Summary

  • The 25 April 2005 joinder gave AFRINIC, as an admitted Regional Internet Registry, a contractual entitlement to a three-person contingent in the NRO Number Council. The incorporated NRO MoU divided the exercise of that entitlement: one person was appointed by the AFRINIC Board and two were selected by the regional policy forum.
  • Those three people simultaneously constituted AFRINIC-originating members of the ICANN ASO Address Council. Their work was advisory and procedural within a multi-gate policy chain; the seats did not give AFRINIC or the individuals sovereign, legislative, punitive, routing-control, or resource-ownership authority.
  • The AFRINIC Board had an express right to replace its own appointee. Removal of a member originating from the region otherwise depended on a published regional procedure, and no operative AFRINIC procedure for that purpose was located for 2005. A later interim appointment establishes a vacancy, not that a removal occurred.
  • The 2005 record identifies Jean Robert Hountomey and Sylvia Geha Kezengwa as the two forum-selected members and Alan Barrett as the Board appointee, but it does not close the voter denominator, ballot design, Board appointment rationale, candidate-screening practice, cause of the later Kezengwa vacancy, or one official collision over the elected members' term-end years.

Analysis

The right was institutional before it was personal

The decisive sentence in the 25 April 2005 NRO-AFRINIC Memorandum of Understanding did something precise. It admitted AFRINIC into the existing Number Resource Organization arrangement and stated that AFRINIC, as an RIR, was entitled to select three individuals to the NRO Number Council as described in the NRO MoU. The entitlement belonged to AFRINIC in its institutional capacity. It did not belong separately to every person who attended an AFRINIC meeting, every member of the company, every network in the service region, every African state, or every user whose connectivity might later be affected by number-resource policy.

That distinction matters because the joinder's compact phrase—AFRINIC could select three individuals—can sound broader than the incorporated rules allowed. The 24 October 2003 NRO MoU supplied the internal architecture. One person in each RIR contingent was to be appointed by that RIR's Board. Two individual members were to be selected by the RIR's regional policy forum through an open, accessible, documented, and transparent procedure. “Select three” was therefore an umbrella entitlement held by the RIR, not a grant of three appointments to the AFRINIC Board.

The arithmetic is simple but the authority map is not. One Board seat plus two forum seats equals three. Within AFRINIC's contingent, the Board-controlled share was one-third and the forum-selected share was two-thirds. Once AFRINIC joined the four existing RIRs, five contingents of three produced a fifteen-member council, giving the AFRINIC contingent three-fifteenths, or 20 per cent, of the institutional seats. Those fractions describe seat allocation. They do not measure population, member organizations, routed networks, addresses, states, capital at risk, or end users.

Equal weight among five RIR service regions was a design choice in a private coordination compact, not evidence of proportional democratic representation.

The legal and organizational character of the entitlement also sets its limit. In its March 2005 recognition application, AFRINIC described itself as a registered regional not-for-profit, non-governmental membership organization managed from Mauritius. The earliest archived bylaw PDF located in the checked corporate collection, from 2007, identifies AFRINIC Ltd. as a Mauritius company and a private company limited by guarantee. The later bylaw corroborates the corporate form; it is not being used to invent a 2005 seat procedure.

A private regional institution can hold contractual rights, make appointments, operate services, and participate meaningfully in international coordination. Its private status neither erases the right nor turns that right into public sovereignty.

The timing reinforces the narrow institutional reading. ICANN recognized AFRINIC as the fifth RIR on 8 April 2005. The NRO-AFRINIC joinder followed on 25 April, seventeen elapsed days later, and the recorded regional selection occurred on 26 April. Recognition was relevant context and an entry gate. It did not make ICANN the selector of the three people, and temporal proximity does not transfer the AFRINIC Board's one appointment, the forum's two choices, or the region's conditional removal competence to the ICANN Board.

One contingent did not mean one selector

The Board-appointed seat had the clearest corporate control line. The NRO MoU gave each RIR Board one appointment and expressly allowed that Board to replace its appointee with immediate effect. The AFRINIC-2 meeting report says that the AFRINIC Board appointed Alan Barrett for a one-year mandate. Later ASO material identifies him as Board appointed. That evidence closes the fact and seat class of the appointment, but not the process behind it. The checked public record does not supply the appointment resolution, exact date of the Board act, vote, candidate pool, qualification review, or reasons for choosing Barrett.

The other two seats had a different source of authority. They were allocated to selection by the AFRINIC regional policy forum through the open, accessible, documented, and transparent process required by the NRO MoU. AFRINIC's archived April 2005 procedure described the election as open to anybody. Its nomination form requested affiliation, physical address, and country. The published page did not state a candidate residence, employment, citizenship, nationality, or no-RIR-staff condition.

This is an important evidence boundary because a different sentence concerned voters. The AFRINIC-2 report said eligible voters were people residing and working in the AFRINIC service region. The wording does not establish whether both conditions were required, whether “or” was intended, which dates mattered, or how either condition was verified. More fundamentally, a voter rule is not a candidate rule. It cannot be moved from one column of the process into the other merely because the surviving record is incomplete.

Nor can a superseded rule be imported to fill the gap. The 1999 ASO MoU had barred RIR staff from serving, but the relevant 2005 chain ran through the 2003 NRO MoU and the replacement 2004 ASO MoU. The checked operative instruments do not reproduce that blanket exclusion. Without evidence that an old qualification survived into the new instruments or appeared in AFRINIC's own 2005 procedure, it cannot be treated as the governing rule.

The defensible statement is narrower: the published regional candidacy process was open to anybody and printed none of the candidate restrictions just listed, while any private screening practice or rejected nomination record remains unknown.

This split between the Board and the forum is not a drafting curiosity. It defines who can answer for each selection. The Board's fiduciary or institutional reasons could apply to its own appointee, though the record does not disclose those reasons. The regional forum supplied the two other selections through a meeting process. Calling all three “AFRINIC representatives” may be convenient as a provenance label, but it can conceal that only one came from a Board act.

Conversely, describing the two forum selections as if they were appointments by every African operator would inflate participation into authorization that the record does not show.

What the April 2005 result proves—and what it does not

The AFRINIC-2 report records five candidate tallies: Jean Robert Hountomey received 33 candidate-votes, Sylvia Geha Kezengwa 27, Ali Drissa Badiel 23, Pierre Ouedraogo Atohoun 16, and Michuki Mwangi Ochieng 5. The report declared Hountomey and Kezengwa elected. It separately recorded Barrett as the Board appointee. The top two rankings are therefore established as the official report's recorded result. Hountomey led Kezengwa by six candidate-votes, and Kezengwa led Badiel by four.

Adding the five tallies gives 104 candidate-vote units. Nothing in the checked material makes that 104 unique people. The ballot design, number of votes allowed per voter, number of eligible voters, ballots issued, ballots cast, invalid ballots, and abstentions were not disclosed. Attendance of more than 100 at the meeting is not an electorate denominator. Without unique voters and an eligible population, there is no defensible turnout percentage. Without a known ballot design, even a percentage of the 104 units would describe allocation among candidate-votes, not share of an electorate or strength of a political mandate.

The official result records an institutional act; it does not resolve the larger representation question. The people who participated could contribute knowledge, preferences, and judgment. Their votes validly filled the two seats under the recorded process. But being affected by registry policy makes someone a stakeholder; it does not automatically make that person a principal authorized to bind everyone else who may be affected. The meeting did not produce evidence of powers of attorney from all resource holders, operators, users, states, or residents in the service region.

The winners therefore became the individuals selected through that forum for the two institutional seats—not sovereign delegates of a continent.

That boundary protects rather than belittles the selection. A procedural role can be genuine without carrying universal political authorization. The two winners could serve, deliberate, advise, review process, and perform the defined ASO functions. The error would be to convert those specified functions into a claim that the election created public jurisdiction. A meeting can choose officeholders for a private coordination body; it cannot manufacture powers that neither the participants nor the contracting institutions possessed.

The initial mandates were described in staggered terms. The contemporary record largely supports Hountomey for three years through 2008, Kezengwa for two years through 2007, and Barrett for one year through 2006. An ASO update presented at APNIC 20 uses that sequence. A separate ASO presentation to RIPE 50 reverses the 2007 and 2008 end years for Hountomey and Kezengwa. The AFRINIC-2 report, the 28 April ASO notice, the APNIC 20 material, and the later vacancy notice collectively support Hountomey as the three-year/2008 seat and Kezengwa as the two-year/2007 seat, but the conflicting official table remains part of the evidence record.

It should be disclosed, not silently corrected as if official documents could not disagree.

Three people, two institutional names, one set of seats

The 21 October 2004 ICANN-NRO ASO MoU made members of the NRO Number Council the members of the ASO Address Council. Selection to the NRO NC therefore supplied service in the ASO AC without a second election or a second seat. Each person had a dual institutional role, not two separately conferred mandates. Counting the NRO NC and ASO AC titles as six AFRINIC places would double-count the same three people.

In the NRO setting, the Number Council was advisory to the NRO Executive Council on proposed global IP number-resource policies and served as a consultation point for external parties. It developed its business procedures for approval by the NRO EC. The council members could elect a chair from among themselves through a majority secret ballot for no more than one year, but chair status did not expand the underlying authority. Internal procedure and advice remained distinct from the power to select or remove a regional member.

In the ASO AC role, the same people participated in the global-policy process, advised ICANN on number-allocation policy, made recommendations concerning recognition of new RIRs, and performed specified ICANN selection functions. These were meaningful institutional tasks. Advice can affect what reaches a decision maker; process review can expose whether regional steps converged; recommendations and appointments can shape later institutional action. But influence is not unilateral command.

The operative framework defined a global policy as one requiring agreement through all RIR processes and ICANN action because it called for action by IANA or another external ICANN-related body. No single AFRINIC seat-holder could enact such a policy. The three together could not do so. The full council could not bypass all RIR and ICANN gates simply by treating its advice as legislation. That multi-gate structure is why the seats should be understood as a bridge into global coordination rather than a transfer of sovereign regulatory power.

The NRO Executive Council also sat at a different control layer. It represented the NRO, approved Number Council procedures, and acted within the MoU chain. It did not choose AFRINIC's two forum-selected members. The ICANN Board recognized AFRINIC and acted at the ICANN end of the process, but the checked instruments did not make it selector or remover of the three AFRINIC-originating members. Authority was distributed among bodies, and accurate analysis depends on resisting the temptation to collapse them into a generic “Internet governance community.”

The removal question exposes the asymmetry

Entry into the three seats was divided, and exit was divided too. For the Board-appointed member, the NRO MoU was explicit: the selecting RIR Board could replace its appointee, with the successor taking the predecessor's place immediately. This was a direct replacement path for one seat. Its precision is also a limit. The clause did not say that the Board could use the same power against either of the two forum-selected members.

For a member originating from an RIR region, the 2004 ASO MoU allowed removal by that region according to its published procedures. The global instrument thus recognized regional removal competence but did not itself name the local decision body, grounds, threshold, notice, hearing, review, or appeal. Those details had to exist in the region's published procedure. No operative AFRINIC removal procedure for the two forum-selected seats was located for April 2005.

The absence claim must remain carefully bounded. It concerns the checked public, indexed, and local record through the evidence cutoff. It does not prove that no private, unindexed, lost, or later-discovered document ever existed. It also does not prove that removal was impossible under every conceivable source of law. What it does mean is that the available record cannot assign summary removal power to the AFRINIC Board, the regional forum, the membership, the NRO EC, or ICANN. A removal clause that depends on a published local process cannot be completed by intuition.

Later AFRINIC material sharpens the finding without retroactively supplying the missing rule. Board minutes dated 26 November 2017 recorded that the AFRINIC ASO AC/NRO NC Election Process and Guidelines did not include a removal procedure and resolved that one should be drafted for community comment. That statement came more than twelve years after the 2005 joinder. It is strong evidence that the relevant guidelines lacked a procedure at that later point and supports the diagnosis of a public accountability gap. It is not proof of every practice throughout the interval.

The draft subsequently published in 2018 is evidence of a proposal and of the acknowledged need. It was presented for comment, so it cannot be treated as the operative rule in 2005 or automatically as a final rule in 2017. A 2019 appointment and removal procedure said that no NRO NC/ASO AC-specific removal process existed and routed those seats through general provisions. That is later governance evolution. Applying it backward would replace the actual 2005 evidence with a procedure developed for a different time.

The distinction between replacement, removal, and vacancy is equally important. The 2004 ASO MoU allowed an individual member to resign by written notice to the ASO AC, NRO Secretariat, and ICANN Secretary. When death, resignation, or removal caused a vacancy, the affected-region RIR could make an interim appointment until the next scheduled regional election; the election would then fill any remaining term. The same continuity mechanism followed several possible causes. Observing an interim appointment does not reveal which cause occurred.

That is exactly the limit of the January 2007 notice. The ASO reported that the AFRINIC Board reappointed Barrett for a three-year term and appointed Ali Drissa Badiel to complete Kezengwa's term. The notice establishes a vacancy and an interim Board action. It does not say that Kezengwa resigned, died, was removed, became unable to serve, or left for another reason. It cannot be used as proof of removal, proof that a valid removal procedure existed, or proof that the Board owned the underlying forum-selected seat. Continuity power after a vacancy is not the same as power to create that vacancy.

The strongest case for the design

The most persuasive contrary case is not that the missing procedure did not matter. It is that the 2005 arrangement was a reasonable federal compromise for technical coordination. AFRINIC entered with the same three-person contingent as each established RIR rather than as a lesser participant. Two seats had to come through an open regional process, preventing the corporate Board from monopolizing all input. One Board appointment could offer institutional continuity and had a clear, immediate replacement route.

The same people served in the NRO NC and ASO AC, avoiding duplicated councils, while multiple RIR and ICANN gates constrained unilateral action.

The initial staggered one-, two-, and three-year mandates also reduced the prospect of all three positions expiring simultaneously. The ASO MoU's interim-vacancy rule protected continuity after death, resignation, or removal. Its decision to leave removal to published regional procedures can be understood as subsidiarity: the global compact did not impose one recall code on every region. On this reading, the design blended institutional equality, local process, continuity, and limited authority.

That steelman fits the surviving instruments and should not be caricatured. A mixed selector system is not inherently undemocratic merely because one seat is Board appointed. There is no basis in the sealed record to call the April 2005 election fraudulent, captured, or invalid, and no basis to infer incompetence, conflicts, alliances, or political blocs from names or vote ranks. The absence of a removal procedure does not show that any officeholder deserved removal.

The residual concern is narrower and harder to dismiss. Subsidiarity works only when the local procedure on which the global rule depends is public and usable. The checked AFRINIC record did not close that loop for the two forum-selected seats, and the later Board record acknowledged the omission in the election guidelines. The design therefore combined a well-specified entry route with an under-specified accountability off-ramp. That is a governance weakness, not evidence of wrongdoing.

The same restraint applies to institutional equality. Giving AFRINIC three of fifteen seats avoided excluding the new RIR's channel from global coordination. It did not prove that the three people proportionally represented African population, networks, resource holdings, member organizations, states, or economic exposure. The arrangement can be defended as equal participation among RIR institutions without being promoted into a democratic theory it was never shown to satisfy.

Official records are evidence, not a source of unlimited authority

The record in this case is dominated by institutional documents: executed memoranda, meeting reports, election pages, Board minutes, presentations, recognition material, and role explainers. Each source has a proper evidentiary use. The executed joinder proves its terms and recorded agreement. The incorporated MoU proves the allocation of selectors and functions within that compact. A meeting report is evidence of what the institution recorded as the tally and outcome. Board minutes establish that the Board recorded a procedural gap and resolved to draft a response.

Official branding does not make every characterization self-proving. The ASO's 28 April 2005 notice said AFRINIC members voted Hountomey and Kezengwa into the NRO NC. That is an attributed official statement. It does not erase the broader voter wording in AFRINIC's process and result records, and it does not prove that all voters were corporate members if the underlying material does not close that question. Likewise, a presentation table can be official and still collide with another official table over term-end years.

Nor do official documents prove metaphysical authority. A memorandum can create contractual rights among its parties. It cannot, by the force of institutional language alone, turn a service region into a sovereign people, attendees into principals for absent parties, or a ledger coordinator into a government. Executable power must be traced to a competent contract, law, court, or enforceable right. Symbolic importance and practical influence cannot create powers the parties did not own.

The current AFRINIC-2 report URL returned a 404 during the package audit, while a dated Internet Archive capture preserved the report. That preservation detail matters because it shows why source closure sometimes requires both a current institutional URL and an archived copy. It does not change the content boundary: the report records the official account of the election, but it does not supply the ballot denominator or candidate-screening ledger that it never contained.

The evidence cutoff is 10 August 2026 in Asia/Shanghai. Findings of absence are limited to the checked public, indexed, archived, and local record through that date. If a contemporaneous removal procedure, ballot register, appointment resolution, or vacancy record later appears, the analysis should change in proportion to what the document actually proves. Evidence discipline is not indecision. It is the difference between a conclusion that can survive new records and a claim built from the silence between them.

A coordination bridge is not a licence to rule

The Heng Lu Note supplies the controlling interpretive boundary: a registry may record, coordinate, and protect uniqueness, but it may not turn its ledger function into sovereignty or punishment. AFRINIC's seat entitlement sits squarely within the coordination side of that line. It connected the institution to the NRO and the ASO, put regional process knowledge into review and advice, and gave AFRINIC-originating individuals defined roles in recommendations and selection procedures.

Nothing in the checked instruments gave the three members ownership of number resources, command over routing, or the ability to fine, punish, prosecute, or bind all operators in Africa. Nothing made AFRINIC a legislature or court. Independent operators still make routing decisions. Any later allocation, revocation, RPKI, WHOIS/RDAP, contract, price, or continuity consequence would need to be traced to its own competent policy, registry act, agreement, or law. It cannot be imputed to the three-seat entitlement merely because the council discussed global policy.

This limit also protects office disputes from escalating into resource punishment. A valid selecting body may replace or remove an officeholder within the applicable procedure. That action concerns service in the council. It does not follow that the individual, the individual's employer, or unrelated operators can be penalized at the resource layer. Office accountability and registry administration are separate legal and operational surfaces.

NRS appears in the source record as an advocate, researcher, convener, and representative of explicitly authorized members. It does not operate AFRINIC's registry, RPKI, WHOIS/RDAP, appeals, settlement, elections, custody, or continuity. Its material helps articulate the distinction between a voluntary bookkeeping institution and sovereign regulation; it does not replace the executed joinder as evidence of the April 2005 entitlement or tally.

The doctrine does not erase practical leverage. Advice may change timing, framing, and recommendations. Institutional recognition may make one channel more influential than another. Registry policies and contracts may have serious operational consequences when adopted through competent processes. The point is to name each source of effect. The three-seat right created access to a coordination chain. It did not itself create the downstream acts that might later affect a resource holder.

Implications for operators and resource holders

For operators, the immediate value of the joinder was access rather than command. An AFRINIC-originating contingent could bring regional knowledge and objections into global-policy review, ICANN-facing advice, and discussions about the coordination system. Exclusion would have deprived the new RIR of that institutional channel. Inclusion reduced one form of coordination asymmetry, even though it did not prove that every affected network had authorized the selected individuals.

The mixed selector design could diversify the channel. Two forum-selected members might carry perspectives visible in an open regional process, while one Board appointee could maintain continuity with the institution responsible for registry operations. That is a plausible structural benefit, not a documented account of the founders' motives or the conduct of the initial members. The surviving record does not show how actively each person attended, voted, reported back, declared conflicts, or responded to selectors.

For resource holders, the danger lies in rhetorical compression. If all three seats are described as Board-controlled, the forum's two selection rights disappear. If the forum is described as “Africa,” absent networks and principals are written into a mandate they did not grant. If council advice is described as regulation, later costs can be blamed on the wrong institutional act. Accurate labels are not semantic fussiness; they are a map of where an affected party can seek records, accountability, or remedy.

The removal gap matters because unclear off-ramps can weaken accountability for nonperformance or conflicts, while ad hoc removal can destroy independence. The 2004 framework attempted to balance those risks by requiring a published regional procedure and preserving interim continuity after a valid vacancy. AFRINIC's public record did not complete the first element in 2005. The correct response is to identify the missing process, not to invent a power or infer misconduct.

The economic impact path is indirect but potentially real. Council members could shape information, process assessment, and recommendations. Those inputs might affect later decisions in the RIR-ICANN chain, and later decisions might affect registry practices or operator costs. Each step, however, requires its own evidence. No specific allocation, revocation, routing, security, fee, contract, or continuity outcome in the sealed package can be causally assigned to the mere existence of AFRINIC's three seats.

The bounded finding

The defensible conclusion is institutional equality with split authority and incomplete accountability. AFRINIC received a contractual three-person contingent. Its Board controlled one appointment and the immediate replacement of that appointee. Its regional policy forum controlled two selections through the stipulated open process. The selected people served simultaneously in the NRO NC and ASO AC, where their functions were advisory, procedural, recommendatory, and limited by other institutional gates.

The public record does not establish a 2005 AFRINIC procedure for removing a forum-selected member. Later acknowledgement of that omission strengthens the gap finding but does not retroactively supply a rule. The 2007 appointment to complete Kezengwa's term demonstrates continuity after a vacancy, not the cause of the vacancy and not Board ownership of the elected seat. The candidate tallies demonstrate rank, not voter turnout or continental mandate.

These distinctions preserve both halves of the truth. The joinder was consequential: it placed the fifth RIR on equal institutional footing inside an established global coordination structure. But three seats were not sovereignty. “AFRINIC selected three” was not “the Board appointed three.” Regional participation was not universal authorization. Advisory influence was not punishment power. A serious account of registry governance begins by keeping those boundaries visible.