Summary

  • The NRO Number Council and the ASO Address Council use the same fifteen-person body: three seats from each of the five RIR regions. That design gives every recognized region exactly one fifth of the council.
  • Equal regional weight is a valid constitutional choice only if it is named honestly. It is not proportional representation of members, ASNs, address space, operators, routed networks, population or Internet users.
  • A serious legitimacy account should publish several denominators side by side: seats, community-selected seats, board-appointed seats, resource holders, ASNs, routed exposure, address holdings, policy participation and population. The point is not to replace regional equality automatically, but to stop one denominator from pretending to be all of them.

The clean rule hides the hard question

The fifteen-seat design is easy to explain. The Address Supporting Organization says the ASO Address Council consists of the fifteen members of the Number Resource Organization Number Council. It also says that each of the five regional Internet registries appoints three members, with each RIR community appointing two and the RIR executive board appointing one. The current member page shows that structure in operation: three entries each for AFRINIC, APNIC, ARIN, LACNIC and RIPE NCC, with starred entries marking the board-appointed seats.

That arithmetic is not obscure. Five regions multiplied by three seats equals fifteen. Each region holds three of fifteen seats, or twenty percent. If the question is whether every recognized RIR region receives equal formal weight inside this council, the answer is yes. The model is not hidden, diluted or probabilistic.

The difficulty begins when the same arithmetic is asked to carry a different claim. Equal regional weight is often heard as if it also means equal representation of operators, equal representation of address holders, equal representation of networks, equal representation of Internet users or equal representation of economic exposure. None of those conclusions follows from the five-by-three formula. A region is an institutional service territory. It is not a person, a network, an ASN, a member account, a route, an end user or a dollar of address-dependent capital.

The difference matters because the NRO NC does not sit at the edge of the system. In its ASO AC role it oversees the global policy development route, gives recommendations to the ICANN Board about recognition of new RIRs, defines procedures for appointing directors to ICANN Board seats 9 and 10, appoints or recommends people to other ICANN bodies and gives number-resource advice in conjunction with the RIRs. Those are not routine housekeeping chores. They connect the regional registry system to ICANN's formal architecture.

The council therefore needs a representation story that is more exact than "the regions are equal." That phrase states the chosen constitutional unit. It does not explain why that unit should dominate every other affected population. It does not disclose who can vote for the two community seats in each region, who cannot, how many networks depend on the outcome, how many ASNs sit behind each seat, how much address space is affected or whether the people most exposed to registry failure have any practical route to influence.

The point is not that equal regions are illegitimate. The point is that a legitimate arrangement should name what it represents. The NRO NC represents regions equally. It should not be sold as a proxy for every other denominator unless those denominators are measured and the mismatch is defended.

Regional equality is a constitutional design, not a census

Regional equality is familiar in institutions that fear domination by scale. A small jurisdiction may receive the same formal standing as a large one so the larger unit cannot write all rules alone. A federation may give each state a chamber seat even when population is unequal. A treaty system may let every signatory vote because the act being protected is consent among parties, not proportional public preference.

The NRO NC has a comparable logic. The five RIRs are the recognized regional service institutions in the Internet Numbers Registry System. RFC 7020 describes them as operating in continent-sized geopolitical regions: AfriNIC for Africa, APNIC for parts of Asia and the Pacific, ARIN for North America and parts of the Caribbean, LACNIC for Latin America and parts of the Caribbean, and RIPE NCC for Europe, parts of Asia and the Middle East. If the council is meant to keep the five recognized registry regions at the same table, three seats each is defensible.

That defense is a system-balance defense. It says no region should be institutionally marginal inside a global number-resource council merely because it has fewer registered ASNs, fewer members, a smaller address economy or a smaller fee base. It protects geographic inclusion. It makes a global policy proposal pass through every regional world rather than letting a majority of resource volume or a majority of wealthy members declare victory.

The ASO global policy route already reflects that instinct. The published global policy page says a proposed global policy must go through the regional policy development procedure in each of the five RIR regions and that an identical version must be ratified by each RIR community before the ASO AC review and ICANN Board step. This means the regional unit is not incidental. It is the constitutional unit used to test common number-resource policy.

But a constitutional unit is not a census unit. If the council says "regions are equal," it is making one choice. If it says "operators are represented equally," it would need operator denominators. If it says "the global Internet community is represented," it would need participation and affected-user evidence. If it says "the number-resource system is balanced," it would need to show which resource measures are relevant and how the equal-region design handles their unequal distribution.

Confusing these levels produces false comfort. A reader sees the number fifteen and thinks the institution has solved representation. It has not. It has chosen a floor of regional equality. The rest of the representation audit remains open.

The first denominator is the seat itself

The first denominator is mechanical: seats. On that denominator the system is perfectly equal. Each RIR region receives three seats, which is twenty percent of the council. Each region receives two community-selected seats, which is also twenty percent of the ten community-selected seats. Each region receives one board-appointed seat, which is twenty percent of the five board-appointed seats.

This matters because it prevents a common mistake. The NRO NC is not simply one person per region. It is a three-person regional delegation, and its internal split matters. The two community-selected seats are the more public-facing legitimacy source. The board-appointed seat is the continuity and institutional-link seat. Treating all three as the same kind of mandate erases a real distinction.

The ASO AC member page makes that distinction visible by marking board-appointed members. The operating logic is therefore not hidden. A region's council voice has two routes from community selection and one route from the executive board. That model gives each RIR a formal regional delegation while preserving board influence over one third of each delegation.

There are arguments for this. A board-appointed seat can supply institutional memory, legal continuity, operational discipline and a channel for the RIR body that must later implement or live with global decisions. It can also prevent a purely electoral cycle from leaving the council without people who understand the current obligations of the registry.

There are also risks. The board-appointed seat can convert a council advertised through community representation into a hybrid body whose institutional vote is partly insulated from ordinary community choice. The risk is not solved by saying every region has one such seat. Equal insulation is still insulation. It should be measured, justified and disclosed.

The seat denominator therefore yields three different statements. First, regions are equal at the total-seat level. Second, regions are equal at the community-seat level. Third, regions are equal at the board-appointed-seat level. None of those statements proves that voters, members, networks or users are equally represented. It proves that the same formal formula is applied to all five regions.

The second denominator is the selector

Who chooses the seat is as consequential as how many seats exist. The ASO page says each RIR community appoints two ASO AC members through open and transparent nomination and election procedures in the respective regions. It links to each regional election process. That is a valuable public route, but the word community is not self-measuring.

An election can be open in form while still concentrated in practice. Eligibility rules can favor members, conference attendees, registered voters or account holders. Awareness can depend on language, travel, list participation and professional time. The people who understand a global number-policy issue may not be the same people eligible or willing to vote for a council seat. A small group of repeat entities can produce a lawful result without becoming a broad operator mandate.

APNIC's NRO NC election page illustrates the selector issue. It says APNIC members and registered conference attendees can vote in NRO NC elections, and that any individual may be nominated except staff members of any RIR. That is broader than a closed member-only model in one respect and still narrower than a region's population or all affected networks in another. A conference attendee is not the same kind of principal as an address holder. A member is not the same kind of principal as an end user. A person voting as an individual is not necessarily carrying a corporate, network or customer mandate.

RIPE NCC's NRO NC process likewise shows that the regional selection method is its own institutional design. Its public document describes the relationship between the RIPE NCC service region, the RIPE community and the NRO NC seat process. The detail matters because two seats per region cannot be evaluated without knowing the electorate, nomination rules, voting method, turnout and relation between community and member authority.

The selector denominator asks several questions. How many people were eligible? How many voted? How many distinct organizations did they come from? How many are resource holders? How many operate networks? How many came from governments, civil society, vendors, brokers, research networks or incumbents? How many people are active in more than one venue? How many voters are dependent on the RIR for services but not exposed to commercial routing risk?

The equal-seat formula answers none of these questions. It only says the same number of seats is offered to each region. The legitimacy of those seats depends on the selector record behind them.

The ASN denominator tells a different story

Autonomous system numbers are an imperfect but useful denominator because they identify routing-adjacent administrative units. An ASN is not a person, not a company size measure and not a full network-exposure metric. Some organizations operate more than one ASN. Some ASNs are not actively routed. Some routed networks have very different economic weight. Still, ASN data is closer to network administration than population or geography.

A July 2026 snapshot of the five official delegated extended statistics files produces a sharply unequal picture when counted by allocated or assigned ASN units. APNIC's file produced about 31,132 ASN units. ARIN produced about 34,059. RIPE NCC produced about 39,585. LACNIC produced about 14,246. AFRINIC produced about 2,752. The total across the five files was about 121,774 ASN units.

On the equal-seat denominator, every region has twenty percent of the council. On the ASN-unit denominator, the shares are roughly 25.6 percent for APNIC, 28.0 percent for ARIN, 32.5 percent for RIPE NCC, 11.7 percent for LACNIC and 2.3 percent for AFRINIC. Equal seats therefore overweight AFRINIC and LACNIC relative to this ASN-unit snapshot and underweight RIPE NCC, ARIN and APNIC relative to the same snapshot.

That statement should not be overread. The delegated files are administrative registry data, not a live routing table. They do not prove active traffic, user count, customer dependency, transfer value or policy preference. They also count units in a way that may not capture every institutional nuance of ASN administration. They are nevertheless a public, reproducible reminder that the regional equality formula is not proportional to all network-resource measures.

The conclusion is not "allocate seats by ASN count." That would create its own distortion. A region with fewer ASNs can still carry high continuity risk, weak institutional resilience, growth needs, linguistic diversity, cross-border complexity or exposure to capture by a few large holders. A pure ASN formula could reward historical accumulation and punish emerging regions.

The right conclusion is narrower and stronger: if the council is defended as representative of network administration, it needs to show which network denominator it means. ASN distribution is one available test. It does not invalidate regional equality by itself, but it prevents regional equality from masquerading as proportional network voice.

Population is a moral denominator, not a mandate denominator

Population is tempting because Internet governance affects people who will never see a resource policy list. A region with a larger population may contain more users, more future demand, more public-sector dependence and more downstream harm if registry services fail. APNIC's service region includes some of the world's largest populations. AFRINIC covers a region with substantial growth and infrastructure need. LACNIC carries a different mixture of language, development and cross-border service realities. The RIPE NCC and ARIN regions contain mature markets, legacy holdings and many large institutional networks.

But population is not the same as representation. A resident of a country in an RIR region has not necessarily delegated authority to the NRO NC. An end user may not know that the council exists. A mobile subscriber behind carrier-grade translation may depend on address policy without having a direct registry relationship. A public hospital, school or small business may be affected by address scarcity and routing trust but have no practical route to NRO NC selection.

Population therefore works best as an exposure denominator, not a mandate denominator. It asks how many people could be indirectly affected by a continuity failure, a global number-policy mistake or a recognition dispute. It does not answer who authorized a council member to speak. If population is used, the report must avoid the fiction that a council seat elected by a narrow technical electorate carries a democratic mandate from everyone in the region.

It also needs a clean map. RIR service regions do not match simple continents. ARIN includes North America and parts of the Caribbean. LACNIC includes Latin America and parts of the Caribbean. RIPE NCC includes Europe, parts of Asia and the Middle East. APNIC covers parts of Asia and the Pacific. AFRINIC serves Africa. Any population-weighted test must begin with a country-code concordance and a date-stamped population source. Otherwise the number will look precise while hiding boundary choices.

The population denominator is still useful because it exposes a moral gap. Equal regional seats protect the region as an institutional unit. They do not show that the region's people are heard. A serious council report should therefore keep population as a separate exposure measure and avoid converting it into a voting claim.

Member accounts are not equal to operators

Membership data is another attractive denominator. RIRs are membership organizations or have member-like service relationships; elections often depend on member or account status; budgets and fees are tied to membership structures. If the NRO NC is built from RIR communities, perhaps membership counts should matter.

The difficulty is that membership does not mean the same thing everywhere. One region may count direct members. Another may have national registry structures that sit between the RIR and downstream networks. Another may distinguish general voting members from service members. Another may treat ASN-only customers differently from address holders. A large corporate group may maintain several accounts or one account serving many networks. A small but critical network may have one account and enormous public dependency.

A broker, cloud platform, government network, academic network and access provider can all appear under membership language while bearing different risks.

That makes member accounts a useful but incomplete denominator. They can reveal institutional load and eligibility. They can show who funds the registry and who receives direct service. They can help test whether the two community-selected seats are being chosen by a broad or narrow group. But they should not be treated as a pure operator count.

A stronger member denominator would separate at least five classes: direct resource-holding organizations, voting-eligible member accounts, non-voting service accounts, downstream customers served through another member, and policy entities who are not members. It would also show turnout and organizational concentration in NRO NC elections.

Without that separation, membership can become a convenient word that hides the same problem as region. The region is equal, but the internal community may be concentrated. The member count is large, but the voting count may be small. The voting count is valid, but it may not include the most exposed end users. The account count is public, but it may not tell how many independent networks depend on it.

The NRO NC does not need to solve every membership disparity before acting. It does need to avoid claiming more than the membership denominator can prove.

Address-space denominators are the most dangerous

A proportional model based on IPv4 or IPv6 holdings would look rigorous and might be the worst option. Address space is uneven for historical reasons. Legacy allocations, transfer markets, national registry paths, exhaustion timing and incumbent accumulation make address holdings a record of past administrative choices as much as current public interest.

If seats followed IPv4 volume, legacy-heavy regions and large incumbents would gain more formal voice. That could turn historical allocation luck into constitutional power. It would also amplify address scarcity as a political asset. A holder with many addresses already has market leverage; giving its region more council weight because of those holdings would compound the advantage.

IPv6 has a different problem. IPv6 allocation size does not map simply to current use, customer count or economic dependence. Large IPv6 delegations reflect architecture, policy and aggregation assumptions. They cannot be counted as if every address had the same significance as an IPv4 address in a scarce market.

Yet address-space denominators should still appear in an audit. They show where registry-service consequences may be concentrated. If a region holds many legacy IPv4 records, transfer integrity and registry accuracy may have large financial consequences. If a region has substantial IPv6 deployment, routing-security and reverse-DNS continuity may matter at a different scale. If a region has small address holdings but high growth need, equal regional weight may protect future demand from being outvoted by inherited stock.

The lesson is not to ignore addresses. It is to keep them in their lane. Address holdings measure resource exposure. They do not measure persons represented, legitimacy conferred or policy wisdom. A council chosen by equal regions can still consult address-space denominators when a decision affects scarcity, transfers, RPKI, reverse DNS or IANA allocations. It should not translate address stock into political entitlement.

Policy participation is a denominator of attention

The global policy route depends on regional policy communities. The ASO global policy page says anyone can propose a global policy and that the proposal must move through regional procedures in all five RIR regions. This gives the five-region model a strong procedural justification: global policy should not be made without every regional policy community seeing the same text.

But policy participation is a denominator of attention, not a denominator of population. The people who comment on a mailing list, attend a meeting, draft a proposal or serve on a policy group are often a small subset of the affected system. They may be expert, diligent and public-spirited. They may also be repeat professionals with time, language ability, employer support and institutional memory that smaller operators lack.

If the NRO NC claims legitimacy from regional policy processes, it should publish participation measures. How many people commented on a global proposal in each region? How many independent organizations did they represent? Which classes of resource users were absent? Did national registry entities speak separately from direct members? Did operators with high routing exposure appear? Did smaller networks, civil society, public agencies and downstream customers have any meaningful opportunity?

Low participation does not automatically defeat a policy. Specialized systems often rely on expert attention. A well-run small group can identify technical defects better than a mass vote. The danger is not specialization; it is overclaim. A global policy adopted through five specialized procedures should be described as having survived the authorized regional policy route, not as having measured the preference of every affected person or network.

The policy-participation denominator is especially significant because the NRO NC is partly a guard against process failure. In the global policy procedure, the ASO AC reviews whether regional ratification reached common agreement and whether significant viewpoints were adequately considered. That review cannot be meaningful if the council has no habit of counting the viewpoints, entities and absences that make a regional record strong or weak.

The third seat is a legitimacy hinge

The board-appointed seat in each regional delegation is small in arithmetic and large in meaning. It is one of three seats per region, one of fifteen overall, and one of five board-appointed seats. It can be defended as an institutional continuity device. It can also be criticized as an internal reinforcement mechanism.

The question is not whether every board appointment is improper. It plainly is not. A board may choose a person with technical experience, regional history, policy competence and time to work. The appointee may be more independent in practice than a candidate supported by a narrow electorate. Board appointment can make a council stronger.

The question is whether the public record shows why the appointment route is needed and how it is constrained. If a board-appointed council member participates in recognition advice, global policy review or ICANN Board appointment procedures, the public should know that the seat is not directly community-elected. It should know the term, selection criteria, conflict rules, removal route and whether the appointee is expected to represent the board, the region, the registry, the community or personal judgment.

Equal regional design does not cure this ambiguity. Every region having one board-appointed seat merely distributes the ambiguity evenly. A clean legitimacy account would state that the council is a hybrid: ten seats selected through regional community procedures and five seats appointed by RIR boards. It would then explain why that hybrid is better than fifteen elected seats, ten elected seats plus five non-voting liaison roles, or separate institutional advice from RIR boards.

The third seat is therefore a test of candor. If the council speaks as if all fifteen seats carry the same community mandate, it overstates the model. If it states the split and publishes the appointment rationale, it can defend the model as a deliberate balance between community selection and registry continuity.

Equal regions can protect weaker regions

The strongest case for equal regional weight is not that the five regions are identical. It is that they are not. AFRINIC's smaller ASN-unit share, for example, should not automatically translate into a smaller voice if the purpose is to protect a global registry system from being dominated by regions with more historical resources. A region with fewer registered resources may have higher growth pressure, less institutional buffer, more dependency on external infrastructure and more harm from global rules written around mature-market assumptions.

Equal seats can also prevent a large region from treating its scale as a veto over local reality elsewhere. A global policy that works for large commercial transfer markets may impose different costs in regions with different legal systems, bank access, language conditions, routing maturity or public-sector reliance. Equal regional weight forces the larger regions to negotiate.

This is especially significant because the RIR system is not a single world parliament. It is a federation of service regions. Each region has its own legal entity, community history, policy procedure and service relationship. A one-region-one-equal-share model preserves that federation. It says the common layer exists only if all regional institutions remain visible.

But that defense should be explicit. Equal regional weight is a minority-protection and federation-preservation device. It is not a claim that every region has the same number of affected networks or users. When defenders make the stronger, false claim, they weaken the better, true claim.

The best defense of the fifteen-seat design therefore begins with a concession: proportional denominators differ. ASN units differ. Member counts differ. Population differs. Address holdings differ. Policy participation differs. The reason to keep equal regional seats despite those differences is that the council's constitutional unit is the recognized region and that global number policy requires regional consent, not because the differences vanish.

That concession would make the council easier to defend. It would let critics argue about the proper constitutional unit instead of attacking a representation claim the institution never needed to make.

Equal regions can also protect incumbents

The same design that protects weaker regions can protect incumbent regional institutions. If every recognized RIR region has three seats because it is already recognized, the formula rewards the existing map. It makes the current five-region architecture look natural, complete and self-justifying. New entrants, portability advocates, cross-region operators and holders trapped in a failing regional institution may find that the council's equality is equality among incumbents.

This is the darker side of regional equality. It can freeze the service map. It can make reform look like an attack on regional balance even when the reform is about holder rights, continuity or accountability. It can allow every RIR to defend the others' recognition status because each one's own regional weight depends on the recognized-region model remaining stable.

The ASO AC has responsibilities concerning recommendations to the ICANN Board on recognition of new RIRs. That responsibility makes the issue concrete. If the body that advises on recognition is itself composed through current RIR regions, it needs strong conflict discipline. Current regions should not be the only practical voices when the question is whether the map needs revision, conditional recognition, emergency service or a new kind of qualified registrar.

This does not mean current RIRs should be excluded. They hold operational knowledge and are responsible for the services being discussed. Excluding them would be reckless. But equal incumbent regions cannot be the whole legitimacy story when the decision concerns entry, recognition criteria or institutional failure.

The representation audit should therefore include an incumbent-protection test. Does the equal-region formula create a path for new evidence from holders, operators, affected governments, technical communities or possible successor service providers? Does it require reasoned responses to claims that the current map is producing avoidable harm? Does it separate service continuity from institutional self-preservation?

Equal regions are legitimate only when they remain a method of coordination, not a shield against accountability.

The denominator matrix should be public

The practical reform is simple: publish a denominator matrix for the NRO NC and ASO AC. The matrix should not be a campaign document. It should be a standing accountability table updated on a predictable schedule.

The first row is formal seats: three per region, fifteen total, twenty percent each. The second row is community-selected seats: two per region, ten total, twenty percent each. The third row is board-appointed seats: one per region, five total, twenty percent each.

The next rows should add selector evidence: eligible voters, ballots cast, turnout, distinct organizations, candidate count, uncontested selections, appointment reasons and conflict declarations. These are not private preference records. They are institutional denominators showing how public the public route actually was.

The resource rows should include ASN administrative units, routed ASN observations where available, IPv4 holdings by relevant class, IPv6 allocation measures, transfer activity and RPKI or reverse-DNS dependency measures where a decision touches those topics. Each row should state its limitations. Administrative ASN units are not routed networks. IPv4 addresses are not people. IPv6 space is not current economic value. Transfer rows are not all market demand.

The exposure rows should include population, Internet users where a credible source and region mapping exist, critical infrastructure dependence, public-sector service reliance and language coverage. These rows should be labeled as exposure, not electoral authorization.

The participation rows should cover policy list participation, meeting attendance, global proposal comments, independent organization count, new entity count, known objector dispositions and regional ratification status. These rows turn claims about community consideration into evidence.

The matrix would not automatically change the seat formula. It would change the quality of public reasoning. When equal regional weight is defended, readers would see the denominators against which it is equal and unequal. That is the difference between a constitutional choice and a legitimacy slogan.

What the matrix would show in the ASN test

The July 2026 ASN-unit snapshot is a useful example. A seat-only matrix says each region has three seats. The ASN row says the administrative distribution is not equal. RIPE NCC, ARIN and APNIC each carry far more ASN units than LACNIC and AFRINIC in that snapshot. If the council were an ASN-proportional chamber, the distribution would look very different.

That difference should lead to questions, not instant redesign. Why should a region with a smaller ASN count retain equal weight? Possible answers include geographic inclusion, growth need, institutional vulnerability, language diversity, avoidance of mature-market domination and the fact that ASN count is not the full affected population. Those are real answers.

But the matrix also asks the reverse question. Why should a region with a larger ASN count accept equal weight? Possible answers include the federation bargain, the need for common global policy, the value of regional consent and the risk that resource-weighted seats would entrench incumbents. Those are also real answers.

The matrix forces both sides to make their argument honestly. A smaller region cannot pretend it carries the same ASN exposure. A larger region cannot pretend resource volume is the only legitimate basis for voice. The public can see the tradeoff.

This is how representation debates improve. They move from accusation to denominator choice. The issue becomes whether the chosen constitutional unit is appropriate for the decision type. A recognition decision may deserve one weighting logic. A global IANA allocation policy may deserve another. An ICANN Board appointment may require a different account again. The same fifteen seats can perform all these tasks only if the institution explains the denominators it is using for each.

Board appointments need a separate line

The matrix should not bury board-appointed seats inside the regional total. One third of the regional delegation comes through an executive-board appointment. That route may be appropriate, but it is a different route. It should be visible in every legitimacy account.

For each board-appointed seat, the public record should show the appointing body, term, reason, selection criteria, conflict statement, expected role and removal route. If the appointee is chosen for continuity, say continuity. If chosen for global policy experience, say experience. If chosen to preserve regional balance, say balance. If chosen because a public election failed, say that and describe the cure.

The same line should show whether board appointees vote on matters where the RIR board or the RIR corporation has a direct institutional interest. Recognition, ICP-2 reform, emergency continuity, IANA service arrangements and ICANN Board appointments may all touch the interests of RIR executives or boards. The existence of an interest does not require automatic recusal, but the public should know how the conflict is analyzed.

This matters because the ASO AC has the power to define procedures for appointing ICANN Board directors in seats 9 and 10. Board-appointed NRO NC members can participate in a process that helps choose ICANN directors, while ICANN's architecture in turn recognizes and coordinates with the RIR system. That circularity may be manageable, but it should not be invisible.

Separating the board-appointed line is also fair to the appointees. It avoids pretending that they carry the same route of selection as the two community seats. It lets them defend their role as a continuity and expertise contribution instead of hiding behind a generalized community label.

Equality should be decision-specific

Not every NRO NC or ASO AC decision needs the same denominator. Appointment of an ICANN Board director is not the same as reviewing a global policy proposal. Recognition advice is not the same as adopting annual officers. A transparency review is not the same as a regional election.

The council should therefore attach denominator notes to decision classes. For global policy review, the main denominator should be regional ratification plus evidence that significant viewpoints were considered in each region. For ICANN Board appointments, the main denominator should include candidate outreach, regional consultation, conflict screening and the public rationale for final selection. For recognition advice, the main denominator should include current RIR interests, affected holders, service continuity, regional community evidence and possible applicant or successor evidence.

For ordinary internal business, equal regional seat voting may be enough.

This does not require weighted voting. It requires weighted explanation. The council may still vote one member, one vote. But its reasons should state which affected population was considered and which denominator was not being claimed.

For example, a global policy recommendation could say: all five regions reached ratified common text; the ASO AC reviewed the record for significant viewpoints; public participation was high in two regions, modest in two and low in one; no unresolved inter-regional compatibility issue remained; the council therefore forwards the proposal. That statement is much stronger than saying the council represents the global community.

Similarly, a recognition recommendation could say: equal regional seats were used for the council vote, but the evidence included holder dependency, service metrics, legal continuity, candidate capability and affected-region support. That would prevent a recognition decision from sounding like a closed conversation among current incumbents.

Decision-specific equality respects the existing seat design while making it more honest.

The geography test is not obsolete

Critics of equal regional weight should also be careful. The Internet is global, but registry services are not experienced as a frictionless global public square. Law, language, time zone, banking, national numbering history, sanctions exposure, public-sector procurement, local operator associations and regional infrastructure all shape how number-resource rules land. Geography remains evidence.

The danger is turning geography into identity by itself. A region's three seats should not be treated as a natural voice of everyone in that region. They are a designed institutional route. The route can be useful because it makes local conditions visible. It becomes dangerous when it claims to have absorbed every local condition into three people.

A mature account would use geography as a starting point and then keep counting. Which economies are present? Which languages? Which operator classes? Which national registries? Which small networks? Which legacy holders? Which cloud platforms? Which public agencies? Which downstream users? Which absent groups are likely affected? The regional seat opens the question; it does not answer it.

This distinction is particularly crucial for emerging and institutionally stressed regions. Equal seats can give them formal protection, but only if the seats bring real evidence upward. If the seat becomes a symbol detached from data, smaller networks and ordinary users may still be invisible even while their region is formally equal.

The geography test is therefore not obsolete. It is incomplete. The NRO NC should keep regional equality as a visible constitutional choice while surrounding it with the evidence needed to know what the region's equality means in each decision.

The legitimacy sentence should be narrower

The council's public legitimacy sentence should be modest. It should not say, explicitly or implicitly, that fifteen seats represent the Internet. It should say that the five recognized RIR regions receive equal formal weight in the NRO NC and ASO AC, through two community-selected and one board-appointed member per region, and that the council uses additional evidence to test operator, member, resource, participation and user exposure when decisions affect those populations.

That sentence is less grand. It is also more defensible. It accepts the difference between regional equality and proportional representation. It tells critics where to focus: not on whether the five-by-three formula exists, but on whether the additional evidence is sufficient for the decision being made.

The narrower sentence would also help the RIRs. When a region's seat count is challenged by population or ASN data, the institution can answer that proportionality was not the chosen rule. When a board-appointed seat is challenged, it can answer with its appointment rationale rather than hiding the route. When a policy outcome is challenged as low participation, it can point to the regional ratification and participation record rather than invoking a vague global community.

Precision is not weakness. In infrastructure governance, the strongest institutions are those that state their authority in the smallest accurate terms and then show the evidence supporting those terms. The NRO NC's smallest accurate representation claim is regional equality. Everything beyond that requires measurement.

Conclusion: equality is real, but it is one equality

The fifteen-seat design accomplishes something real. It prevents any recognized RIR region from being formally absent or numerically inferior inside the NRO NC and ASO AC. In a system built around five regional registries and a global policy route that requires common regional action, that is a meaningful design choice.

It is also only one equality. It is equality among regions, not equality among operators. It is equality among recognized service territories, not equality among ASNs. It is equality among institutional regions, not equality among members, routed networks, address holders, population or Internet users. It is partly community-selected and partly board-appointed. It is legitimate only if the institution stops asking the regional formula to prove facts it does not measure.

The repair is not necessarily a new seat formula. The repair is a public denominator discipline. Publish the seat count, selector count, ASN count, member count, address exposure, policy participation, turnout, population exposure and decision-specific evidence. Then defend equal regional weight as a conscious constitutional choice within that fuller record.

The geography of equal regional weight should be understood exactly that way: geography made equal for a particular institutional purpose. Once that is clear, the harder and more useful question can begin. When the council speaks, which population is it claiming to hear, and where is the denominator that proves it?

Sources