The Promise of Uniqueness and the Claim of Ownership

Summary

• Uniqueness is a property of a coordinated ledger: at a given level and time, one number resource should not be allocated to more than one party. It prevents conflicting official entries but does not identify an ultimate owner.
• Registration, contractual use, operational use, transfer eligibility, route acceptance and property are distinct legal and technical relationships. Evidence about one cannot silently answer the others.
• The historical record from RFC 790 to RFC 7020 shows continuity in the coordination task, not a continuous chain of title vested in the institutions that maintained successive registers.
• Current RIR agreements use deliberately bounded language. RIPE NCC rejects ownership arising from registration, ARIN defines rights within its database and AFRINIC combines an exclusive right of use with an express denial that the resources are ordinary property.
• Economic scarcity can make a recognised registration position valuable. The value may attach to contract rights, transfer eligibility, operational continuity or a business transaction without converting the registry's uniqueness function into ownership of the underlying numbers.

One entry, six different questions

The attraction of ownership language is easy to understand. Internet numbers are finite within a protocol, organisations pay substantial sums in transactions associated with IPv4 registrations, and a registry can accept or reject changes to a recognised record. Those features resemble control over an asset. Yet resemblance is not classification. Before asking who owns an address block or an Autonomous System Number, the inquiry must identify which relationship is actually in dispute.

The first question is uniqueness. Does the recognised registry show one non-overlapping allocation at the relevant level and time? This is a statement about the integrity of a coordinated record. If two parties receive the same resource from the same authoritative branch, the system has failed at its central task.

The second is registration. Whose name or organisation appears in the authoritative record, under what status, and through which contractual or historical basis? Registration can be exclusive without being absolute. A land register, company register and telephone directory each record different interests; the fact of recording does not make every entry the same legal kind.

The third is contractual use. What may the registered party do under its agreement, and what services must the registry provide? These rights can include record maintenance, reverse DNS, routing-security support and a policy-compliant change of registrant. They are rights against identified counterparties, subject to the text, incorporated policies and governing law.

The fourth is operational use. Which organisation configures addresses on systems, announces the associated prefix, serves customers or authorises another network to act? The operational user may be the registrant, an affiliate, a customer, a lessee or a managed-service provider. The registry entry does not expose every downstream arrangement.

The fifth is transfer eligibility. Will the relevant registry recognise a change under applicable policy and update its records? Recognition can be necessary for a clean transaction, but the registry does not thereby become the seller or purchaser.

The sixth is property. Does a particular legal system recognise a proprietary, contractual, statutory, possessory or other protectable interest, and in what entity? That answer may vary by jurisdiction, agreement history and remedy sought. Uniqueness supplies evidence to this inquiry because exclusivity matters. It does not complete it.

Collapsing these questions gives every entity too much rhetorical power. A holder may present operational use as absolute title. A registry may present record authority as ownership. A buyer may present a paid transaction as a conveyance of the numbers themselves. A network may present route acceptance as proof of entitlement. Each may have a defensible interest, but none should obtain the other five by vocabulary alone.

RFC 790 shows coordination before the modern corporate structure

The historical starting point is useful precisely because it predates today's institutions. RFC 790, Assigned Numbers, published in September 1981, assembled assigned values for networks, protocols, ports and other technical fields. It credited the maintenance of those assignments to the central administrative arrangements of the period. The document looks unlike a modern RIR agreement, but the basic coordination need is recognisable: entities needed a common account of which values had already been assigned.

That record did not need an ownership theory to be useful. Its force came from shared technical practice and institutional recognition. If implementers independently reused numbers that were already assigned, communication would become ambiguous or fail. The register reduced that risk by making an accepted allocation visible. It was closer to a common coordination table than a warehouse containing the assigned values.

The distinction matters when later institutions describe themselves as successors in a global hierarchy. Continuity of function is not automatically continuity of property. A succession from one record keeper to another can preserve entries, operational expectations and administrative responsibilities without conveying ownership of every recorded value. To prove title, one would need an instrument and a legal theory identifying the asset, transferor, transferee and applicable law. A historical list does not do that work.

Nor can the 1981 record settle modern holder rights. Contemporary arrangements include regional companies, membership terms, service agreements, transfer policies, routing-security services and diverse legacy positions. Many did not exist in RFC 790's world. Reading current contractual categories backward would conceal the institutional changes that require explanation.

The defensible historical conclusion is narrower. Assigned-number administration began from a practical requirement for non-conflicting values. The early register supports the ancestry of the uniqueness function. It does not show that every later administrator inherited ownership of the entries, or that every listed recipient acquired identical rights under a timeless global rule.

This is not to diminish the administrator. Maintaining a trusted register is a form of power. Decisions about eligibility, documentation and changes can shape access to scarce resources. But administrative power should be described by the authority that actually supports it. In 1981, that support included technical coordination and the institutional arrangements of the early Internet. The later corporate and contractual settlement needs its own evidence.

RFC 7020 defines the promise with precision

The clearest modern statement appears in RFC 7020, The Internet Numbers Registry System, published in August 2013. It describes uniqueness as ensuring that an IP address or AS number is not allocated to more than one party at the same time. The temporal and relational limits are important. The promise concerns allocation within the registry system, not metaphysical possession of a number forever.

The document also describes registration as providing accurate information about the party to which a resource has been allocated or assigned. Accuracy supports troubleshooting, coordination and policy administration. It makes the record useful to parties beyond the direct holder. Yet a reliable statement about allocation remains a statement about allocation.

RFC 7020 expressly separates route announcement from registry administration. Whether addresses are announced, and how advertisements are made, sits outside the Internet Numbers Registry System. That boundary defeats one common ownership shortcut. A registry can record one party without commanding every network to accept that party's route. Conversely, a network can observe an announcement without proving that the announcer holds the recognised registration.

Uniqueness therefore operates at a defined layer. It reduces the chance that two parties will receive conflicting official claims from the same coordinated hierarchy. It helps networks and service providers plan without negotiating the identity of every number from first principles. It supports reverse DNS and routing-security services that refer back to registered authority. But it does not turn the hierarchy into the operator of all networks.

The document also recognises that the interests of registry users and network operators may not always align. That observation is incompatible with an account in which the registry's entry conclusively determines every operational question. If interests can diverge, governance needs mechanisms for policy, review and coordination rather than a simple assertion of dominion.

The exact promise can be stated as follows: the system will strive to maintain one accepted allocation position for a resource at one time, with accurate associated information, under the rules of its hierarchy. That is already a major institutional commitment. Expanding it into ownership is unnecessary and analytically costly. It converts a testable duty of record integrity into a contested legal claim that the technical description neither needs nor makes.

Hierarchy distributes responsibility, not a single proprietary estate

The IANA Number Resources description presents a hierarchy. Global pools are coordinated at the IANA functions level; large blocks are made available to the RIRs; regional systems may work through national or local registries, providers and other intermediaries; users ultimately receive number resources through these arrangements. The hierarchy scales administration and preserves non-overlap.

It is tempting to read the diagram as a chain of ownership descending from a global proprietor. The language and function support a different reading. Each level records, allocates or assigns within a defined scope. The parent level must avoid conflicting delegations to its children; each child must avoid conflicts within its own service region or customer base. A coherent result emerges from nested record duties.

This resembles delegated administrative competence more than repeated sale. When IANA records a block to an RIR, the block is not consumed as goods would be. When an RIR records a smaller range to a holder, the higher-level record remains. The entries identify responsibility and recognised allocation at different levels. They do not necessarily describe successive transfers of the same legal title.

The hierarchy also explains why an exclusive registration has practical weight. Other entities accept the coordination structure because competing official records would undermine interoperability. An RIR's recognition is therefore not merely private opinion. It occupies a singular place in the accepted system for its region. A holder can reasonably attach value to being the one recognised party for a particular block.

That value still does not prove registry ownership. An institution can possess authority to maintain a record without owning what the record concerns. Corporate registrars can recognise companies without owning them. Standards organisations can maintain identifier spaces without owning every product that carries an identifier. Analogies are imperfect, but they expose the missing premise in the proprietary claim.

What the hierarchy does prove is stewardship responsibility. A mistake at one level can create conflicting signals below it. A delayed correction can burden operators and counterparties. A disputed change can affect services tied to the record. The appropriate consequence is stronger accuracy, chain-of-custody controls, review and continuity. Calling the hierarchy an owner may obscure those measurable duties behind a broad label.

Institutional boundaries matter as much as technical ones

RFC 2860, published in June 2000, records a memorandum concerning technical work of the IANA between ICANN and the IETF. Its value here lies in its limits. It distinguishes the covered technical tasks from general policy issues concerning IP addresses, and it denies that the parties become partners, joint venturers, principals or representatives of one another merely through the memorandum.

Those clauses do not define the entire numbers system. They do demonstrate why institutional relationships must not be inferred from functional proximity. Two organisations can cooperate on globally important technical administration without becoming one legal person or acquiring all of one another's powers. A document may authorise one set of tasks while reserving another set to different arrangements.

The same discipline should govern ownership claims. ICANN's institutional connection to the IANA functions does not by itself make ICANN the owner of all Internet numbers. IANA's global coordination role does not by itself make every RIR its property manager. Recognition of an RIR does not by itself make holders tenants. Each proposition requires evidence from the relevant instrument and legal setting.

This is especially important because the public often uses IANA, ICANN and the RIRs as if they were interchangeable names for a single registry. Operationally, they participate in a connected hierarchy. Legally and corporately, they are distinct institutions with different documents, constituencies and duties. The hierarchy's coherence should not be mistaken for unity of personality.

The bounded scope of RFC 2860 also warns against using a technical memorandum to settle general address policy. If the instrument excludes that subject from its own scope, it cannot serve as a hidden conveyance of proprietary authority over it. Its actual contribution is to show an agreed division of technical responsibility and an express refusal to create broader status relationships.

Institutional legitimacy improves when these lines are visible. Parties can ask which body made a decision, under what authority, with what review and against which counterparty. Ownership rhetoric tends to merge the bodies into a single sovereign. The documents instead reveal a plural system held together by coordinated functions and accepted records.

Recognition of an RIR is authority to coordinate, not a deed

ICP-2, accepted on 4 June 2001, sets criteria for establishing new Regional Internet Registries. It addresses service region, community support, neutrality, impartiality, technical competence, continuity, procedures and financial stability. It also reflects a model of one recognised RIR for a large geographic region.

Recognition has substantial consequences. A successful institution receives a defined place in the global distribution hierarchy. Other registry entities coordinate with it. Holders and networks rely on its records and services. The status is not equivalent to that of an ordinary commercial directory that anyone can replace by opening a competing site.

Yet the criteria describe fitness to perform a public-interest coordination role. They do not convey ownership of the region's unallocated or allocated number resources to the registry company. Their objectives include fair distribution, conservation and route aggregation. These are administration goals. They define how a recognised institution should act, not what proprietary estate it acquires.

The difference has governance consequences. If recognition were ownership, institutional discretion might be framed as the exercise of a proprietor's private choice. If recognition is stewardship authority, discretion remains tied to purpose: accurate administration, impartial treatment, continuity and community-supported policy. The latter account fits the criteria more closely and offers a clearer standard of review.

One RIR per region also intensifies the need for precision. A holder cannot readily switch the authoritative regional record to a competitor when it disagrees with a decision. That singularity creates dependency and gives the record economic significance. But market power arising from recognition is not proof of title. It is a reason to demand transparent procedures and bounded authority.

Recognition should therefore be understood as conferring a role in the accepted hierarchy. The institution may set service terms within its powers, apply community policy, maintain records and recognise eligible changes. It may have contractual remedies when holders breach agreements. None of these powers needs the proposition that the company owns the numbers. They can be explained by delegated function, contract, policy and system-wide coordination.

RIPE NCC states what registration does not do

The RIPE NCC Standard Service Agreement, ripe-812, dated November 2023, is unusually direct. It states that registration of Internet Number Resources does not constitute property and does not confer ownership. The agreement nevertheless creates a substantial relationship. It identifies the services, incorporates policies and procedures, places duties on the member and provides for suspension, termination and cooperation with deregistration in stated circumstances.

This combination is analytically valuable. It shows that an institution can exercise consequential record and service powers without claiming that registration transfers title. The powers arise from the agreement, the association's role and the policies governing the service. The member's position can be exclusive and operationally valuable while remaining bounded by those arrangements.

The denial of ownership should not be overstated. A contract between RIPE NCC and a member cannot conclusively classify every conceivable interest against every third party under every legal system. Mandatory law may affect characterisation. A legacy history may complicate which terms apply. A court may protect a contractual position or grant relief without declaring the resource itself ordinary property.

Nor does non-ownership mean non-value. A member with an established registration, infrastructure and customer relationships may have a powerful commercial position. The right to continued services subject to the agreement can be important. Eligibility to transfer a registration under policy can attract payment. Loss of recognition can have serious consequences. None of those facts is made unreal by refusing the language of ownership.

The agreement instead encourages a more exact account. The resource remains part of a coordinated number space. RIPE NCC maintains the authoritative regional entry and associated services. The member receives contractual service and a recognised registration position. Networks make independent routing decisions. Downstream users may receive service under separate contracts. Transfer policy governs whether a change of registration will be accepted.

This layered description is less rhetorically satisfying than saying either "the member owns it" or "the registry owns it." It is more useful. It identifies which party can change which relationship and which evidence would be needed in a dispute. It also prevents the no-property clause from becoming a claim that the member has no protectable interest at all.

ARIN locates the right inside the registry relationship

The ARIN Registration Services Agreement, version 14.0, dated 15 August 2025, uses a carefully bounded vocabulary. It describes an exclusive right to be the registrant of included number resources within the ARIN database, a right to use those resources within that database, and a right to transfer the registration under applicable policy. The repeated connection to the database and services matters.

An exclusive database position is not trivial. ARIN's record is relied on as the recognised regional account. Associated services include record maintenance, reverse name service and RPKI. A buyer seeking a clean transfer wants ARIN to recognise the resulting registrant. A holder facing an adverse change has more at stake than a line in an informal contact list.

At the same time, the text does not need an absolute property claim. The agreement specifies what ARIN promises and what the holder may do in that relationship. It can define conditions, representations, service limits, suspension and termination. The right is intelligible as a contractually protected registration position with operational consequences.

This drafting exposes a recurring error in transaction language. When parties say that an address block was bought or sold, they may be describing the economic result rather than the legal entity. The transaction might include a policy-compliant change of registered holder, contractual promises by the former holder, corporate assets, customer arrangements, historical records or technical cooperation. Price alone does not reveal which component carried the value.

The current ARIN form also cannot answer every historical case. Legacy registrations may have different documentation. Earlier agreements may apply. Corporate reorganisations and bankruptcy proceedings can introduce additional law. A universal conclusion based on version 14.0 would exceed the evidence.

What the agreement does establish is a useful model of precision. Exclusivity can be stated without treating the database operator as owner. Use can be defined within a service relationship without pretending to command route acceptance. Transfer can mean recognised registration change under policy without becoming proof that the number itself is an ordinary chattel. This is not evasive drafting; it is an attempt to match legal words to institutional control.

AFRINIC combines exclusive use with an express property denial

The AFRINIC Registration Service Agreement, dated 27 November 2017, offers another revealing combination. It describes an exclusive right of use during the agreement while stating that the number resources are not real, personal or intellectual property. The two clauses can coexist because exclusivity and property are not synonyms.

A contract can promise that one applicant will be recognised for use under defined conditions. That promise can be enforced as a contractual interest even if the subject is not classified as owned property. Licences, memberships and regulated permissions often create valuable exclusive positions without transferring an underlying thing. The correct analogy depends on the law and facts, but the general distinction is ordinary.

The duration language is equally important. An exclusive right during an agreement is temporally bounded. Termination, policy compliance and service conditions affect the position. This does not mean the registry may act arbitrarily; contractual powers remain subject to their terms, governing law and applicable duties. It means the source of the right has been identified more precisely than a permanent ownership assertion would allow.

The express denial of real, personal and intellectual property also closes several misleading routes. An IP address is not land merely because it has an address-like notation. It is not a physical entity possessed like equipment. It is not automatically a copyright or patent right. Scarcity and transfer value do not force it into one of those classes.

As with RIPE NCC, the agreement cannot decide all possible legal interests worldwide. A court might recognise a protectable contractual entitlement, an expectation relevant to insolvency, or another form of relief. The point is not that property analysis is forbidden. The point is that it must identify the jurisdiction, the asserted interest and the remedy rather than treating operational exclusivity as a universal answer.

AFRINIC's wording also demonstrates why the registry need not claim ownership to enforce policy. It can rely on the service agreement and its recognised administrative role. The holder can rely on the promised exclusive use and procedural protections. Networks can continue to assess routes independently. Keeping these relationships separate makes authority easier to test.

Routability is evidence of operation, not title

Routing creates the most visible temptation to collapse the layers. If a prefix is announced and widely accepted, the announcing network appears to control it. If a route disappears after a registry-related change, the registry appears to control it. Neither observation is enough on its own.

An announcement shows that a network originated or propagated routing information and that other networks accepted it under their own policies. It does not necessarily show authority from the registered holder. Route leaks and hijacks are possible precisely because operational announcements and recognised registration can diverge. A technically successful route can therefore be unauthorised.

The reverse divergence also occurs. A party may remain the recognised registrant while its route is filtered, misconfigured or withdrawn. Connectivity can fail because of equipment, transit, security policy or human error without any change to the authoritative record. Registration is not a guarantee of global reachability.

RPKI adds a powerful signal but does not erase the distinction. A Route Origin Authorisation can help relying networks evaluate whether a particular origin AS is authorised for a prefix. Operators decide whether and how to use validation states. A registry's certificate services can affect the available signal, yet the final routing choice remains distributed.

These facts support two simultaneous conclusions. First, registries should not claim that record authority equals direct control of every route. Second, the operational consequences of registry-linked services can be serious because many networks voluntarily rely on them. Accurate analysis follows the causal chain rather than choosing one extreme.

In an ownership dispute, route observations may prove practical use, continuity, reliance or harm. They can show which network served customers and when. They do not alone prove the legal source of entitlement. In a registry dispute, the record may prove recognised status but not who actually operated the network. Both forms of evidence are necessary when the case crosses layers.

The separation protects the wider Internet. If route acceptance automatically created title, a successful hijacker could convert technical abuse into legal entitlement. If registry status automatically compelled routing, network operators would lose essential autonomy. The system works because authoritative coordination and distributed operation interact without becoming identical.

Transfer markets price a position, not necessarily the numbers themselves

IPv4 scarcity has produced transactions in which parties pay for access to recognised address space. The commercial language often speaks of buying and selling addresses. That shorthand describes a real exchange of value, but it does not settle the legal entity transferred.

At minimum, a successful transaction generally requires the former holder's cooperation and a registry-recognised change consistent with applicable policy. It may require evidence of corporate authority, clean records and the absence of conflicting claims. The recipient wants a registration position that counterparties will recognise and that can support continued operations. Those attributes are scarce and valuable.

The price may also reflect features outside the registry entry: the reputation of the address range, existing routing history, geolocation treatment, customer contracts, equipment, a business acquisition or the cost of renumbering. Without transaction documents and allocation of consideration, a headline price cannot be attributed entirely to the abstract numbers.

Economic value is therefore relevant but underdetermined. It rebuts any suggestion that registration is clerical trivia. It does not prove that IANA or an RIR owns the resources, nor that the holder owns them in the same way as land or hardware. Markets routinely price contract positions, licences, memberships, priorities and expectations.

Transfer policy gives the registry meaningful power. If recognition is withheld, a contemplated transaction may lose much of its practical value. That control should be exercised under clear criteria, reasoned decisions and review. But a gatekeeper to recognised change is not necessarily the proprietor of the subject. A corporate registrar can refuse a defective filing without owning the shares being transferred.

The safest description of a transaction is factual. Identify the registered block, the parties, the applicable agreement and policy, the record change, any associated corporate assets, the operational transition and the price if reliably disclosed. Then identify the legal characterisation advanced by the parties or adopted by a court. Do not replace that evidence with the word sale.

A global denominator remains absent. There is no single public inventory classifying every resource as legacy, directly contracted, sponsored, leased, downstream-used or previously transferred, with the governing instrument and jurisdiction attached. That absence makes universal property statements especially weak.

Legacy status prevents one contract from answering every case

The modern agreement model can create an illusion of uniformity. A current form is easy to find and read, so its vocabulary is projected onto every registration. The history is more complicated. Some resources were recorded before the present RIR companies and policies. Some holders later signed agreements; others may have entered different arrangements. Transfers, mergers and service relationships can alter the evidence.

Legacy status does not itself prove ownership. Age is a fact about chronology, not a legal classification. An early entry may support continuity and reliance. It may affect which agreement applies or what representations were made. It does not automatically establish that the original administrator conveyed permanent title.

Nor should a current registry dismiss historical interests merely because contemporary forms use different words. If an institution claims authority over an old entry, it should identify the succession, policy basis, accepted terms and procedural protections supporting the action. Institutional continuity cannot be invoked selectively: preserving the authority of the record while ignoring the expectations attached to its history.

The right method is resource-specific. Establish the original entry if available, later changes, corporate succession of the holder, agreements accepted, fees or services, transfer history, operational use and relevant communications. Determine which law and forum govern the contested relationship. Only then can the nature of the interest be assessed.

This approach may produce different answers for different resources. That is not a defect. A global technical identifier system can coexist with varied legal relationships. The uniqueness rule supplies a common coordination baseline, while contracts and law supply more particular rights.

The missing global inventory is a governance problem. Registries can protect confidential details while publishing aggregate counts by status and contract family. They can explain how legacy records are treated for services, transfers, disputes and succession. Better classification would reduce opportunistic claims by both holders and institutions.

Without that evidence, two categorical statements should be rejected: that every old holder owns its block outright, and that every current registry may treat every old entry solely under its newest standard form. Both conclusions skip the documents that would make them defensible.

Downstream use complicates every bilateral account

The registered holder is often not the only organisation using a number resource. Providers assign addresses to customers. Corporate groups centralise registrations while affiliates operate networks. Hosting companies, cloud services and managed operators may place many businesses behind a registered range. Leasing or sponsorship arrangements can separate record status from day-to-day control.

This structure explains why ownership language can mislead even within one branch. If the holder says it owns the block, what is the customer's interest? If the customer pays for exclusive use of a range, what survives termination of the service contract? If the registry changes the holder's status, what happens to parties that never contracted with the registry?

The uniqueness function answers only the highest relevant record question. It helps ensure that the recognised holder's allocation does not conflict with another allocation. It does not document every downstream permission. Those permissions arise from separate agreements and technical arrangements.

Downstream reliance nevertheless matters to registry governance. A record change can affect contact data, reverse DNS, security signals or transfer expectations used by many parties. An institution should consider foreseeable effects, preserve continuity where possible and provide rapid review. That responsibility arises from consequential administration, not ownership of the customers' networks.

The holder also has duties. It should maintain accurate records, secure credentials, document delegated use and avoid representing customer access as a stronger right than the contract supplies. Customers should understand whether they receive portable addresses, provider-dependent service or another arrangement. Clear language at each level reduces disputes when a relationship ends.

Evidence should follow the chain. Registry documents establish recognised status. Holder-customer contracts establish downstream rights. Configuration and routing data establish operational use. Payment and transaction records establish economic exchange. No single item substitutes for the others.

This chain also prevents a false binary. The choice is not only between registry ownership and holder ownership. Several parties can possess different enforceable interests in the same coordinated resource relationship: the registry's administrative authority, the holder's registration rights, the operator's technical control and the customer's service rights. Their interests can conflict without any one party owning every aspect.

Scarcity strengthens accountability but does not decide classification

IPv4 scarcity changes bargaining power. When replacement space is costly and renumbering is disruptive, a registry decision or failed transaction can impose substantial losses. Scarcity makes recognised registration positions valuable and makes governance more consequential.

Yet scarcity is not a legal category. Rare concert tickets, spectrum permissions, domain registrations and contractual options can all be valuable under different legal structures. The fact that people pay does not tell us whether the subject is property, a licence, a contract right or a bundle of interests.

Scarcity does establish why dismissive language is inadequate. A registry cannot reasonably say that a record is merely administrative when its policies make recognised transfer a condition of accessing a scarce position. The institution's power should be matched by clear criteria, secure procedures, reasons and review.

Likewise, a holder cannot use scarcity to erase public-interest conditions. The number space remains a shared coordination environment. Conflicting allocations would harm others. Accurate contact and security information has wider value. Transfer rules may protect integrity and prevent duplicate claims. A protectable interest can remain conditional and relational.

The policy task is to separate legitimate stewardship from proprietary overreach. Conservation rules should be justified by the technical and distribution goals they serve. Transfer controls should target record integrity, eligibility and non-conflict. Fees should fund services and governance transparently rather than imply rent extracted by an owner. Enforcement should be proportionate to the actual breach.

As IPv4 becomes more valuable, empirical transparency also matters. Registries should publish aggregate transfer counts, processing times, rejection grounds, reviews and reversals. Researchers need transaction evidence that separates registration value from attached businesses and infrastructure. Courts and parties need the exact agreement in force.

Scarcity therefore raises the standard of institutional conduct. It does not supply the missing deed. A registry can be a powerful steward of a scarce coordination position without owning the underlying numbers; a holder can have a valuable and protectable position without possessing absolute title.

A disciplined dispute matrix prevents category errors

Any serious dispute should begin with a matrix rather than a slogan. The first row identifies the resource and level: global pool, RIR allocation, holder registration or downstream assignment. The second identifies the relevant date, because records, contracts and policies change.

The next rows separate parties and capacities. Which institution maintained the record? Which company signed the agreement? Which organisation was shown as registrant? Which network originated routes? Which customers used services? A group name or brand may conceal several legal persons.

The documentary rows then identify the historical entry, current registry record, governing agreement, incorporated policy, transfer request, notices and decisions. Each document should be used only for what it can establish. A public record may prove registration but not contract acceptance. A route collector may prove observation but not authority. A payment may prove value but not the entity conveyed.

The technical rows identify reverse DNS, RPKI state, route origin, propagation and service impact over time. They establish consequence and causation. They should be dated closely enough to distinguish a registry action from unrelated operational changes.

The legal rows identify the claimed interest and remedy. Is the party seeking correction of a record, continued service, recognition of a transfer, damages, an injunction, insolvency treatment or return of control? Different remedies can depend on different classifications. A court may protect a contractual right without deciding universal ownership.

Finally, the matrix records uncertainty. Missing agreements, incomplete legacy histories, undisclosed downstream contracts and absent technical traces should be stated. Uncertainty is not a victory for either side. It limits the conclusion and identifies the evidence needed next.

This method turns an ideological contest into a sequence of answerable questions. It may reveal that the decisive issue is identity verification rather than property, transfer policy rather than title, or route causation rather than registration. It also makes review possible because each conclusion is tied to a dated fact and a defined relationship.

Better registry language would improve legitimacy

Institutions should describe their authority in terms that match their functions. Public materials can say that the registry maintains the recognised regional record, provides defined services, applies community policy and prevents conflicting allocations. They can explain that networks make independent routing decisions and that downstream use may be governed by separate contracts.

Agreements should state positively what the holder receives, not rely only on denials of property. Exclusive registration, service continuity, correction rights, transfer eligibility, notice, review and restoration can be described with precision. A bare statement that resources are not property may protect against one misconception while leaving members unsure what is protected.

Transfer materials should distinguish a change of registered holder from sale of an underlying business or infrastructure. Public statistics should identify what the registry measured. Incident communications should separate account status, registration status, certificate state and observed routing effects.

Historic and legacy positions deserve a clear account of applicable terms. If the registry relies on a later agreement or policy, it should identify how that instrument became binding in the case. If a holder claims immunity from current administration, it should identify the historical basis and subsequent conduct supporting that claim.

The global hierarchy can also state institutional boundaries more clearly. ICANN, the IANA functions and the RIR companies cooperate within a recognised system but are not one undifferentiated legal body. Naming the responsible institution and instrument improves accountability.

None of this requires a universal renunciation of property analysis. Courts and legislatures may recognise interests appropriate to their systems. The discipline is to avoid presenting uniqueness as if it had already made that judgment. Registry legitimacy is strongest when the institution relies on demonstrable authority and accepts review within its actual sphere of control.

Clear language also benefits members. It protects valuable registration positions from being dismissed as nothing, while preventing those positions from being oversold as unconditional ownership. It tells customers what depends on the holder and what depends on the registry. It gives operators a reason to treat registry data as trusted evidence without confusing it with a command.

What remains unknown should stay visible

The available documents establish the architecture and several contractual choices. They do not provide a complete global classification of resources by legacy status, agreement version, operational user, downstream arrangement, transfer history and jurisdiction. That missing denominator prevents confident statements about how common any legal position is.

There is also no complete cross-jurisdiction digest in the materials considered here that classifies registration, use and transfer interests under every relevant law. A conclusion reached in one proceeding may turn on its remedy, parties and local doctrine. It should not be projected globally without analysis.

Transaction evidence is incomplete. Publicly reported prices may combine address scarcity with corporate assets, reputation, contracts and technical assistance. Reliable comparison requires the actual transaction perimeter and consideration allocation. Without it, market value proves demand for a recognised position but not the legal character of the thing priced.

Operational evidence is similarly fragmented. To understand disputed changes, researchers need longitudinal registry status, certificate state, reverse DNS and route acceptance. A record change may be highly consequential in one case and operationally quiet in another. The causal path should be observed, not assumed.

These gaps do not leave the analysis empty. They support bounded findings. Uniqueness has a clear technical-administrative meaning. Routing remains distributed. Published agreements use distinct vocabularies and often reject simple ownership. Recognition criteria focus on capable, neutral and continuous service. Economic value exists around exclusive registration positions.

The gaps also suggest a public research programme: aggregate status inventories, contract-family counts, transfer evidence, anonymised dispute outcomes and technical timelines. Better evidence would allow institutions to calibrate rights and remedies without forcing every controversy into the ownership frame.

Until then, categorical claims should carry the burden of proof. A registry asserting ownership should identify the instrument and law. A holder asserting absolute title should do the same. A network asserting entitlement from route acceptance should establish authority. The uniqueness record remains central evidence, but it is not the final verdict.

The promise should remain both strong and narrow

The Internet needs an accepted answer to a practical question: which party has the recognised allocation of a number resource at a given level and time? Without that answer, duplicate claims would increase, coordination costs would rise and security signals would become harder to interpret. The promise of uniqueness is not modest in consequence.

Its legitimacy depends on remaining narrow in claim. The record keeper should not convert the duty to prevent collisions into ownership of every entry. The holder should not convert exclusive registration into unconditional dominion over routing, downstream users and future transfers. Operators should not convert route acceptance into proof of legal entitlement.

The historical line from RFC 790 to RFC 7020 supports a durable coordination function. The IANA hierarchy explains how that function scales. RFC 2860 shows that connected institutions can preserve legal boundaries. ICP-2 recognises regional administrators for competence, neutrality and continuity. The RIR agreements then define particular service and holder rights with language that is deliberately more bounded than absolute title.

Together, these materials support a plural account. IANA coordinates global pools. RIRs maintain recognised regional records and services. Holders possess registration and contractual interests that can be exclusive and valuable. Operators decide routes. Customers receive downstream rights. Transfer policy governs recognised changes. Courts may classify particular interests when a remedy requires it.

This account is not weaker than ownership rhetoric. It is more demanding. Every institution must identify its authority; every holder must identify its right; every operational claim must show its technical evidence. Power cannot hide inside a metaphor.

Uniqueness should therefore be defended as a public-interest commitment to accurate, non-conflicting coordination. Its stewards deserve authority sufficient to keep the record coherent, and they owe procedures proportionate to the value others place on that record. The holder's recognised position deserves protection appropriate to its contract, history and law. Neither conclusion requires a fictional universal owner.

The most credible registry is not the one that claims the greatest dominion. It is the one that can explain, for every consequential act, which record it maintains, which rule authorises the act, which interest is affected, how the decision can be reviewed and what remains outside its control. That is the institutional promise behind uniqueness. Ownership, where it matters, must still be proved.