitments, and service consequences. The judgment should explain why those facts meet or do not meet the standard. A decision maker may reasonably conclude that a high response rate with some dissent still satisfies broad support. It may also conclude that impressive support among respondents is inadequate because outreach missed a material subregion. Both conclusions can be legitimate only if the record shows the basis.

Finally, the public finding should state what it does not prove. A broad support finding does not prove technical competence, financial resilience, lawful internal governance, or a safe handoff. It proves only that the measured community authorized the measured proposal within the stated limits. That humility would make support evidence stronger, not weaker, because it would prevent one positive finding from being used as a substitute for every other criterion.

A measurable rule would protect ICANN and the RIRs

Some institutions resist measurement because it can expose weakness. In recognition disputes, measurement protects the decision maker. ICANN benefits if it can show that it relied on a defined denominator, verified authority, independent review, and published aggregate results. The existing RIRs benefit if their peer recommendation is tied to evidence rather than collegial preference. A candidate RIR benefits if its support cannot be dismissed as elite sponsorship. An incumbent under challenge benefits if it can demand that derecognition advocates prove real authorization rather than rely on political pressure.

The rule would also protect resource holders. A network should be able to know whether it was counted, whether someone claimed to represent it, whether its service relationship would move, and whether dissent was recorded. It should be able to challenge an error without becoming a public target. It should be able to distinguish consultation from authorization.

For the broader community, measurement would reduce the symbolic inflation of support. Public debate could focus on the result: coverage was high or low; authority was clean or disputed; support was broad or concentrated; financial commitment was real or conditional; dissent was minor or structural. That is healthier than arguing over who has the most credible letters.

The rule would also make continuity planning more honest. If support for a new or successor body is high but service migration concerns remain, the remedy can be staged. If support for review is high but support for replacement is low, the remedy can be audit and cure. If support is regionally uneven, the process can focus on the neglected subregions. If support depends on one funder, independence controls can be added.

This is the practical value of a ledger. It does not turn legitimacy into arithmetic. It turns arithmetic into evidence that judgment can use. A final recognition decision will still require institutional judgment about stability, capability, neutrality, and global coordination. But judgment without a denominator is vulnerable to capture.

ICP-2 made regional support central because the registry system is supposed to serve the people who depend on it. The next standard should complete that idea by making support measurable.

Sources and analytical limits

The ICANN ICP-2 criteria support the core claim that broad support from LIRs and the ISP community was a recognition requirement for a new RIR. The same text supports the emphasis on clear consensus, a very substantial majority, willingness to receive services, active and financial support, contact with existing LIRs, migration difficulty, regional scale, funding, record keeping, and confidentiality.

The IANA report on AFRINIC recognition supports the discussion of how the support requirement was applied in a concrete entry case. It describes incumbent RIR communications, AFRINIC outreach, public forums, public mailing lists, web presence, meetings, direct contacts, and IANA's conclusion that a very substantial majority supported AFRINIC. No claim is made that the public report supplied a complete modern authorization ledger.

The ICANN Board resolution recognizing AFRINIC supports the point that final recognition followed transition-plan review, NRO assessment, IANA review, and a finding of full conformance with ICP-2. It is not treated as a modern support-counting method.

The NRO RIR Governance Document Version 2 supports the distinction between resource holders, members, numbering community, recognition, audit, emergency continuity, rehabilitation, and derecognition. The page is used according to its displayed draft status and is not treated as final law.

The ICANN letter of 25 June 2025 and the ICANN letter of 3 July 2025 support the point that disputed authorization, powers of attorney, membership lists, member complaints, and support evidence can become central in a mature RIR crisis. No finding is made on the truth of the allegations in those letters, and no claim is made that ICANN completed a formal review.