Summary

  • RFC 7258 carries the IETF's Best Current Practice label and a plain verdict — pervasive monitoring is an attack — but the reviewed text names no party to decide whether a given specification met the requirement.
  • Around the verdict sit a threat model its own record says has no formal standing in the IETF standards process, a group that concluded as a BOF and was never chartered, and an IESG statement that names the reviewer and then excludes this subject from the mechanism by example.

The title is the verdict: Pervasive Monitoring Is an Attack. The front matter of RFC 7258, as published in plain text attaches the machinery that makes those four words more than a position — Internet Engineering Task Force, Request for Comments 7258, BCP 188, Category: Best Current Practice, May 2014, S. Farrell of Trinity College Dublin and H. Tschofenig of ARM Ltd. The status line is a single sentence: "This memo documents an Internet Best Current Practice." The abstract then attaches the qualification that governs everything after it: "Pervasive monitoring is a technical attack that should be mitigated in the design of IETF protocols, where possible." The HTML rendering prints the same front matter and the same operative passages; it is a second rendering of one document, never a second document.

What follows is a ladder of graded verbs, and it is worth reading each rung for the party it lands on. "The IETF community's technical assessment is that PM is an attack on the privacy of Internet users and organisations." Then the organisation commits: "The IETF will work to mitigate the technical aspects of PM, just as we do for protocol vulnerabilities in general." Then the authors are addressed: "Those developing IETF specifications need to be able to describe how they have considered PM, and, if the attack is relevant to the work to be published, be able to justify related design decisions." Then the whole apparatus compresses into a conditional: "It means that, if asked, there needs to be a good answer to the question 'Is pervasive monitoring relevant to this work and if so, how has it been considered?

'"

Read that sentence twice. The answer must exist only if a question is asked, and no sentence in the reviewed text names who asks, when the asking happens, or what follows a bad answer. The Datatracker record supplies the administrative facts — also known as BCP 188, document type RFC - Best Current Practice (May 2014), RFC stream Internet Engineering Task Force, IESG Responsible AD Jari Arkko, formerly draft-farrell-perpass-attack as an individual submission — together with a last-updated date of 20 May 2026 that refreshes the record rather than the standard.

The document's own timing is not a clock. "An appropriate balance will emerge over time as real instances of this tension are considered." It warns that "Revisiting these architectural decisions late in the process is very costly", which prices a late decision without ever naming the day it comes due. And the artefact a reader would go looking for was ruled out in advance: the RFC Editor's info page records the document's statement that a separate "pervasive monitoring considerations" section is not required in IETF documentation. A heading in the next specification was therefore never the test.

The perimeter is drawn honestly, and it is wide.

The IETF, as a standards development organisation, does not control implementation or deployment of its specifications, does not standardise all layers of the protocol stack, and places the non-technical legal and political aspects of mitigation outside its scope. "The broader Internet community will need to step forward to tackle PM, if it is to be fully addressed." The closing commitment keeps the same register: "The IETF will strive to produce specifications that mitigate pervasive monitoring attacks." Strive is the operative verb, and it is a verb about effort, not about outcome.

Correction, by contrast, carries a complete vocabulary. The errata view for RFC 7258 printed "No matching errata found" when checked on 21 November 2026, and it prints the statuses it would use if it had any: Verified+Reported, Verified, Reported, Held for Document Update, Rejected, with types Editorial and Technical. Bound that observation to the view itself — one search view, whose filter state the page does not display, on one date, which says nothing about whether an erratum was ever filed under a different query. The contrast is what matters: the channel that fixes wording publishes its verdicts by name, while the channel that would test whether the verdict was met publishes a conditional sentence.

Around the verdict, the programme's documentary trail is thin in a very specific way. The perpass group record prints WG Name Handling Pervasive Monitoring in the IETF, acronym perpass, Security Area, State "BOF Concluded", Charter "(None)", and the line "Charter for Working Group Not chartered yet", with Sean Turner and Stephen Farrell as chairs and a public mailing list and archive. A BOF that concluded is a meeting that ended; it is not a standing assessor, and the record says so in its own field.

The method arrived a year later and without standing. RFC 7624 is dated August 2015, is filed as Informational, comes from the Internet Architecture Board (IAB) stream, and lists R. Barnes, B. Schneier, C. Jennings, T. Hardie, B. Trammell, C. Huitema and D. Borkmann as authors. It states that it "is not an Internet Standards Track specification", and that "Documents approved for publication by the IAB are not a candidate for any level of Internet Standard". Its Datatracker record for RFC 7624 is blunter still: "This RFC is not endorsed by the IETF and has no formal standing in the IETF standards process." It supplies the threat model and the problem statement — the method the verdict lacks — and it inherits no power to decide anything. The verdict has no method. The method has no standing.

The IAB Statement on Internet Confidentiality is dated 13 November 2014, is typed as an IAB Statement, is marked Active, and carries the same date in its body. Its core sentence is an urging: "The IAB now believes it is important for protocol designers, developers, and operators to make encryption the norm for Internet traffic." The urging sentence names four addressees — protocol designers, developers, network and service operators, and firewall policy administrators — and none of them is an assessor of IETF specifications. The statement sits six months after the RFC at month precision: the RFC prints a month and the statement prints a day, so no day-level interval exists to measure between them, and none is asserted here.

One document in this set does name a reviewer and a mechanism, and then removes this subject from it. DISCUSS Criteria in IESG Review, updated 7 May 2014 on a 2007 original, prints that "Members of the IESG have the option, when they review a document, of stating a DISCUSS position", resolvable only to the satisfaction of the Area Director who issued it. That is a named decider at last. The same statement then supplies the carve-out in its own example: "a lack of consideration for pervasive monitoring in an existing specification would not justify holding a DISCUSS on the extension or minor update." The statement prints a day; the RFC prints only a month; which came first is not established by the documents reviewed.

Put the pieces side by side and the shape of the instrument appears.

There is a verdict with a standards-process status; a ladder of considerations that lands on the organisation, on authors, and on a review whose reviewer is unnamed; one interrogative left in the passive; a timing clause with no unit; a required-section question already answered no; an errata channel with a full verdict vocabulary and nothing in it; a group that carries the subject's own name and never chartered; a threat model that disclaims standing; an urging addressed to four outsiders; and a review mechanism that names its decider and explicitly declines to carry this failure into extensions and minor updates.

Nothing in that list is a violation. Nothing in it is an assessment either. Across the material reviewed for this piece, the only artefact of compliance the verdict produces is a sentence that begins "if asked" — and that is a design finding about an instrument, not an allegation about any author, working group or Area Director.

The classification would change with four additions, none of which needs a new process. Name the assessor inside the Best Current Practice itself. Make the considerations section required rather than explicitly optional, with a stable heading a stranger can search for. Add a field on the draft's record carrying the answer and its date. Publish a tally of the drafts where the question was put. All four are field-level changes to registers that already exist, which is precisely why their absence is worth the space: the gap is a design choice, not a constitutional problem.

This desk has read the correction channel before. My briefing of 18 November 2026 followed who may amend a published text and where the amendment lives; this piece follows the different question of who decides that a published verdict was met. The errata vocabulary appears here as the contrast, not as the subject. The card this article is attached to, Handling Pervasive Monitoring in the IETF, had no prior coverage at all.

Twelve years after May 2014, the most enforceable sentence in RFC 7258 is still its shortest conditional. A standards body said plainly what it believed the network was facing, and set a test for its own specifications that begins with the word if. What it did not do, in any document reviewed here, was name the person who would have to answer.