Summary

  • RFC 6410, BCP 9 of October 2011, cut the standards track from three maturity levels to two, and its Section 1 justifies the cut with a claim about practice — that most Standards Track documents never leave Proposed Standard and that very few advanced in a decade — rather than with a count, an interval or a denominator.
  • The interoperability test was not abolished. The formal report disappeared in Section 3.2, while the requirement for at least two independent interoperating implementations, now joined to widespread deployment and operational experience, became criterion one of Section 2.2 for Internet Standard.

A header block that answers the first questions

The plain-text rendering of RFC 6410 begins with a header block that settles jurisdiction before any prose arrives. It prints "BCP: 9", "Updates: 2026", "Category: Best Current Practice", "ISSN: 2070-1721" and the date October 2011, and it names three authors with their affiliations: R. Housley of Vigil Security, D. Crocker of Brandenburg InternetWorking and E. Burger of Georgetown University (RFC 6410, BCP 9). The RFC Editor's catalogue record repeats the same metadata and the same document text, identifying it as "BCP 9: Reducing the Standards Track to Two Maturity Levels", IETF process, October 2011, Best Current Practice (RFC Editor catalogue record).

The abstract states the job in one sentence: the document "updates the Internet Engineering Task Force (IETF) Standards Process defined in RFC 2026" and "reduces the Standards Process from three Standards Track maturity levels to two." What follows that sentence is the part worth reading closely, because a process document that deletes a step is also a document about what evidence a standards body will accept for its own work.

Quotations here come from the retrieved page text of the plain-text RFC; letterhead, pagination and section anchors were not independently verified. The retrieved text ends inside Section 4, on Security Considerations, so Sections 5 and later — references, acknowledgments, author addresses, copyright and BCP 78 boilerplate — were not inspected.

The justification is a judgement, and it says so

Section 1 does not pretend to measure. It states that "the vast majority of Standards Track documents are published as Proposed Standards and never advance to a higher maturity level", and that "Very few specifications have advanced on the maturity ladder in the last decade" (RFC 6410, BCP 9).

Neither the header block nor Section 1, as retrieved, supplies a number of documents, a stated date range for the count, or a denominator behind either sentence. That is the finding, and it is a narrow one: the deleting document rests its case on an observation about how the process had been used, and it does not publish the observation's arithmetic. A reader who wants to check the premise cannot check it from this document. They can check only that the premise was asserted, in those words, in that section.

The same section does date the politics. Adoption is attributed to IESG discussion in May 2010 and to a plenary discussion at IETF 78 in July 2010 that "demonstrated significant support for transition from a three-tier maturity ladder to one with two tiers" (RFC 6410, BCP 9). Those are meetable facts: a discussion, a meeting, a month.

The rung, in the words it lost

The rung that disappeared had its own published entry criteria, and they were not thin. RFC 2026 Section 4.1.2 admitted a specification to Draft Standard when "at least two independent and interoperable implementations from different code bases have been developed" and "sufficient successful operational experience has been obtained", defining "interoperable" as "functionally equivalent or interchangeable components of the system or process in which they are used" (RFC 2026, Section 4.1.2).

That section pushed the requirement across the whole specification: it applied "to all of the options and features of the specification", required unimplemented options and features to be removed before advancement, and made the Working Group chair responsible for documenting the qualifying implementations and the interoperation testing in the protocol action request. It also supplied the reason anyone cared. "Elevation to Draft Standard is a major advance in status, indicating a strong belief that the specification is mature and will be useful", and a Draft Standard was "normally considered to be a final specification" (RFC 2026, Section 4.1.2).

The test moved; the paperwork died

This is where the folklore usually goes wrong. The October 2011 document does not announce that interoperability stopped mattering. Section 1 describes the older arrangement as an interim step requiring "an interoperability report covering at least two implementations from different code bases", and says the new arrangement measures interoperability "through widespread deployment of multiple implementations from different code bases, thus condensing the two separate metrics into one" (RFC 6410, BCP 9).

Section 2.2 then carries the substance forward as criterion one of advancement to Internet Standard: "There are at least two independent interoperating implementations with widespread deployment and successful operational experience." The same section requires an IETF-wide Last Call of at least four weeks and provides that, on consensus, "the RFC will be reclassified without publication of a new RFC" (RFC 6410, BCP 9).

What Section 2.1 says about the entry level is equally blunt: "No new requirements are introduced; no existing published requirements are relaxed." And Section 3.2 states plainly what was actually withdrawn: "The IETF Standards Process no longer requires a formal interoperability report, recognizing that deployment and use is sufficient to show interoperability", while pointing to RFC 5657 as an aid to conducting interoperability testing (RFC 6410, BCP 9).

So the audit trail changed shape rather than vanishing. Two implementations still have to exist and interoperate; deployment and operational experience now have to exist beside them. What no longer exists is the document that the Working Group chair once assembled to prove it. The test survives as a condition. The exhibit does not.

A name chosen to avoid a collision

The new top level was not a fresh invention, and the document says so. Section 2 states that Internet Standard "is a merger of Draft Standard and Standard as specified in RFC 2026", and that the name was chosen "to avoid confusion between 'Draft Standard' and 'Internet-Draft'" (RFC 6410, BCP 9). A reader expecting a new maturity criterion at the summit will not find one there; the summit is the old summit with a clearer sign.

The residue: documents that keep a dead label

Deleting a rung for future advancement does not empty it. Section 2.3 handles the occupants, and the handling is asymmetric. Documents then at Standard level "shall be immediately reclassified as an Internet Standard", while documents at "the abandoned Draft Standard maturity level will retain that classification, absent explicit actions"; after two years the IESG may reclassify such a Draft Standard as Proposed Standard (RFC 6410, BCP 9).

That is a deliberate piece of untidiness: the ladder has two rungs, and for a while the catalogue can still show a third. Anyone reading a standards status in the years after October 2011 had to know which of the two documents governed the label on the page.

The review cycle that stopped

Buried in Section 3 is a second deletion, quieter than the rung. Section 3.1 states: "No review cycle is imposed on Standards Track documents at any maturity level" (RFC 6410, BCP 9). Advancements, in other words, no longer come with a scheduled obligation to look again. That is a governance change with no counterpart anywhere in the header block, and it is the part of the document least likely to be quoted.

Who signed, and what the record dates

The author line is printed, and so is the origin. The IETF Datatracker records RFC 6410 as part of BCP 9, document type "RFC - Best Current Practice (October 2011)", "Updates RFC 2026", originating as draft-housley-two-maturity-levels — an individual submission in the general area, in the IETF stream — with Jari Arkko as responsible AD and authors R. Housley, D. Crocker and E. Burger (Datatracker record for RFC 6410). The same page carries a maintenance date of 20 May 2026; that is page churn, not the publication date, which the page itself gives as October 2011.

The office record is a separate instrument and is dated separately. The IETF's roster of past IETF Chairs lists "Russ Housley (2007-2013)", giving tenure by year only (IESG past members). The IETF 82 Administrative Plenary minutes from Taipei, dated 16 November 2011, record the agenda item "4.1. IETF Chair Report by Russ Housley (see slides)" (IETF 82 plenary minutes).

Those two records do not add up to a document of October 2011. No document reviewed for this article and dated in that month states in its own text that Housley held the chair while RFC 6410 was published; the year span and the November minute place him in the role either side of it. Nor does any reviewed document state how the drafting labour was divided among the three named authors.

The honest reading of the record is narrower and still striking: a document that re-wrote the IETF's own advancement ladder carries on its author line the name of a man the IETF's own roster places in its chair from 2007 to 2013, in an individual submission rather than a working-group product.

What a reader can still check

Two checks survive the deletion, and neither requires trusting a summary. The first is the premise: open Section 1 and confirm for yourself that the case for removing a rung is stated as a judgement about practice, with no count behind it — the pages above are the pages that were read, and Section 5 onward were not. The second is the test: read Section 2.2's four criteria beside Section 3.2's withdrawal of the report, and see that interoperability was relocated in the evidence chain rather than struck from it.

What cannot be checked from this document, because it never printed it, is how many specifications actually stalled on the ladder, or for how long. The rung is gone; the number that justified removing it was never on the page.