Need Base and Pre-Approval — The RIR Veto over IPv4 Transfers

Why ISPs and telecom operators should not need their recordkeeper’s permission to invest

An Internet service provider (ISP) does not engage a registry to become its investment committee.

Consider an operator expanding its network. Its engineers have designed the deployment. Its finance team has arranged the money. A seller is ready to transfer the IPv4 resources. The parties have identified the block, verified each other and agreed on commercial terms.

Then the registry asks the operator to justify its business need.

This is the issue addressed here: “need base,” or needs-based IPv4 transfer assessment, and the pre-approval processes built around it. The subject is not the authentication of a transfer. It is the requirement that a purchaser satisfy a Regional Internet Registry’s (RIR’s) assessment of how much address space it should acquire and when it should use it.

For the operator, the question is prior to processing speed or compensation: where did the recordkeeper acquire a right to make that investment decision?

My answer, under Running-Code Primacy, is that the registry function supplies no such right. Technical verification and commercial permission are different functions. The existence of the first does not establish the legitimacy of the second. That is the boundary developed in Notes 64 and 65 and stated in Article VI of Note 72, *The Bill of Rights of Uniqueness Coordination*.[^doctrine]

The first-principles test: change the forecast, hold the network constant

Begin with the actual coordination problem, not the institution’s preferred description of itself.

Number-resource records need to distinguish authorised changes from forged instructions and incompatible claims. Participants need reliable evidence of control, an auditable history and coherent security information. Those requirements concern whether a change can be trusted.

Now compare two proposed transfers. The prefix is identical. The parties and their authority are identical. The transfer instrument is authentic. There is no conflicting claim. The proposed routing and security transition are identical.

Only the buyer’s forecast changes: deployment within eighteen months in one case, deployment over thirty-six months in the other.

What changed in the uniqueness problem?

Nothing.

A needs assessment may produce a different commercial eligibility result. But the evidence authenticating the change of control has not become weaker. No second holder has appeared. No forged signature has become genuine. No routing-security defect has been cured by shortening a business forecast.

When a rule changes commercial eligibility without changing the relevant technical conditions, its justification cannot simply be “uniqueness.”

This does not erase history. RFC 7020 describes finite-pool management, hierarchical allocation and registration as distinct functions of the Internet Numbers Registry System. Operational need was part of that allocation framework.[^rfc7020] The question is whether that history establishes a continuing right to adjudicate purchases of resources already held by somebody else. It does not establish that right merely by repetition.

A rule for distributing an unallocated pool is a rule for deciding how that pool is issued. A secondary-market transfer is a different event. The registry is not supplying the seller’s consideration or financing the buyer’s investment. Applying an allocation-era test to it requires a separate justification.

My principle is narrower than a claim that every market transaction is wise: commercial judgment does not belong in the mandatory coordination layer unless the choice actually threatens a shared technical invariant.[^doctrine]

Fraud requires evidence controls. Conflicting claims require resolution. Binding legal restrictions require compliance by those subject to them. None of these creates a general registry mandate to decide whether an ISP’s expansion plan is sufficiently convincing.

 What “need base” means in the published rules

The differences between registries matter. “Need base” is not one identical rule applied everywhere.

APNIC makes the distinction explicit in section 11.1.3 of its Internet Number Resource Policies, APNIC-127, version 015, dated 20 February 2025. Transfer recipients must demonstrate a detailed plan for using the resources within twenty-four months. Existing holders must also show past usage and compliance concerning earlier delegations. Section 11.2.3 applies those recipient conditions to incoming inter-RIR transfers.[^apnic-policy]

ARIN separates the minimum initial transfer from larger acquisitions. Its Number Resource Policy Manual, section 8.5.5, provides a route to a larger initial or additional block through documentation of at least 50% use within twenty-four months. Section 8.5.6 requires at least 50% efficient utilisation of cumulative holdings for organisations with an ARIN allocation seeking additional addresses. Other provisions supply specific alternatives and exceptions. These are not simply checks that the seller authorised the transfer.[^arin-policy]

RIPE NCC’s October 2023 Resource Transfer Policies do not impose a general future-needs demonstration on ordinary intra-region transfers. Section 3.2 separately requires a plan to use at least 50% within five years for incoming transfers from regions requiring the receiving region to maintain needs-based policies. Other transfer restrictions remain; absence of a general needs test is not absence of rules.[^ripe-policy]

This variation does not prove that every restriction is unlawful. It proves something more fundamental to the technical argument: a general needs examination is not inseparable from maintaining a functioning transfer registry.

A needs test can also travel through inter-registry compatibility conditions. That is a condition attached to institutional cooperation, not evidence that a buyer’s forecast is part of an address’s technical identity.

The distinction matters to operators because our deployment plans do not become more or less authentic when the source of a block changes registry.

The historical record already contains the objection

This argument did not begin with the present note.

At APNIC 31 in 2011, Geoff Huston, explicitly speaking as an individual, warned that artificial barriers to registry entry could push transactions outside the registry and damage its accuracy. In the same transcript, Nigel Titley explained that needs-based conditions had been retained in the earlier RIPE transfer policy to secure its passage: “It was a concession.”[^apnic31]

That is evidence of a contested institutional choice, not a timeless engineering requirement.

In September 2017, APNIC discussed prop-118, which proposed removing needs assessment for transfers, subject to an inter-RIR exception. The Secretariat’s summary recorded arguments about recordkeeping, the distinction from free-pool distribution and disclosure of sensitive business information. It also recorded the opposing concerns: speculation, hoarding, abuse and fraud. The proposal did not reach consensus at that meeting and was ultimately abandoned in February 2019.[^prop118-history][^prop118-discussion]

The summary also recorded a defence that only one transfer had been refused for failure to demonstrate need in the preceding twelve months.[^prop118-discussion]

A low refusal count does not measure the cost of the process. It does not reveal the time consumed by successful applicants, the commitments rearranged, the applications reduced or the transactions never attempted.

Nor does a concern about fraudulent transfers answer the distinction between authenticating a seller’s instruction and evaluating a buyer’s demand forecast. A proposed control should be tested against the particular failure it claims to prevent.

The historical question is therefore not whether anyone noticed the boundary. The record shows that participants did. The unresolved question is why commercial demand review remained a condition of recording transfers.

An operator buys readiness, not only current utilisation

The registry’s forecast window and an operator’s investment horizon are not the same thing.

Consider a regional ISP preparing to serve a new enterprise customer. The customer wants evidence that capacity will be available before committing to migration. The ISP wants to acquire the resources before promising delivery. A needs assessment that insists on stronger evidence of committed demand can place the input behind the commitment that the input is needed to secure.

This is not proof that every application encounters that problem. It identifies a failure path: requiring the completed business as evidence for acquiring an input needed to build it.

The same logic applies to reserve capacity, staged network expansion and acquisitions made to avoid repeated procurement. Unused capacity today can have option value tomorrow. Its carrying cost gives the operator a reason not to acquire it casually. A registry’s utilisation threshold cannot, by itself, determine whether that option is worth its price.

The commercial downside remains with the operator. If the forecast is wrong, the operator holds excess inventory. If it is right but acquisition is delayed, the operator may lose the customer. The registry does not take either side of that investment position.

This is the point of Localized Future Decision: these commercial choices should remain outside the common validation rules, not be returned to the operator only after central approval.[^doctrine]

Willingness to pay is not a complete test of social benefit. It is evidence that an identified party is prepared to commit capital. An assertion of hoarding is not a complete test of harm either. Where there is actual fraud, market manipulation or anticompetitive conduct, the relevant facts and legal standards should be examined. Purchasing before deployment is not, by itself, proof of any of them.

There is a seller’s side as well. An operator selling surplus addresses may be converting idle capacity into funding for fibre, transit or equipment. Blocking the buyer can therefore delay two businesses’ capital decisions at once.

The seller may be a registry member too. The institution cannot establish that it protects members merely by restricting what one member may buy from another.

Pre-approval does not answer the authority question

Pre-approval moves an assessment earlier. It does not establish why the assessment belongs there.

APNIC offers recipients the option of having their needs evaluated before locating a seller. Its guide states that approval lasts twenty-four months and that a transfer within the approved size does not require the recipient to supply justification again.[^apnic-preapproval]

ARIN likewise offers pre-approval based on twenty-four-month projected need and says qualifying transfer requests submitted within that period will not be subject to re-verification of the needs assessment.[^arin-preapproval]

Those are useful commitments within the existing process. An operator should preserve them when making commercial arrangements. A later dispute about reopening an already settled question is different from the broader objection to mandatory needs review.

But even a perfectly honoured pre-approval leaves the underlying question untouched: why should the operator’s acquisition plan require the recordkeeper’s commercial assent?

There is also a distinction between a response deadline and a completion deadline. ARIN’s pre-approval page promises responses within two business days, including subsequent question-and-response cycles; that is not a promise to complete the whole assessment within two days.[^arin-preapproval]

A process can meet every response target while taking much longer to resolve. Whether a particular duration is justified depends on the actual requests and evidence. An operator cannot evaluate execution risk from response speed alone.

Faster permission is not the same product as freedom from an unnecessary permission requirement.

Membership does not make the registry the member’s principal

An RIR member has several separate relationships: with its customers, suppliers, investors and registry. Collapsing them into “community” obscures who made which commitment and who bears which loss.

A technical employee may be authorised to maintain registry records without being authorised to delegate the company’s procurement strategy. The scope of any authority has to be established from the actual appointment or agreement. Participation in discussion is not a substitute for that inquiry.

Equally, membership does not mean that every other member is entitled to approve an operator’s investment. Commercial competitors can share a technical reference system without sharing an investment committee.

The practical source of registry leverage is substitution. Can the operator obtain the same recognised update from another provider while preserving the resource and its continuity? Or must it abandon the transaction, replace the resources or accept a different operating position?

Buying a different asset is not the same as replacing the administrator of the asset already contracted for.

That distinction is central to the portability argument in Note 67: resource-level portability concerns the record and proof of control, not merely moving a company’s membership or legal address.[^portability]

A small institution can control an important dependency. Market power is assessed through competitive constraints and alternatives, not staff count. The Australian Competition and Consumer Commission explicitly identifies barriers to entry and freedom from customer or competitor constraints as relevant considerations.[^accc-market]

The ability to withhold an update establishes practical leverage. It does not, by itself, establish a legal entitlement to every condition imposed through that leverage.

Legal exposure: challenge both the conduct and the rule

The responsibility principle is straightforward: an institution that unlawfully obstructs commerce must answer for the recoverable economic loss its conduct causes. Australian competition legislation expressly provides a damages route for loss caused by contraventions of its competition provisions.[^cca-remedies]

A lawsuit still needs a cause of action, jurisdiction, causation and evidence. The architectural argument against needs-based approval is not a substitute for those elements. It identifies what should be challenged; law determines the available remedy.

There are two different disputes. One concerns conduct that departs from an applicable contract or undertaking. The other concerns a restriction that may itself be unlawful even when staff apply it exactly as written.

Contractual obligations and reliance

APNIC’s published membership agreement requires it to provide rights and services in accordance with its membership schedule and other APNIC documents. The agreement also selects Queensland law and courts, subject to its dispute-resolution document.[^apnic-agreement]

The member’s legal file should therefore identify the actual signed agreement, incorporated documents and specific approval. Did the registry demand something not required by those instruments? Did it reopen an assessment the applicable commitment had settled? Did it fail to perform an obligation that had become due?

These are questions of contractual construction and performance, not merely whether an applicant remained polite in a support ticket.

Contract damages ordinarily compare the claimant’s actual position with the position proper performance would have produced, subject to the applicable limits, including remoteness and mitigation. They are not automatically confined to the price of the administrative service.[^contract-damages]

Relying on an existing undertaking to mitigate exposure does not concede that needs-based approval is justified as the permanent architecture. Operators can enforce today’s obligations while disputing the design that made them necessary.

Market power and harm to competition

Section 46 of Australia’s Competition and Consumer Act 2010 prohibits conduct by a corporation with substantial market power where it has the purpose, effect or likely effect of substantially lessening competition in a market covered by the section.[^cca46]

The inquiry must identify the relevant market and its connection to the registry’s services. It must examine substitutes, entry barriers and the competitive consequences of the restriction. An operator’s inconvenience alone does not prove a competition-law violation. The ACCC makes the distinction clear: having market power is not itself unlawful, and refusal to deal is not automatically prohibited.[^accc-market]

A serious claim would ask whether the challenged conditions prevent otherwise capable operators from entering or expanding, exclude a competing supply model, or restrict resource mobility in ways that materially impair competition. These are hypotheses to test through transactions, customer evidence and market analysis—not findings established by this essay.

The relevant comparison is not “no coordination.” It is coordination with authenticated control and accurate records, but without the challenged commercial-needs restriction. That comparison tests whether the additional restraint contributes something necessary or simply restricts participation.

Where the statutory contravention and loss are established, section 82 permits damages proceedings. Section 80 supplies an injunction mechanism for relevant contraventions.[^cca-remedies] A member need not treat an internal policy discussion as the only conceivable place to contest a commercial restriction.

The exclusion clause is another question, not the answer

APNIC’s agreement seeks a broad exclusion of liability, to the extent permitted by law and subject to specified exceptions. It simultaneously requires members to indemnify it for losses resulting from their breaches.[^apnic-agreement]

The wording establishes the risk allocation the institution seeks. Its enforceability and application to a particular claim remain to be tested.

For eligible Australian standard-form small-business contracts, the unfair-contract-terms regime is particularly relevant. The post-November 2023 framework includes businesses with fewer than 100 employees **or** annual turnover below A$10 million, subject to the contract and commencement requirements. It examines significant imbalance, necessity to protect legitimate interests and detriment. A court considers the contract as a whole; an unattractive term is not automatically an invalid one.[^accc-contracts]

An operator should examine an alleged commercial veto together with the claimed ability to change its conditions, the available exit and the limitation of remedy. Their combined effect matters more than any clause read in isolation.

The defence cannot end at “you signed.” Equally, the claim cannot end at “we had no alternative.” Both must be tested against the actual agreement and applicable law.

Private technical institutions have faced liability before

The broader proposition that technical expertise or nonprofit status removes commercial conduct from legal scrutiny has already been tested elsewhere.

In *American Society of Mechanical Engineers v. Hydrolevel Corp.*, 456 U.S. 556 (1982), the United States Supreme Court held the nonprofit standards organisation civilly liable under antitrust law for its agents’ antitrust violations committed with apparent authority. A technical interpretation had been used against a competing product.[^hydrolevel]

In *Allied Tube & Conduit Corp. v. Indian Head, Inc.*, 486 U.S. 492 (1988), the Court rejected Noerr antitrust immunity for a manufacturer’s manipulation of a private standards process. The association’s influence and the adoption of its code by public authorities did not turn the challenged private conduct into protected governmental action.[^allied-tube]

Neither judgment determines the validity of an RIR needs test. They establish that private standards machinery is not a general legal safe harbour. An operator’s case must prove the relevant conduct, not borrow the outcome of a different dispute.

The absence of a successful challenge to a particular registry practice cannot be treated as a judgment upholding it. Non-enforcement is a circumstance, not an immunity.

The loss belongs in an evidence file

A registry delay should be analysed through the operator’s critical path.

Suppose an operator has committed US$2 million of borrowed funds that cannot be returned or deployed during an additional sixty-day hold. At an assumed annual financing rate of 10%, simple interest over that period is approximately US$32,877. This is an illustration, not an observed transfer loss or a current financing quotation.

A real claim would establish whether the money was actually tied up, what interest was incurred and whether the challenged conduct caused the additional period. Capital that remained freely available cannot be treated as trapped merely to enlarge the claim.

The same discipline applies to substitute-address costs, rescheduled implementation and lost customer business. Lost revenue is not automatically lost profit. The purchase price is not automatically a loss if the purchaser retains its money. Costs avoided and successful mitigation must be accounted for.[^contract-damages]

The decisive record connects four things: the requirements and evidence; the registry’s action and stated reason; the transaction that would otherwise have proceeded; and the incremental financial consequences.

Preserve the versions of the rules and approval relied upon. Separate time spent obtaining missing authenticating evidence from time spent answering commercial-needs questions. Establish when the seller, financing and downstream deployment were ready. Record independent obstacles as carefully as registry-created ones.

For a competition claim, add evidence of effects beyond one balance sheet: whether customer choice, entry or expansion was impaired. For a contract claim, identify the particular obligation breached. The claims can overlap factually without becoming interchangeable legally.

Most importantly, do not mistake an eventual approval for an absence of loss. A financing charge already paid or a genuinely lost commercial opportunity does not disappear when a ticket is closed.

The amount must be proved. The institution should not be presumed exempt from answering it.

Liability does not purchase a right to rule

There is a trap in stopping at compensation.

A registry could accept more liability while retaining the same unnecessary commercial veto. That would price some consequences. It would not cure the authority problem.

My position is not that RIRs may approve our business plans once they carry sufficient insurance. It is that business-plan approval does not belong in the common coordination function. Liability concerns harm already caused; architecture concerns removing the mechanism that can cause it again.

The three principles remain the design constraints. Minimum Initial Specification limits common rules to the verifiable requirements of uniqueness, control, security and interoperability. Localized Future Decision leaves investment, deployment timing and commercial arrangements outside those common rules. Voluntary Adoption prevents later institutional preferences from becoming compulsory merely through publication.[^doctrine]

The longer-term target is independently verifiable, portable state—not a replacement committee administering the same needs test. Ordinary validity should not require a standing institution’s opinion of an operator’s commercial plan.

That target carries engineering obligations. Authentication, conflicting histories, compromised keys, replay, secure succession and compatibility boundaries must be solved. Note 64 expressly recognises security and fragmentation risks; it does not promise that decentralisation removes every failure mode.[^doctrine]

During transition, existing registries can continue maintaining useful records and publication services. Those functions should remain separable from commercial adjudication. Replication of a database alone is not sufficient for RPKI succession: signing, certificate and relying-party continuity also require an operational transition. That distinction is developed in Note 70.[^continuity]

The objective is not to remove evidence. It is to remove the recordkeeper’s unnecessary discretion over what the evidence is allowed to describe.

The operator’s conclusion

For an ISP, telecom carrier or other network operator, the issue is not whether a needs-based approval department can become more agreeable. It is whether a coordination supplier should occupy the decision point at all.

We bear the cost of acquisition, unused capacity, deployment and customer delivery. The registry’s job is not to substitute its forecast for ours. Its useful role is to make control legible and transfers trustworthy without becoming the commercial principal.

Need base is not proof of control. Pre-approval is not the source of an operator’s right to invest. Membership is not a delegation of the member’s business to its recordkeeper.

Where actionable obstruction causes economic loss, the responsible institution should face the claim under ordinary law. That does not legitimise the obstruction once a price is attached to it.

Keep the evidence. Preserve uniqueness and continuity. Remove the commercial veto.

Our networks need reliable records. They do not need a registry’s permission to have a future.

Sources and historical record

The policy sources below establish published requirements and recorded debate, not adjudicated findings that a particular registry acted unlawfully. The legal authorities identify applicable tests and remedies; the analysis does not report a judgment invalidating RIR needs-based transfer assessment. Documents were reviewed on 18 September 2026. Dates below identify the document or event, rather than assuming that a current webpage describes every earlier transaction.

[^doctrine]: Lu Heng, [Note 64: Minimum Initial Specification, Localized Future Decision, and Voluntary Adoption]

 [Note 65: Running-Code Primacy]

[Note 72: The Bill of Rights of Uniqueness Coordination] 1 July 2026, especially Article VI. Authorial principles were taken from the supplied *Heng.lu Notes Full Export*, exported 3 July 2026; these principles are the analytical framework, not judicial holdings.

[^rfc7020]: R. Housley, J. Curran, G. Huston and D. Conrad, [RFC 7020: The Internet Numbers Registry System], August 2013, especially section 2. Describes finite-resource management, hierarchical allocation and registration; it is a historical technical description, not a judgment about private transfer liability.

[^apnic-policy]: APNIC, [APNIC Internet Number Resource Policies], APNIC-127, version 015, 20 February 2025, sections 11.1.3 and 11.2.3. Page marked active when reviewed.

[^arin-policy]: ARIN, [Number Resource Policy Manual], sections 8.5.2–8.5.7. The retrieved page displayed “Version 2025.1 – 3 March 2026”; both elements of that heading are preserved here without silently reconciling them. Initial minimum-size eligibility, larger/additional transfers and alternative criteria are separate provisions.

[^ripe-policy]: RIPE, [RIPE Resource Transfer Policies](https://www.ripe.net/media/documents/ripe-807.pdf), October 2023, sections 2 and 3.2. Archival identifier note: the retrieved URL ends in `ripe-807.pdf`, while the PDF cover prints “Document ID: ripe-806.” The policy is therefore identified here by its title, date, sections and retrieval URL. This source was visually checked on all three pages.

[^apnic31]: APNIC 31, Hong Kong, [Policy SIG Sessions 1 & 2 transcript](https://conference.apnic.net/31/31/policy/transcript.html), 2011, discussion of transfer needs requirements. Geoff Huston explicitly identified his intervention as an individual contribution. Nigel Titley’s quoted words concern the earlier RIPE policy compromise, not a legal ruling or a statement of all later RIPE policy.

[^prop118-history]: APNIC, [prop-118: No need policy in APNIC region](https://www.apnic.net/community/policy/proposals/prop-118/), proposal history: APNIC 44 discussion on 14 September 2017; abandonment on 28 February 2019.

[^prop118-discussion]: APNIC Secretariat, [Community discussion of prop-118](https://www.apnic.net/community/policy/proposals/prop-118/discussion-118/), summary of APNIC 44 discussion and mailing-list contributions, September 2017. Listed supporting and opposing points are recorded arguments. The one-refusal statement refers to the preceding twelve months in that historical discussion; it is not a current statistic or a measure of total commercial harm.

[^apnic-preapproval]: APNIC, [IPv4 Transfer Guide](https://www.apnic.net/manage-ip/manage-resources/transfer-resources/transfer-of-unused-ip-and-as-numbers/transfer-guide/), “Transfer pre-approval” and expiry provisions, reviewed 18 September 2026.

[^arin-preapproval]: ARIN, [Submitting a Transfer Pre-approval Request](https://www.arin.net/resources/registry/transfers/preapproval/), projected-need assessment, re-verification, review response times and approval duration, reviewed 18 September 2026.

[^portability]: Lu Heng, [Note 67: When the Water Company Says Your House Belongs to It](https://heng.lu/when-the-water-company-says-your-house-belongs-to-it/), 4 June 2026. Resource-level portability distinguished from membership migration in the supplied notes export.

[^accc-market]: Australian Competition and Consumer Commission, [Misuse of market power](https://www.accc.gov.au/business/competition-and-exemptions/misuse-of-market-power), reviewed 18 September 2026. Market power, competitive constraints, refusal to deal and restrictions on essential inputs.

[^cca46]: Australia, [Competition and Consumer Act 2010, section 46](https://www.legislation.gov.au/C2004A00109/2026-08-27/2026-08-27/text/original/epub/OEBPS/document_1/document_1.html), compilation dated 27 August 2026. The market-connection and substantial-lessening-of-competition requirements must be established for a particular claim.

[^cca-remedies]: Australia, [Competition and Consumer Act 2010, sections 80 and 82](https://www.legislation.gov.au/C2004A00109/2026-08-27/2026-08-27/text/original/epub/OEBPS/document_2/document_2.html), compilation dated 27 August 2026. Relevant injunctions and private damages proceedings; relief depends on statutory conditions and the facts.

[^apnic-agreement]: APNIC, [Membership Agreement](https://www.apnic.net/about-apnic/corporate-documents/documents/membership/membership-agreement/), clauses 3.1, 3.3, 5.2 and 5.3, reviewed 18 September 2026. This is the published template; a claimant must establish which executed agreement and incorporated documents govern its own relationship.

[^accc-contracts]: Australian Competition and Consumer Commission, [Contracts](https://www.accc.gov.au/business/selling-products-and-services/contracts), unfair contract terms, small-business coverage and changes commencing 9 November 2023. The commencement rules, standard-form requirement, statutory exclusions, territorial application and circumstances of each contract require separate consideration.

[^contract-damages]: Supreme Court of New South Wales, [paper on the measure of damages for breach of contract](https://supremecourt.nsw.gov.au/documents/Publications/Speeches/2025-speeches/20254CLQ3-5.pdf), *Commercial Law Quarterly*, March–May 2025, especially the discussion of the compensatory principle, the performance counterfactual, remoteness, mitigation and proof of loss. This is judicially hosted explanatory material, not an RIR case.

[^hydrolevel]: United States Supreme Court, [American Society of Mechanical Engineers, Inc. v. Hydrolevel Corp., 456 U.S. 556](https://supreme.justia.com/cases/federal/us/456/556/), 1982, especially pp. 565–576. Primary judgment reproduced by Justia. Addresses antitrust liability for agents acting with apparent authority; it is not a decision concerning an RIR.

[^allied-tube]: United States Supreme Court, [Allied Tube & Conduit Corp. v. Indian Head, Inc., 486 U.S. 492](https://supreme.justia.com/cases/federal/us/486/492/), 1988, especially pp. 499–511. Primary judgment reproduced by Justia. Concerns the limits of Noerr immunity for the challenged private standard-setting conduct, not a general rule invalidating technical standards.

[^continuity]: Lu Heng, [Note 70: The Registry Continuity Fallacy — Protect the Ledger, Not the Gatekeeper](https://heng.lu/the-registry-continuity-fallacy-protect-the-ledger-not-the-gatekeeper/), 24 June 2026. Registry-function continuity, dispute isolation and RPKI succession, from the supplied notes export.