Summary
- The IANA Numbering Services Review Committee is advisory by an express clause: it has no power to commit the NRO EC to any recommendation or decision of its own, except where it holds express delegated authority.
- The rule that decides whether the committee can speak at all was changed on 11 December 2023: quorum moved from a member from each of the five regions to a member from at least four of them, while eight members and five voting members stayed fixed.
- Across nine indexed review years the published record shows two community comment submissions, and the most recent verdict sits above a compliance sentence that reports reverse DNS at 80 per cent, with no stated conversion rule in the sections read.
The clause that says what it cannot do
The committee's function, in the charter published by the Number Resource Organization, is 'to advise and assist the NRO EC in its periodic review of the service level of the IANA Numbering Services provided to the Internet Number Community' (charter). The same document then removes the sharpest instrument a reviewer can hold: 'The Review Committee does not have any power to commit the NRO EC to any recommendation or decision made by it, except where and to the extent that it has express delegated authority from the NRO EC' (charter).
Read that carefully, because it is narrower and stranger than the word 'advisory' suggests. The limit is on power, not on process. The committee owes an annual report, and it owes a public consultation before that report becomes final: 'Before the Review Committee submits any final report and/or recommendation to the NRO EC, it must first consult and seek the feedback and comment from the Internet Number Community about such proposed report and/or recommendation' (charter). An advisory body with a compulsory consultation step is not a passive one. It is a body whose output is mandatory and whose effect is optional.
Fifteen seats, three kinds of relationship
Composition carries the rest of the story. The charter sets fifteen members: 'two community appointees from each RIR region (who must not be RIR staff)' plus 'one RIR staff from the region (who will be a non-voting member)' (charter). The current operating procedures restate that arrangement, and the arithmetic is exact: three seats from each of five regions (operating procedures v2.1).
Who fills those seats is deliberately unsettled at the centre. 'Each RIR shall appoint its Review Committee members by a method of its own choosing', and the charter leaves the term of appointment to each registry as well (charter). One region's page spells out the local mechanism: 'As proposed by the RIPE Chair, the two community-elected NRO Number Council members and one RIPE NCC staff member who is appointed by the RIPE NCC Executive Board will participate on the RIPE community's behalf' (RIPE NCC). That page is undated and its printed roster lists terms that ended years ago, so it is evidence about a selection method, not about who sits today.
The arithmetic of a quorum
Here is the part a reader can check without trusting anybody. The 2019 procedures said: 'Quorum will consist of 8 Review Committee members, of whom there must be at least one member from each region and five voting members' (operating procedures v2.0). The 2023 revision says: 'Quorum will consist of 8 Review Committee members, of whom there must be at least one member from at least four regions and five voting members' (operating procedures v2.1). Both texts survive in the open, and v2.1's own title block flags 'changes in quorum requirements'.
That is a real transfer of margin. Under the older rule, a finding needed presence from all five regions. Under the newer one it needs four, which means the committee can now speak with one region absent from the room. The documents state that the requirement changed. They do not state why — whether to survive absences, to move faster, or for a reason nobody wrote down. I read both versions on 19 November 2026 and the reason is not in either.
The cycle, and who supplies its raw material
The operating procedures describe a machine with a fixed rhythm. The registries publish the IANA Numbering Services Review Matrix each January, the committee passes it to the five communities 'for a comment period of 30 days', and the comments are reviewed and summarized into the performance report (operating procedures v2.1). The NRO's own committee page describes the same annual round (NRO committee page).
Two facts about that machine matter for anyone who wants to audit it later. The matrix — the primary evidence the committee reasons from — is compiled and published by the registries, not by the committee. And the procedures themselves are a registry-side instrument. The reviewer inherits both its inputs and its rulebook.
Nine years, two comments
The published record of the cycle is a single table, and the table is startling. The reports index lists nine review years, 2017 through 2025, with a comment period, a comment count and a final report date for each (reports index). Two of those nine rows record a submission: one in the 2019 cycle and one in the 2020 cycle. The other seven print 'None'.
The table also carries three loose threads that its own text does not tie off. The 2022 row prints 'Not applicable' where every other row prints a consultation window. The 2018 row prints a period opening on 21 January and closing on 22 January 2019, which does not read like a thirty-day round. And the 2024 cell gives a closing date without a year. Any of those could be a drafting slip on a page last modified on 30 March 2026; none of them is explained there (reports index).
What that record supports is narrow, and worth stating precisely: the public comment route has been used twice in nine years. It does not support the claim that nobody looked at anything. The index counts submissions, not scrutiny.
The verdict and the number beneath it
The 2025 report concludes: 'We, the RC, conclude that the performance of the IANA Numbering Services Operator related to delivery of the IANA Numbering Services and performance under the SLA in 2025 was acceptable' (2025 report). It states the basis one line later, in a sentence it attributes to registry staff: 'RIR staff have reviewed the IANA performance reports for 2025 and confirm that the SLA was met 100% of the time in all cases except for reverse DNS performance, where the SLA was met 80% of the time' (2025 report).
And the reconciliation, where a reader would look for it: 'In reaching our conclusion that the SLA was met, we evaluated the data sets noted above and reviewed the SLA requirements' (2025 report). That is an assertion of diligence, not a formula. In sections 2, 4.1, 4.3 and 5 the report does not print the rule that allows a metric sitting at 80 per cent to coexist with an overall finding that the service level was met. Perhaps that rule lives in the agreement or in the matrix; I did not obtain either in this review, so the gap is in my evidence as much as in the document.
This is a documentary point, and it deserves to be kept separate from a serious one. Nothing in the material read for this article alleges a misstated figure, a captured review or an officer acting improperly, and I make none of those allegations. The defect is reproducible reading: a member of the community who accepts every number in the report still cannot arrive at the verdict from those numbers.
The remedy is one paragraph long
The fix does not require a new committee, a new charter or a new secretariat. It requires the document that carries the verdict to carry its rule: state how the individual metrics combine into the overall finding, and state whether a single metric below target is tolerable and on what basis. That paragraph belongs in the report the communities are asked to comment on, because clause 10 requires them to comment on it (charter) — and a comment period on a conclusion without its rule is not a consultation on much.
Because the registries compile the matrix and publish the procedures, the drafting burden is partly theirs as well: whatever rule governs the conversion has to exist in an instrument the committee can cite and the community can read. Wherever it is written, the test is the same. A reader holding the same matrix should be able to reach the same verdict, or a different one, and say exactly where they parted company.
What this review did not obtain
The boundary belongs in the open. Not obtained for this article: the IANA numbering services agreement between ICANN and the five registries, and therefore its review, dispute and termination clauses; the IANA Performance Matrix and its Summary Report, named in the 2025 report as its primary data set; the two comment submissions the index records; any committee minutes or transcripts; a printed publication date on the charter extraction; and the full appendix tables of the 2025 report, which arrived garbled in the rendering and supply no figure here. The committee itself is catalogued in our directory.
Member Briefing
Read the full trend briefing
Members read the full briefing with the source notes behind every finding.
Strategic Circle
A daily briefing in your language, built from changes we have verified.
Join for USD 20/monthCancel anytime.
Members also get the full research library, a daily briefing on the networks they follow and the private forum.






