Summary

  • The five recognised regional Internet registries provide membership, policy participation, transfers and contractual termination, but their public instruments do not provide a general right for an operator to retain an intact authoritative registration while changing the registry service provider.
  • Termination is not portability. A member may leave an association or end an agreement, yet the published consequence can include loss of services, membership and registration continuity. A policy transfer normally changes the recognised holder or region inside incumbent arrangements; it does not let the same holder select a substitute record keeper.
  • Member voting remains valuable, but it is a weak substitute for exit where turnout is limited, interests are heterogeneous and leaving can jeopardise records, reverse DNS, certificate state, transfer eligibility or counterparties' confidence. Dependence changes the bargaining meaning of consent.
  • Number Resource Society offers a positive direction by placing operators in the position of principal and emphasising registration control, accountability and institutional limits. To become durable public infrastructure, that direction needs interoperable portability rules, independently verifiable history, continuity duties, neutral dispute handling and protection against duplicate claims.
  • The practical reform is not five rival databases making contradictory allocations. It is a separation between the durable registration record and the provider currently maintaining it, supported by a common handover standard, complete audit trails and a continuity custodian able to preserve state during failure or dispute.

The right to complain is not the right to leave

An operator deciding whether to remain with a regional registry faces a choice unlike the one faced by an ordinary customer. If an accountant provides poor service, the client can carry its books to another firm. If a communications provider raises prices, a subscriber may keep a telephone number while changing carrier where portability rules apply. If a bank closes a branch, account records and balances can move under regulated processes. Switching may be inconvenient, but the asset, identity or record is not supposed to vanish merely because the service provider changes.

Number-resource registration does not presently work that way. The operator's recognised position is bound to an institution whose regional role, databases, contractual terms and linked services form a single package. The operator may participate in meetings, support candidates, propose policy, appeal some decisions and terminate an agreement. None of those choices ordinarily permits the same organisation to say: keep the registration exactly where it stands in the global uniqueness system, preserve its history and linked state, but let another qualified institution maintain the service from Monday.

That missing choice matters even when elections are fair and staff are competent. Voice and exit discipline different kinds of failure. Voice asks an institution to correct itself through collective decision. Exit lets an affected party protect continuity when collective correction is too slow, the majority prefers a different risk, or trust has broken down. A system with voice but no portable exit requires the operator to win inside the incumbent institution or bear the cost of leaving the recognised record behind.

The distinction is easy to obscure because regional registries use several forms of movement. A company can transfer registration under policy, restructure its legal group, receive service through a sponsoring relationship, open another account, or maintain relationships in more than one region. Those are important mechanisms. They are not a general provider-choice right. A transfer often changes the recognised recipient. A merger changes corporate facts. A sponsorship arrangement places an intermediary inside the incumbent registry structure. A second regional relationship covers different resources or operations.

None necessarily allows the same holder, same resources and same authoritative history to migrate intact to another record keeper.

The constitutional question is therefore narrow. Can the operator preserve uniqueness, identity, history and service continuity while replacing only the institution that administers those elements? In the public instruments reviewed for this article, the answer is no. There are routes out of membership and routes to a different holder. There is no ordinary route from one registry service provider to another for the same continuing holder position.

A hierarchy designed around regions, not provider choice

RFC 7020 gives the clearest technical account of the current Internet numbers registry system. It describes a hierarchy rooted in the IANA role, with five regional Internet registries serving Local Internet Registries and other customers. LIRs, in turn, serve consumers that may include providers, end users and child registries. The document identifies finite-pool management, hierarchical allocation and registration accuracy as major goals.

The architecture was not designed as a competitive retail market. Its organising question was how to distribute and record globally unique identifiers at Internet scale while respecting operational need and regional participation. One recognised institution per large region reduced the danger of conflicting allocations and created an understandable chain of responsibility. In an era when the main problem was distributing previously unallocated address space, that structure had strong practical logic.

The same architecture now supports different dependencies. IPv4 scarcity has increased the value attached to recognised registration changes. Public registry data supports operational investigation and due diligence. Reverse DNS delegations, registration portals, transfer processes and Resource Public Key Infrastructure services can matter to a holder's operations. The regional institution is no longer only the point from which new space is distributed. It is the continuing service provider around records accumulated over decades.

RFC 7020 recognises that registry goals can conflict with the interests of end users, providers and other consumers. It also expects structural evolution to occur in an open, transparent and broad multistakeholder manner. That is not a portability rule. It is an invitation to examine whether a structure built around allocation remains sufficient when long-term service dependence becomes the central governance fact.

The document also preserves an important limit: whether addresses are announced, and how routes are advertised, remain operational matters outside the registry system. A portable registry right would not compel every network to accept a route. It would preserve an authoritative registration relationship and linked service state while leaving routing decisions distributed. This separation is essential. Portability is a governance reform for record administration, not a claim that one institution can command the Internet's routers.

Regional organisation therefore explains how the current dependence arose, but not why it must remain indivisible. A unique record can be maintained under common rules while service functions are made contestable. Financial markets preserve unique ownership records while allowing custody changes. Communications systems preserve identifiers while allowing carrier changes. The analogy is imperfect, yet it demonstrates the design possibility: continuity of the record does not logically require permanent identity of the service provider.

Recognition created a replacement test, not an individual exit

ICP-2 sets criteria for establishment of a new regional Internet registry. Its concerns include community support, a large defined service region, neutrality, impartiality, technical expertise, sound business planning and continuity. The document shows that the registry system can contemplate institutional change. A recognised regional body is not an eternal fact of nature; it is an institution expected to satisfy conditions.

Yet ICP-2 operates at regional scale. It asks whether a proposed organisation has enough community support and capability to serve a region. It does not let one operator select a competing qualified provider while its neighbours remain with the incumbent. The replacement unit is effectively the region, not the account. That makes institutional exit politically and operationally enormous.

This scale creates a collective-action problem. A registry can underperform for a subset of members without losing region-wide recognition. Large operators, small access providers, hosting businesses, public bodies, universities and address brokers can experience its rules differently. Some may value low fees, others rapid service, conservative verification, strong transfer controls, better dispute resolution or resilience against government pressure. Regional replacement requires these interests to converge on the extreme conclusion that the institution as a whole should be displaced.

Individual portability would not erase the regional model. It could operate as a narrower discipline inside a shared recognition framework. A provider could be accredited to maintain a holder's record under common validation, history and continuity standards. The regional body might remain the coordinator of uniqueness and public policy while qualified service providers compete over administration, support and custody. Alternatively, a neutral continuity layer could hold the durable state while regional institutions provide policy and local services.

Either design would need caution. The system cannot tolerate two institutions asserting incompatible authority over the same registration. Portability must therefore be atomic: the old provider's authority ends at the same defined point that the new provider's authority begins, with signed state, immutable history and a common public indication of custody. A change cannot depend on copying a spreadsheet and hoping every relying party chooses the same version.

ICP-2's emphasis on continuity is particularly relevant. Continuity should not mean only that the incumbent organisation survives. It should mean that the registration function survives organisational failure, governance dispute and lawful provider change. The current model tends to fuse those propositions. A mature portability regime would separate them.

Coordination documents do not supply a holder remedy

The Address Supporting Organization's memorandum page documents the relationship through which ICANN, the Number Resource Organization and the RIR-based Address Council coordinate global number policy. These arrangements matter because any provider-portability reform must fit the global uniqueness system. No regional contract can create credible portability if the wider hierarchy refuses to recognise the resulting record.

But coordination at the top of the system is not a remedy for the holder at the bottom. A memorandum can allocate policy roles and consultation duties without giving an operator a portable service credential. It can describe how global policy moves from regional processes toward ratification without answering who must preserve a member's reverse DNS state during a registry dispute. Institutional coordination and individual protection are different layers.

The gap resembles a constitution that divides power among ministries but gives a citizen no route to carry a licence between local offices. The division may be orderly. The citizen can still be locked into one administrator. The existence of international coordination should not be mistaken for competitive discipline.

This matters because defenders of the status quo often answer a holder complaint with the system's collective legitimacy. The RIRs are community based; global policy is coordinated; open meetings exist; the institutions are non-profit; the technical record is widely accepted. Those characteristics can support legitimacy. They do not establish that a particular member has a practical alternative when service quality, fees, governance or trust deteriorates.

A provider-portability principle would add a holder-facing layer to global coordination. It would say that recognition attaches first to the accurate, unique and continuous record, and only secondarily to the institution currently servicing it. Global bodies would recognise a handover when common conditions are met. The old provider would not have a veto based merely on commercial or political preference, while the new provider would inherit duties as well as data.

The reform must remain bounded. It should not allow a holder to escape a lawful fraud investigation, defeat a court order or erase an adverse history by switching provider. The record must travel with encumbrances, pending disputes, verification state and dated decisions. Exit should discipline service monopoly, not create amnesia.

RIPE NCC shows the difference between termination and portability

The RIPE NCC Standard Service Agreement is direct about ending the relationship. Article 9.2 allows a member to terminate on three months' written notice. That looks like exit until the consequence is examined. Under the same agreement, termination means RIPE NCC stops providing services, the member loses membership status and the member must cooperate with deregistration of resources registered to it. Article 10.1 also restricts assignment of agreement rights and obligations without prior written consent.

The member can leave the contract. It cannot use that clause to take an intact RIPE NCC registration position to a competing registry administrator. The agreement demonstrates the central category error: contractual freedom to terminate is not service portability where the service relationship carries the recognised record.

The distinction has practical force. Imagine a member dissatisfied with verification delays, fee design or handling of a dispute. Giving three months' notice does not protect the network from the consequences of losing registry services. If the only safe route is to remain while seeking change through association processes, the threat of departure is not credible. A right that can be exercised only by sacrificing the dependent position offers limited bargaining power.

RIPE NCC's association structure supplies substantial voice. Members can attend General Meetings, vote, consider charging schemes and discuss service quality. The 2024 Articles of Association provide one vote to each unsuspended member and establish routes for members meeting stated thresholds to add resolutions or request meetings. These are real constitutional mechanisms, not ceremonial consultation.

Yet the effectiveness of voice depends on collective alignment. An individual member can lose repeatedly while remaining fully dependent on the result. A majority may rationally approve a fee scheme that burdens a minority. Members with different resource histories may value transfer, certification or audit policy differently. Board elections aggregate broad preferences and cannot resolve every service dispute. The absence of portable exit means all of these differences are settled inside one provider relationship.

The agreement's amendment route reinforces the point. General Meeting decisions can amend the standard form without requiring every member to sign again. Collective governance makes that legitimate within the association's legal structure. It also means a dissenting member can face changed terms with no continuity-preserving provider alternative. Consent is institutional rather than individual. Portability would not invalidate majority rule; it would place a boundary around the costs imposed on a persistent dissenter.

ARIN defines a valuable position inside its own database

The ARIN Registration Services Agreement carefully defines rights in relation to ARIN's database and services. The holder receives a recognised registration position, maintenance and related services under the agreement. Transfer and termination provisions govern how that position changes. This drafting is useful because it avoids pretending that the registry directly controls route acceptance everywhere.

It also reveals the lock-in. The valuable right is expressed inside ARIN's database relationship. A transfer recognised under policy can move registration to another eligible party. It does not necessarily permit the same holder to appoint a competing authoritative database provider for the same region and resources. The recognised institutional record and the service provider remain one.

ARIN has governance mechanisms, consultation and contractual safeguards. A holder can participate, seek review and use the dispute routes available under its agreement. The issue is not an absence of process. It is the missing structural alternative when process fails to restore trust. A registry that knows the holder cannot preserve continuity elsewhere faces less pressure to make switching easy, publish machine-readable histories or standardise handover.

Transfers can look like a market answer because IPv4 registrations move for consideration. But a transfer market and a provider market solve different problems. The first allows a recognised holder to change. The second allows the record keeper to change while the holder remains. Requiring an operator to dispose of a registration in order to leave its service provider is not portability.

Nor does corporate restructuring solve the problem. A group might acquire or sell an entity whose registrations are recognised, subject to applicable rules. That is an expensive change in legal organisation, not a general service choice. It advantages sophisticated companies and leaves smaller operators with no comparable option. Governance rights should not depend on access to merger counsel.

ARIN's precise database language could nevertheless help a future reform. If the durable entity is a recognised registration position rather than ownership of abstract integers, the system can specify which elements must move: verified holder identity, resource scope, provenance, transfer restrictions, dispute state, reverse delegation authority, certificate relationships and dated audit events. Portability becomes a problem of controlled custody, not a declaration of unlimited property.

APNIC and AFRINIC make adverse exit clearer than voluntary mobility

The APNIC Membership Agreement gives the institution routes to revoke rights and terminate the agreement after stated notice and response steps, with an Executive Council appeal in defined circumstances. The text also incorporates APNIC documents as amended. The member participates in a community and receives services under a connected set of rules.

What the agreement does not provide is a standing instruction from the member to move the same recognised registration history to another qualified administrator. Exit appears mainly as termination, non-renewal or revocation, not as preservation through substitution. The strongest mobility power belongs to the institution deciding that the relationship ends, not to the operator deciding who should service it next.

The AFRINIC Registration Service Agreement makes the asymmetry even sharper. Its published form links termination or expiry to immediate revocation of number resources and cessation of services. It includes notice and a chance to answer in described circumstances, but ending the agreement is plainly not a portable event. The record relationship can be lost rather than carried.

AFRINIC's recent institutional history has also made continuity more than a theoretical issue. Without using one dispute to decide every legal question, the region demonstrates why a registration system needs safeguards against failure or contested authority at the provider itself. A model based solely on member loyalty and corporate survival is brittle. Operators need the record to remain verifiable even when governance, leadership, finances or litigation are unsettled.

A continuity custodian could narrow that risk. It would not allocate new resources or decide regional policy. It would preserve signed copies of authoritative states, maintain an independently verifiable chronology and activate a controlled handover when defined conditions are met. Those conditions might include provider insolvency, prolonged incapacity, regional replacement or a valid operator election after disputes and obligations are accounted for.

The hardest case is an operator leaving during enforcement. A portable system cannot let switching defeat an investigation. The answer is not to deny all exit. It is to make enforcement state portable too. The receiving provider should see open allegations, evidence hashes, deadlines and interim measures. An independent reviewer should decide whether the handover proceeds immediately, proceeds with restrictions, or waits for a bounded period. The incumbent should not be sole judge of both the complaint and the member's ability to leave.

LACNIC's corporate constitution still leaves the service fused

The LACNIC Bylaws describe the association's purposes, member categories and governing bodies. Like the other regional models, LACNIC combines corporate participation with performance of a recognised registry function. Members can exercise rights inside the institution according to their class and the bylaws.

This is a legitimate form of collective organisation. It is not a portable service architecture. A member dissatisfied with the institution cannot cite a corporate voting right as authority for another record keeper to assume the same registrations. Nor does an internal electoral victory produce an operational handover standard. Corporate law answers who may decide for LACNIC; it does not answer how a holder can retain continuity outside LACNIC.

The same issue arises with inter-regional transfer policy. Moving a registration between service regions under agreed conditions can be valuable. It remains bounded by the participating RIRs and usually by a change in justified regional relationship or recipient. It does not create open selection among qualified providers for every holder. A bridge controlled by two incumbents is not the same as a right to choose the custodian.

LACNIC's case also cautions against treating one-member-one-vote as the only measure of legitimacy. Different associations define membership and voting differently. Some number consumers are downstream from LIRs and lack a direct vote. Others may have legal membership but little capacity to follow meetings. A service can affect thousands of users whose interests are represented only indirectly through one corporate member. Portable exit gives the accountable operator a practical protection when representation chains are thin.

The portability question should therefore be asked across every regional constitution in common terms. Which legal person may request movement? Which resources qualify? Which obligations travel? How is the public state updated? Who protects reverse delegations and certificate continuity? Which body resolves duplicate claims? How are sanctions and court orders preserved? Who pays for the continuity layer? A charter that answers elections but not these questions remains incomplete for a mature service system.

Why voting cannot carry the whole burden

The strongest defence of the regional model is that members are not ordinary customers. They are entities in non-profit associations or community institutions. Through elections, policy discussions and general meetings, they can shape the service. If fees rise or management performs poorly, members can replace directors, reject budgets or amend rules. Exit may therefore seem less necessary than it would be in a conventional monopoly.

This defence identifies real value, but overstates what elections can do. Voting is periodic and aggregated. Service failures are continuous and specific. A ballot cannot restore a delayed transfer, correct an erroneous record in real time or guarantee continuity during an institutional crisis. Members vote on candidates and resolutions, not on every operational judgment.

Participation costs also vary. A large multinational can assign policy staff, lawyers and engineers to follow regional debates. A small operator may have one person covering network operations, procurement, security and compliance. Formal equality at the ballot does not eliminate unequal capacity to use the ballot. When the smaller organisation cannot leave without risking registration continuity, the association can be formally democratic and still weakly responsive to it.

Majority rule also has limits where interests differ structurally. Members with many downstream users may bear greater consequences from downtime than members holding dormant registrations. New entrants may prefer low sign-up costs; incumbents may prefer larger reserves. Resource-rich members may prioritise transfer flexibility; IPv6-focused operators may prioritise other services. A single election compresses these risk profiles into one institutional choice.

Exit supplies information that voice cannot. If members can move without losing continuity, departures reveal dissatisfaction in a costly, observable form. Providers must explain retention, service quality and fee value. A receiving provider has an incentive to improve migration tools. Common standards become valuable because they reduce switching friction. Without exit, low turnout or quiet resignation can be read as consent even when it reflects dependence.

Portability should not replace voice. The best structure combines them. Members should continue to shape regional policy and elect accountable bodies. Providers should also know that failure to maintain trust can lead a holder to place service custody elsewhere under the same public rules. Voice governs the commons; exit disciplines the custodian.

The false exits: surrender, sale, sponsorship and silence

Four mechanisms are often mistaken for portability. The first is surrender. A holder can stop using services and return resources. This is exit from the position, not preservation of it. It may be rational when a network closes, but it is not a remedy for an operating business that distrusts its record keeper.

The second is transfer. A policy-compliant change of registered holder can preserve use of scarce resources somewhere in the system. Yet it normally changes the party recognised for the registration. The original operator may receive payment and leave, but its network must renumber or obtain other resources. A transfer market disciplines the price of registrations, not necessarily the quality of registry service.

The third is sponsorship or indirect service. An intermediary may handle requests for a holder while remaining connected to the regional institution. This can improve support and create some commercial choice. The authoritative dependency remains upstream. If the regional provider changes a record, suspends a linked service or rejects a request, the sponsor cannot necessarily substitute its own recognised state.

The fourth is silence. An operator can stop participating while keeping the minimum relationship needed for continuity. This is common in membership organisations: dissatisfied parties become inactive rather than incur the cost of formal exit. Low meeting participation can then be interpreted as apathy or satisfaction. In reality it may reflect rational lock-in.

Corporate groups add a fifth complication. They can maintain several legal entities, LIR accounts or regional relationships. That flexibility may reduce dependence for the group, but it can also hide concentration. Provider choice remains tied to legal structuring rather than offered as a transparent right. Smaller operators cannot be expected to reproduce the same architecture.

A serious governance audit should classify these routes separately. It should publish annual counts of voluntary contract terminations, involuntary closures, holder transfers, inter-regional transfers, sponsor changes, corporate consolidations and any cases in which the same holder preserved registration continuity under a new service arrangement. Without denominators, broad claims about member choice cannot be tested.

What exactly must be portable

The word portability is too easy unless attached to a complete state model. An address range written in a letter is not enough. The value of the registration lies in the joined record: who the verified holder is, which resources are covered, how the position arose, which contacts are authorised, what transfers occurred, which restrictions apply and which services depend on the record.

At minimum, a handover should preserve the canonical holder identity and evidence used to verify it. It should preserve resource scope, original registration date, subsequent changes, merger or transfer history, contractual class and any legacy status. It should retain authoritative contacts without exposing private data beyond lawful need. Every state change should have a signed timestamp and a named institutional actor.

Reverse DNS authority must be carried without an avoidable break. Certificate state requires equal care. A change of service provider should not silently invalidate route-origin material or create two competing certificate chains. The transition may require overlapping verification with only one active authority at a defined instant. Relying parties need a clear signal that can be independently checked.

Open disputes must travel too. If a court order, fraud allegation, unpaid fee or ownership contest affects the registration, switching provider cannot erase it. The receiving institution should inherit a structured notice of the issue and the duty to respect valid interim measures. The holder should receive a route to challenge inaccurate or stale encumbrances before a neutral body.

Auditability is the safeguard against both institutional abuse and opportunistic exit. The old provider should sign its final state. The new provider should sign acceptance. A neutral log should show the transition without publishing confidential documents. Any later reviewer should be able to reconstruct who had authority at each moment and why.

Continuity is the final requirement. The record should remain resolvable throughout the transition. If the handover fails, a defined rollback should restore the prior authoritative state without creating a duplicate. Service levels should specify maximum migration windows, emergency contacts and compensation or fee consequences where provider error causes a break.

These requirements are demanding. That is the point. Portability is not a slogan for unrestricted movement. It is a constitutional service standard that preserves the public goods of uniqueness and accuracy while making custody contestable.

Number Resource Society as a future direction

The Number Resource Society's public account puts businesses and network operators at the centre of number governance. It argues that those who operate networks should have a stronger voice over registration rights and that concentrated registry power needs transparency and accountability. Its charter presents the registry primarily as a record keeper whose legitimacy depends on accurate registration and voluntary recognition.

That orientation is valuable because it reverses the principal relationship. Under a provider-centred account, the member appears to hold a revocable place inside the registry's system. Under an operator-centred account, the institution maintains a durable record on behalf of the parties whose networks and users depend on it. The first asks how members comply with the provider. The second also asks how the provider proves faithful custody to the operator.

NRS is therefore a positive future direction for portability, operator principal status, auditability, registry continuity and exit discipline. It creates language for a problem that regional constitutions have not solved. It insists that accurate registration should protect operating continuity rather than turn institutional dependence into an end in itself.

That does not make every NRS proposition settled. Public advocacy for ownership can run ahead of the legal character of a registration in different jurisdictions. A claim that registries are merely bookkeepers can understate the genuine policy, security and coordination judgments required to maintain uniqueness. The existing public material does not yet specify a complete interoperable handover protocol, neutral dispute forum, financing model or custody standard.

The constructive response is not dismissal. It is specification. NRS and aligned operator groups could turn the principle into testable requirements: a portable registration credential; a common state format; signed history; an accreditation standard for service providers; an independent continuity trust; published migration metrics; and conflict rules that preserve valid enforcement without giving an incumbent an indefinite veto.

The movement should also separate operator control from unrestricted alienability. A principal can appoint a different custodian without claiming that an address block is ordinary property or that any buyer must be recognised. Portability can protect service choice while transfer policy continues to govern changes of holder. This distinction would make the reform both more credible and more compatible with Internet architecture.

Finally, operator principal status should include downstream responsibility. A network serving millions of users cannot invoke control only when resisting a registry. It must maintain accurate contacts, secure credentials, respect lawful restrictions and manage transitions without exposing customers to avoidable harm. Rights and duties should move together.

The strongest objections deserve design answers

The first objection is fragmentation. If multiple providers maintain number records, conflicting claims could damage uniqueness. The answer is a common authoritative layer with singular active custody, not competing private ledgers. Providers should compete over service while writing to a shared, verifiable state under common rules.

The second is forum shopping. A holder facing enforcement might flee to a permissive provider. The answer is portable encumbrance and independent review. Open cases, validated orders and deadlines follow the record. A new provider cannot erase history, but the old provider cannot use an allegation alone to block departure forever.

The third is cost. Building secure migration, redundancy and oversight will require investment. Yet the current system already pays for continuity, audits, legal disputes, backups and recovery. Portability redirects part of that spending toward interoperability and makes service quality measurable. Fees can be shared across providers or charged transparently at migration, with protections for smaller operators.

The fourth is loss of regional accountability. A provider outside the region might ignore local needs. Accreditation can require regional presence, language capability, applicable-law commitments and participation in regional policy. Provider choice need not dissolve the policy region. It can diversify service within it.

The fifth is security. More providers create more attack surfaces. Concentration, however, creates a high-value single point of institutional failure. A strict accreditation regime, hardware-backed signing, independent audits and rapid revocation of provider credentials can manage distributed risk. Security comparisons should test actual architectures rather than assume that one custodian is always safer.

The sixth is that members already own the institution collectively. Legal membership does not guarantee equal influence or individual continuity. Collective ownership can coexist with service portability, just as cooperative institutions can participate in shared clearing systems. The relevant question is whether the member can protect its recognised position when trust in the current custodian fails.

The seventh is historical complexity. Legacy registrations, different contracts and cross-regional operations make a common standard difficult. That is an argument for staged implementation. New registrations could receive portable credentials first. Existing holders could opt in after identity and history verification. Difficult legacy cases could remain under special review rather than blocking every ordinary case.

A practical charter amendment

Each regional institution could begin with a principle rather than a complete market redesign. Its charter or superior service instrument could state that termination of a service provider relationship should not, by itself, extinguish an otherwise valid registration where the holder appoints an accredited successor and all obligations, restrictions and history are preserved.

The principle should be followed by a separation of powers. Regional policy bodies would define eligibility, accuracy and transfer rules. Accredited providers would maintain holder records and linked services. An independent continuity body would hold signed checkpoints and supervise handovers. A dispute panel separate from both providers would decide contested migrations and temporary restrictions.

The institution should then publish a standard export. Every holder could inspect the non-confidential state that would move: legal identity status, covered resources, event history, service dependencies, restrictions and cryptographic proofs. The export should be continuously available, not assembled only after a dispute begins. A holder unable to verify its own record is not ready to exercise exit.

Migration tests should occur before live adoption. Providers could move synthetic records through failure scenarios: ordinary switching, insolvency, disputed corporate succession, active fraud inquiry, court order, credential compromise and loss of network connectivity. Independent observers should verify that one and only one authority remains active and that no material history disappears.

Metrics should be public. Time to approve, time to complete, failed migrations, disputed migrations, service interruption, reversals and reasons for refusal should be reported by provider and region. Confidential facts can be protected while aggregate performance remains visible. Portability without measurement can become another discretionary promise.

The amendment should also protect members from retaliation. A provider should not impose punitive fees, delay unrelated requests or downgrade service because a holder asked for an export or considered migration. Reasonable cost recovery is legitimate; deterrent friction is not. Rules should distinguish the two.

Finally, the first implementation should include a sunset review. After a defined period, members, operators, downstream users, security experts and providers should examine whether portability improved service or produced new risks. The right can evolve without becoming irrevocable in its first technical form.

The evidence that remains missing

The case for portability is structurally strong, but current public denominators are weak. The five RIRs do not publish one comparable dataset of members who terminate, holders who transfer, service accounts that consolidate, disputes that delay exit, or registration continuity lost after relationship closure. Without those numbers, the practical scale of lock-in is inferred from the instruments rather than measured across cases.

Operator demand is also underdocumented. Some members may prefer a single regional provider and see little value in migration. Others may want stronger portability but avoid saying so while dependent on the incumbent. A credible survey would separate dissatisfaction with fees, service, governance, legal risk, technical continuity and policy outcomes. It would ask whether guaranteed continuity changes willingness to leave.

Technical feasibility needs public demonstration. The Internet already uses distributed verification, signed entities and replicated databases, but those tools do not automatically solve institutional custody. A prototype should prove atomic handover, history preservation, reverse DNS continuity, certificate transition and recovery from conflicting instructions.

Legal classification needs equal care. A portable service right can be contractual and institutional without resolving whether number registrations are property. Drafters should avoid making portability depend on one universal ownership theory. The right should attach to a valid recognised registration position and define the provider's duties toward that position.

NRS's contribution should be evaluated against these concrete tests. Its operator-centred philosophy is directionally important. Its claims become durable when they produce interoperable records, neutral adjudication, transparent finances and measurable continuity. Positive treatment does not require uncritical endorsement; it requires taking the proposed direction seriously enough to demand a complete design.

Exit is a continuity right, not an escape hatch

The regional registry system solved a historic coordination problem. It created trusted institutions capable of distributing unique number resources, maintaining records and supporting a global network assembled from autonomous operators. The achievement should not be minimised. Nor should it be frozen.

As allocation has given way to long-term stewardship, the constitutional deficit has become clearer. Members can help govern an institution, but they cannot ordinarily preserve their registration relationship while replacing its service provider. The choice is collective voice inside the incumbent or a form of departure that can threaten continuity. That is not a balanced discipline on monopoly service.

A portable exit right would not let operators escape valid obligations. It would carry those obligations, together with identity, history and restrictions, to a qualified successor. It would not compel route acceptance. It would protect the recognised record. It would not create five contradictory registries. It would make custody contestable within one uniqueness framework.

The reform also clarifies legitimacy. A registry worthy of trust should not need preventable lock-in to retain members. It should retain them through accurate service, fair fees, reasoned decisions and proven resilience. When a member leaves, the institution should be able to hand over a complete record without treating continuity as its private asset.

Number Resource Society points toward that future by insisting that operators are principals, not merely subjects of a registry relationship. The next step is to convert that principle into audited infrastructure and restrained law. Portability, continuity, transparency and responsibility must arrive together.

The missing exit right is therefore not a demand for deregulation. It is a demand for a stronger form of order: one in which the record outlives the custodian, authority can be traced, service can be changed, and no operator must choose between institutional dissent and the continuity of its network identity.

Public sources