Summary

  • AFRINIC's public Board register limits Resolution 201411.214 to “large resource allocation applications”, while the contemporaneous minute-book wording covers “resource allocation applications” without that qualifier. The surviving official record supplies no basis for choosing one formulation as secretly controlling or for inventing a reconciliation.
  • The defensible purpose of authentication is to verify a fact needed for an accurate registry decision. Discretion over “any available methods” can accommodate changing documentary and cross-border risks, but it does not create unrestricted discretion over outcomes, nor does it confer police, punitive, confiscatory, regulatory or adjudicative power.
  • A legitimate system would make the selection rule, proof question, ownership and timetable intelligible while protecting sensitive anti-fraud tactics. It would record evidence and contradictions, distinguish the authentication finding from the allocation decision, permit correction, provide independent reconsideration and audit delays, false positives and reversals.
  • The contemporaneous discussion shows why verification was on the Board's agenda, but it proves neither fraud nor a dispute. The applicant, application size, resources requested, evidence considered, methods used and ultimate outcome all remain undisclosed and must remain unstated.

A resolution with two edges

The first edge of Resolution 201411.214 is easy to understand. A number registry cannot maintain a reliable allocation record if it must accept every assertion or document at face value. An application can contain an innocent error, an outdated corporate record, a representative whose authority is unclear, an internal inconsistency or a document whose provenance needs confirmation. Scarcity raises the consequence of getting those facts wrong. Once a unique number resource is recorded against the wrong party or on a defective basis, the error can impose costs on other applicants and on the operators who rely on the registry's records.

Verification is therefore not an ornamental task. It belongs to the ordinary discipline of keeping a technical ledger accurate.

The second edge is institutional. The power to decide how a fact will be checked can influence whether an applicant ever reaches a substantive allocation decision. If the trigger for extra checking is unclear, if the requested proof changes without explanation, if contradictory material is never put to the applicant, or if the process has no end date, authentication can become a holding pattern. No formal rejection is needed. Delay itself can disrupt deployment, customer commitments, financing and planning. The distinction between verification and a veto then survives in words but disappears in operation.

That possibility should not be confused with proof that it happened in the case before the Board. The official record does not disclose a delay, rejection, adverse finding, appeal or outcome. It does not reveal the applicant's identity, corporate form, address family, quantity requested or entitlement. It gives no licence to reconstruct a scandal from silence. The institutional task is narrower and more useful: identify what the delegation allowed, identify what it left unresolved, and design the controls that would keep a valid authentication function within its proper boundary.

What the November record establishes

The chronology is compact. A 29 October action-item review carried a resource-abuse and allocation issue forward to the November face-to-face Board meeting. On 24 November, the agenda recorded a presentation on the number-resource allocation process by Registration Services manager Madhvi Gokool. The minutes say that the presentation happened; they do not preserve its contents. It would be unsafe to attribute any particular procedure, concern or recommendation to that presentation.

On 25 November, the Board discussed whether to issue a statement concerning one particular resource application. The minutes record that Registration Services had encountered difficulties and that a third-party statement featured in the discussion. They also attribute to legal counsel Ashok Radhakissoon the view that there was not yet a dispute, and record the chief executive taking the same present position. The minutes say that the chief executive had already confirmed that staff could allocate the resources and reminded the Board that it needed to stay within policy.

Those details locate the decision. They do not prove the truth of the third party's words, the nature of the staff difficulty, falsity by the applicant, fraud, or even the existence of a concluded dispute. Nor does the statement that staff could allocate reveal whether allocation occurred, on what date, in what amount or on what evidence. The discussion explains why authentication had become an immediate governance question. It cannot carry more factual weight than that.

The minutes then record unanimous adoption of Resolution 201411.214 on 25 November. Unanimity establishes the Board decision as recorded. It does not disclose a director-by-director rationale, eliminate the need to construe the delegation narrowly, or validate an unknown future exercise. A unanimous grant of bounded discretion remains bounded; the vote count cannot supply definitions and safeguards absent from the text.

The unresolved word: large

The public Board register places the resolution under November 2014 and says the Board delegated authority to the chief executive to use any available methods, at his discretion, to authenticate “large resource allocation applications”. The minute book records a delegation to authenticate “resource allocation applications”. The difference is one word, but the word determines the apparent reach of enhanced authentication. Under the register's wording, some class of applications falls outside the stated trigger. Under the minutes' wording, the delegation is not expressly limited by size.

There is no responsible way to erase that discrepancy. The records do not explain whether the register summarised, narrowed or corrected the minutes. They do not say that the qualifier was accidentally omitted from one text or accidentally inserted into the other. They do not establish which wording governed internal practice. Treating the two formulations as interchangeable would conceal the central uncertainty; choosing one as authentic would invent evidence.

Even the narrower formulation is incomplete. Neither text defines large. No published number of addresses, prefix length, IPv4 or IPv6 distinction, financial value, novelty test, risk score or measure of organisational complexity appears in the resolution. Size might conceivably refer to several different dimensions, but possibility is not proof. A reader cannot select one. Nor can the absence of a definition in these two records prove that no internal guidance, policy, contract or legal constraint existed. It establishes only that the public act, read through these records, does not state the threshold.

Trigger ambiguity is not a semantic inconvenience. The trigger is the first allocation of procedural burden. Before anyone assesses evidence, someone decides that an application belongs in the enhanced channel. If the public standard is large but the measure is unknown, applicants cannot predict whether identical resource quantities will be treated alike, whether documentary complexity substitutes for size, or whether staff impression does the work of a rule. If the minute-book wording applies to every application, the possible field is broader still. Governance begins at selection, not at the later factual finding.

The discretion that was actually delegated

The phrase “any available methods, at his discretion” is broad about technique. It is not self-executing authority over every consequence that might follow from a check. Method discretion answers a procedural question: how may the institution test a material assertion? It does not, without another identified source of authority, answer who may delay indefinitely, impose new substantive conditions, withhold an otherwise due allocation, punish a speaker, revoke an existing resource or decide a dispute.

The official texts do not list authorised methods. They do not mention document checks, databases, site visits, interviews, surveillance, outside investigators or any particular source category. It would be equally wrong to claim that a specific technique was used and to insist that the phrase placed every imaginable technique beyond constraint. Available cannot sensibly mean unlawful. Discretion cannot sever a method from the purpose for which it was delegated. An authentication measure should be directed to a precise application fact, proportionate to the consequence and uncertainty, and no more intrusive than reasonably needed to answer the question.

This distinction preserves the useful core of the resolution. Methods can adapt to differences in corporate form, jurisdiction, language and recordkeeping. A fixed public checklist may become obsolete or invite evasion. But the institution can protect operational detail while publishing the rule architecture around it. It can say what class of case receives enhanced review, what kind of fact is in doubt, who owns the case, when an answer is due and how error can be corrected without disclosing the specific indicators that reveal a forged record.