Summary

  • AFRINIC’s Board ratified AFPUB-2019-GEN-006-DRAFT03 in January 2022 through Resolution 202201.673, a corporate act distinct from the community-consensus stage that preceded it.
  • The strongest case for separate ratification is accountability: participants in a voluntary policy process can recommend a rule, but they cannot by consensus alone commit AFRINIC’s staff, budget, systems, signing infrastructure, repositories or corporate risk.
  • The public resolution proves acceptance of the named proposal, but it does not prove delegation, implementation, publication of an AS0 object, validator observation, operator reliance or any routing outcome.
  • AFRINIC remains a private member-based technical bookkeeper and coordinator. Ratification did not confer legislative or regulatory power, and a technical Invalid result would not be a legal judgment or a universal instruction to reject a route.
  • A compact, versioned ratification receipt would preserve the useful corporate gate while showing the exact handoff, authority, scope, decision record, risk owner, review terms, correction route and operator-discretion boundary.

The threshold crossed by 202201.673

In January 2022, AFRINIC’s Board adopted Resolution 202201.673, ratifying AFPUB-2019-GEN-006-DRAFT03. The most important thing about that sentence is not the technical acronym. It is the verb and the actor. The Board ratified a proposal that already existed and had already travelled through a community policy process. The resolution did not rewrite the proposal at the moment it was adopted. It changed the status of AFRINIC’s responsibility: a recommendation produced through consensus became a commitment accepted by the private corporation’s governing organ for action within AFRINIC’s own bounded services.

That distinction is easy to lose because the vocabulary of Internet coordination often compresses several institutions into a single story. Authors publish and revise a proposal. Participants discuss it. A process records consensus and deals with any appeal. The exact version then has to be handed to a corporate decision-maker. A Board can ratify it. Management can receive authority and resources. Technical teams can build and operate controls. A repository can expose signed objects. Validators can observe them. Networks can decide what weight to give the resulting validation state. These are not synonyms for one event.

They are separate stages, controlled by different actors and proved by different records.

The Board’s role is most defensible when understood as a restraint on institutional ambiguity. Community participants can examine a proposed technical rule and indicate support, but they do not control AFRINIC’s payroll, budget, contracts, systems, signing keys, repositories or liability merely because consensus has been declared. If consensus were treated as self-executing, the organisation could incur operational and legal commitments without a visible act by the body responsible for them.

Ratification can instead confirm the precise version, test whether it fits the corporation’s limited role, accept bounded risk and authorize an accountable handoff.

The proposal before the Board concerned RPKI ROAs for unallocated and unassigned address space administered by AFRINIC. Draft 3 proposed that AFRINIC create AS0 ROAs over such IPv4 and IPv6 inventory. It required covering AFRINIC objects to be revoked and removed before later allocation. It also preserved the possibility of a distinct Trust Anchor Locator and placed reclaimed resources under the proposal only after reclamation ended. Those features identify what the Board accepted; they do not establish that the organisation later implemented, published or operated any particular object.

The technical sequence also has boundaries. An AS0 ROA expresses negative route-origin intent. If a conflicting ordinary origin is evaluated through route-origin validation, a validator can classify it as technically Invalid. That result is information for a network, not an automatic act upon a route. The operator may monitor, de-preference, reject or make no routing change. Neither community consensus nor a Board resolution converts that independent choice into a command. “Invalid” in this context is not a finding about legal title, abuse or guilt.

The official AFRINIC Board record and policy register establish the numbered ratification and connect it to Draft 3. The proposal page identifies the instrument and the mechanism presented for acceptance. They do not, on the available record, establish the exact day of the Board meeting, its agenda materials, quorum, attendance, conflicts, legal or staff advice, vote count, dissent, delegation terms, budget, effective condition, implementation deadline, review clause or reasons beyond the act of ratification. Those are unknown fields.

Their absence from the public record is not proof that no private record existed, and it is not proof of misconduct.

That evidentiary discipline matters because the consequences of category error run in both directions. If ratification is dismissed as ceremonial, responsibility for committing corporate resources disappears into the word “consensus.” If ratification is inflated into sovereign authority, a private service provider’s internal act begins to resemble public law. AFRINIC is neither a legislature nor a regulator.

It is a private member-based technical bookkeeper and coordinator whose legitimate work is limited but important: maintaining accurate number-resource records, supporting uniqueness, operating bounded certification and publication services, preserving continuity, documenting changes and correcting errors.

NRS brings the member-side question into view: what precisely did the organisation accept, and how can the people affected by its services inspect that responsibility? Heng Lu supplies the decisive institutional limit: a mailing list is not a legislature, a meeting is not a public mandate and coordination does not become sovereignty by repetition. LARUS explains why the line still matters operationally: decisions made at a registry can transmit into diligence and continuity costs for networks even without becoming law.

BTW’s broader AS0 analysis supplies comparison and a boundary, but it is not evidence that this resolution was implemented or caused a route to be rejected.

The proportionate remedy is not to remove the ratification gate. It is to make the gate reconstructible. A useful receipt would name the exact proposal version and hash; the consensus and appeal state; the handoff date and actor; the Board meeting and agenda item; the corporate power relied upon; quorum, attendance, conflicts and recusals; advice received; questions asked; the vote and any dissent; the accepted scope; the implementation delegate; budget and risk owner; effective conditions; review or sunset; implementation evidence; the correction route; and an express statement that operators retain routing discretion.

Such a receipt would not manufacture a regional mandate. It would show the more modest and more credible thing: what a private corporation committed itself to do through its own staff and systems. Resolution 202201.673 therefore deserves neither dismissal nor mythology. Consensus supplied a recommendation. Ratification supplied corporate acceptance. Execution and routing remained later, separately observable decisions.