Summary of active discussions in the MOQ ecosystem during June 2026.

Activity (June 4 06:00 UTC → June 7 19:00 UTC) — London interim agenda PUBLISHED (Martin Duke June 5, “shifted more time to MOQT blockers”); DTS consensus readout flips to “separate WG extension document” (draft-ietf-moq-dts4moq) — Cullen Jennings contests June 6; Tobbe ships MSF/CMSF catalog validator (CUE) that finds draft-01 bugs → files MSF PR #177 + CMSF PR #23; Kota Yatagai debuts Moqtopus (C++/MsQuic Unreal Engine client); thibmeu opens 2 privacy-pass issues; moq-dev/moq adds moq-hls gateway + VP8/VP9 + moq-gst maturation; interop drops to 41 → 37 → 38 pass

TL;DR:

London interim (June 11-12) agenda is now PUBLISHED on Datatracker. Martin Duke posts “New Agenda for London” June 5 — “The revised agenda is posted… shifted a bit more time to MOQT blockers” — and Mike English posts “London interim details” (logistics/remote). Day-1 (interim-08, June 11, 0900-1700 BST) is now almost entirely MOQT-blocker time: Concurrent Subscribe (Ian, #1633), Object Range Filters (Mo, PR #1518), Joining FETCH (Alan+Mo, PR #1642), Request Blocking (Alan, #1519), two “Other MOQT Issues” blocks (Alan+Ian, 1330-1430 + 1500-1600), and SWITCH/DTS Implementation Lessons (Ali, 1600-1700) demoted to the day’s tail. Day-2 (interim-10/-11, June 12) is still conditional on the SWITCH/DTS outcome — both branches keep MSF/CMSF (Will), Privacy Pass (Suhas), Secure Objects (Cullen), LOC (Mo); the not-incorporated branch adds CDN Provisioning (Mike) + MSFTS (Gwendal). LOCMAF and catalog-format are not on either agenda.
DTS consensus readout flips to “separate WG extension document” — and Cullen Jennings contests it. The chairs’ readout of the June-4-close DTS/SWITCH consensus call concludes there is “strong consensus for DTS to be adopted as an extension described by a WG document” — not integrated into the base moq-transport draft — and invites Will Law to publish his wilaw/dts4moq repo as a new WG document draft-ietf-moq-dts4moq at his earliest opportunity (“it is possible that the document later reaches a state where it could be merged into MOQT”). Cullen Jennings June 6 pushes back: “How did you reach the conclusion that there is consensus this should be an extension instead of in the MoQT draft?” — a reversal of the trajectory the wiki tracked (Cullen’s May 27 base-spec-PR preference + wilaw’s PR #1638 base-spec integration, still OPEN). The readout lands closer to Magnus Westerlund’s original separate-draft proposal; Cullen, who argued against a separate draft, is now the dissenting voice on the process question.
Implementations: moq ~20 merges + several OPEN (June 4→7, all kixelated) — new PR #1626 “Add moq-hls: HLS / LL-HLS gateway” (+1695/−49, OPEN), VP8/VP9 codecs (PR #1625 moq-mux + PR #1632 gstreamer), moq-gst maturation (#1627 dynamic catalog follow, #1633 pad-pump refactor, #1646 backwards-timestamp fix), moq-net reshape (#1631 Track→TrackInfo+async TrackConsumer, #1634 one subscription + N concurrent fetches, #1640 RAII announcements), PR #1637 moq-json JSON Merge Patch for catalog snapshot/delta, moq-lite-05 PR #1648 (add TRACK stream, drop SUBSCRIBE_OK/FETCH_OK — supersedes #1609) + PR #1601 MERGED (FETCH past groups), relay #1628 GOAWAY drain + #1630 TLS cert reload OPEN. moqx: afrind PR #386 MERGED (per-session auth filters replace session-keyed grants map) + multithread stack 363 still OPEN; gmarzot+TimEvens static-Boost build fixes; peterchave auto-regression-test CI #379 MERGED. quiche moqt: 3 vasilvv commits June 5-6 (MoqtControlMessageQueue factoring + “has first object in subgroup” header bit). moqlivemock v0.11.0 + warp-player v0.11.0 ship MSF/CMSF draft-01 catalogs (+ LOCMAF 0.2 in mlm). NEW msf-catalog-validator (Tobbe). moqtail, moq-rs (PR 171 untouched), moq-js (PR #72 OPEN), imquic (PR #27 updated June 5), mondain/moqxr (TilsonJoji SRT PR #14 OPEN), mondain/moq2ts quiet.
Interop: 38 / 137 pass/fail at 2026-06-07 00:49:07 UTC (draft-18 target) — sequence June 5 41 → June 6 37 → June 7 38, a −13 to −16 drop from June 4’s 54; total slips 177 → 176. Version breakdown still 0 at target · 0 ahead · 176 behind — no impl has registered draft-18 (Nokia’s announced v17/v18 relay not yet registered; moqlivemock now shows draft-16/14). 20 consecutive daily reports (May 19-Jun 7) = new longest cadence streak. London interim 4 days away (hackathon/interop June 9-10, formal sessions June 11-12).

London interim agenda PUBLISHED — Day-1 is MOQT-blocker time, DTS/SWITCH demoted

Martin Duke’s June 5 “New Agenda for London” message (“The revised agenda is posted… I shifted a bit more time to MOQT blockers”) and Mike English’s June 5 “London interim details” (arrival/remote logistics) finalize the structure for the June 11-12 in-person interim at Cloudflare’s London office. The Datatracker agendas:

Day-1 — interim-2026-moq-08 (June 11, 0900-1700 BST):

Time (BST)	Item	Presenter	Ref
0930-0945	Administrivia	—
0945-1015	Concurrent Subscribe	Ian	#1633
1030-1045	Object Range Filters	Mo	PR #1518
1045-1145	Joining FETCH	Alan + Mo	PR #1642
1145-1230	Lunch (Tongyu demo)	—
1230-1330	Request Blocking	Alan	#1519
1330-1430	Other MOQT Issues	Alan + Ian
1430-1445	Next Interim Planning	Martin
1500-1600	Other MOQT Issues (cont.)	Alan + Ian
1600-1700	SWITCH/DTS Implementation Lessons	Ali

Day-2 — interim-10 (morning) + interim-11 (afternoon), June 12 — still flagged “dependent on the outcome of the SWITCH consensus call”. Morning: Interop Report (Mike) + DOS Design-team readout (Mike, draft-englishm-moq-relay-dos) + conditional SWITCH Issues (Gwendal) / DTS Issues (Will). Afternoon has two branches; both keep MSF/CMSF (Will), Privacy Pass (Suhas), Secure Objects (Cullen), LOC update (Mo); the incorporated branch adds Top-N DDOS (Cullen) + Top-N issues (Mo) + Top-N Implementation (Suhas); the not-incorporated branch adds CDN Provisioning (Mike) + MSFTS (Gwendal).

Structural read: Martin’s “shifted more time to MOQT blockers” is visible — Day-1 gives Alan+Ian two hour-long “Other MOQT Issues” blocks to triage the 14-item draft-18 errata/design backlog, while SWITCH/DTS is reduced to a 1-hour Ali “Implementation Lessons” report at the very end of Day-1, no longer the centerpiece it was at the May 26 interim. LOCMAF (Tobbe) and catalog-format do not appear on either agenda — the LOCMAF initData-carriage discussion would surface, if at all, under Will’s MSF/CMSF slot.

DTS consensus readout — “extension WG document”, and the Cullen reversal

The chairs’ readout of the DTS/SWITCH consensus call (opened May 21, closed June 4) concludes:

DTS: adopt as an extension described by a WG document — “strong consensus for DTS to be adopted as an extension described by a WG document”, not integrated into the base moq-transport draft. Will Law is invited to publish his wilaw/dts4moq repo as the new WG document draft-ietf-moq-dts4moq at his earliest opportunity; the readout leaves the door open: “it is possible that the document later reaches a state where it could be merged into MOQT.”

Cullen Jennings June 6 challenges the readout: “How did you reach the conclusion that there is consensus this should be an extension instead of in the MoQT draft?” This is a 180° turn from the trajectory the wiki tracked: through late May the on-list votes (Will Law, Gwendal, Nokia, Ali Begen) favored integration into MOQT, and Cullen explicitly argued against a separate draft (May 27: “I disagree it should be done as a separate draft… it should be worked as a PR that can be discussed in context”), prompting wilaw’s base-spec PR #1638. The chairs’ readout instead lands on a separate document (closer to Magnus Westerlund’s original proposal), leaving PR #1638’s base-spec integration path superseded-in-direction and Cullen as the dissenting voice on the process question. Carry-forward: draft-ietf-moq-dts4moq is invited but not yet submitted to Datatracker as of June 7; the June-12 “DTS Issues (Will)” slot is conditional on this outcome and may now reopen the extension-vs-base-spec question on the floor. (See interim-meetings for the full vote-tally history.)

Tobbe ships MSF/CMSF catalog validator (CUE) — finds draft-01 bugs

Torbjörn Einarsson announces on #moq (June 5 10:44 CEST) a catalog validator for MSF/CMSF draft-01: “I made a validator for MSF/CMSF catalogs based on the new draft-01. It found a number of issues in the drafts, but should be generally useful. Based on a schema in CUE with some extra code to generate user-friendly error messages.” — live at moqlivemock.demo.osaas.io/msf-catalog-validator, source at Eyevinn/msf-catalog-validator. The validator was built fast (4 PRs merged June 4-5: LOCMAF packaging + bitrate-not-required-for-non-A/V fix, web UI with examples + schema view + Prism highlighting, strict validation against draft-01 definitions rather than examples).

The validator immediately produced spec errata: Tobbe files moq-wg/msf PR #177 “Fix version in examples and minor typos in draft-01” (+24/−24) and moq-wg/cmsf PR #23 “Fix version examples and minor typos in draft-01” (+5/−5), both OPEN June 4. Significance: this is the first time a wiki-maintainer-built tool feeds errata directly back into the WG specs it validates — and it confirms the June-4 thesis that draft-01/draft-18 examples are carrying machine-detectable inconsistencies that an unhurried review would have caught. Tobbe’s moqlivemock v0.11.0 + warp-player v0.11.0 (both released June 4) move Eyevinn’s stack onto MSF/CMSF draft-01 catalogs, with LOCMAF 0.2 implemented alongside 0.1 in moqlivemock.

Kota Yatagai debuts Moqtopus; interop spreadsheet → automated runner

On #moq June 4 16:44 CEST, Kota Yatagai announces Moqtopus — “a C++ client based on MsQuic for use in Unreal Engine. I’ll finish the work needed to support draft-18 in time for the interop.” — a new game-engine-targeted MoQ client (Kota also authored the Moqintosh iOS client). Kota also asks whether the MoQ interop spreadsheet is still maintained (“kota’s moqtail is no longer chasing new drafts so I’ll delete the record”). Mike English confirms the automated interop-runner is intended to replace the spreadsheet (“we can be more precise and accurate about… exactly which versions of each implementation support what features”), notes the initial test set was “intentionally small” and more test cases are coming next week (open to ordering suggestions), and points to a new #moq-interop-runner channel. Alan Frindell asks “Are there new test cases coming?” — the answer being yes. This is the first explicit statement that the spreadsheet is being retired in favor of the runner, and that the runner’s coverage will broaden right after London.

thibmeu opens privacy-pass design issues

Thibault Meunier (Cloudflare; privacy-pass author) opens 2 issues on moq-wg/privacy-pass June 5: #14 “MOqTokenChallenge should be base64 encoded” and #15 “Find how to pass MoQTokenChallenge in reply to a SETUP closure without using ReasonPhrase”. These are the issuer-aware challenge-reply concerns thibmeu raised in his May 30 review of cloudflare/moq-rs PR #169, now filed against the privacy-pass-auth draft itself — moving the challenge-reply wire question from cross-impl review into the WG draft’s own issue tracker ahead of the June-12 Privacy Pass slot.

moq-transport: EXPIRES + sharmafb editorial PRs merge; 3 new design issues

The June-2/3 normative work largely lands: PR #1640 EXPIRES (martinduke) MERGED June 4, sharmafb PR #1645/#1648/#1649 all MERGED June 4, PR #1476 “Delivery timeouts are both Track and Object Properties” (afrind) MERGED June 5 (closes #606). Tim Evens’s #1641 (PUBLISH_DONE wording) + martinduke’s #1639 (EXPIRES) auto-CLOSE; ianswett’s PR #1451 “Allow multiple Subscriptions to a Track” CLOSED June 6 (folded into the Concurrent Subscribe discussion). 3 new design issues OPEN: #1650 (afrind, “Forward Handling SHOULD set Forward=1 upstream is too strong”), #1651 (martinduke, “How does GROUP_ORDER interact with PUBLISH?”), #1652 (Kostya Vasilyev, “SETUP - extensions negotiation”). The fill-replacement PR #1642 + DTS PR #1638 remain OPEN into the Day-1 Joining-FETCH and SWITCH/DTS slots. quiche moqt continues structural refactor June 5-6 (8a11928a + 6b6a9b74 factoring control-message queueing into a MoqtControlMessageQueue class; f3ce1892 “Support the ‘has first object in the subgroup’ bit in the subgroup header”).

Activity (June 3 06:00 UTC → June 4 06:00 UTC) — draft-ietf-moq-cmsf-01 PUBLISHED on Datatracker June 3 (2nd WG draft to land in 2 days); LOCMAF Slack thread engages afrind + wilaw on initData carriage; sharmafb editorial sprint files 5 draft-18 issues/PRs in 18 min; openmoq/moqx paul-mondain CAT token PR #264 MERGES after 33 days; google/quiche moqt starts draft-19 wire prep (new varint, CLIENT/SERVER SETUP merge); moq-dev/moq ~12 merges incl. voice-AI latency range; 2 new external contribs (arielmol SCTE-35, vipyne AI-voice flush); interop 177/54/122/0 (−3 pass)

TL;DR:

draft-ietf-moq-cmsf-01 PUBLISHED on Datatracker June 3 by Will Law (Akamai) — 2nd MoQ WG draft revision in 2 days (MSF -01 June 2 → CMSF -01 June 3). Final pre-submission clean-up was moq-wg/cmsf PR #22 MERGED June 3 11:32 UTC “Fix I-D nits” (+7/−6). wilaw also posts “MSF updates” + “CMSF updates” announcement messages to the IETF MoQ mailing list — first per-spec author-led list updates from wilaw the wiki has tracked.
LOCMAF Slack thread engages afrind + wilaw on initData carriage. Torbjörn Einarsson announces LOCMAF -00 at June 3 08:51 CEST citing msf draft-01’s new initData catalog references + catalog compression as enablers; afrind 18:13 CEST asks “Have we considered init data as a track property?”, then 20:06 CEST “InitData as a separate subgroup?”, then 22:00 CEST “Only Publish the init subgroup in groups where it changed.”. wilaw 21:37 CEST responds “we have added the init structure in MSF as an extensible scheme. if the init is not going to change for the life of the track, then we could add indeed add it as a defined track property… Currently we do not allow track init properties to change once publish has begun.” 5-reply thread surfaces 3-way design space: catalog-referenced vs track-property vs per-group-subgroup. Promotes the stale moq-msf Issue #153 (“initTrack does not work”) to active design engagement within 12h of LOCMAF announcement.
moq-transport: sharmafb (Cisco) 5-item editorial sprint June 3 21:09-21:27 UTC (18-minute window): PR #1645 “Change PUBLISH_OK = 0x1E in table” (PUBLISH_OK was removed); Issue #1646 “Application-specific property codepoints conflict” (§2.5 reserves 0x38-0x3F vs §15.8 reserves 0x78-0x7F); Issue #1647 “TIMESTAMP and SUBGROUP_DELIVERY_TIMEOUT have same value”; PR #1648 “Soften one mandatory response for a REQUEST_UPDATE”; PR #1649 “Modify text about PUBLISH_DONE on a control stream” (companion to Tim Evens’s #1641). Plus Issue #1644 by kixelated “Duplicate Track Properties on repeated SUBSCRIBE_OK / FETCH_OK” — same gap that moq-lite-05 PR #1609 already addresses via TRACK_INFO Track Stream. draft-18 errata-style issues now outpace design-issue resolution: 5 of 5 sharmafb items are wire-table-vs-text/codepoint-collision-style. quiche moqt starts draft-19 wire prep: vasilvv 3 commits June 3 — 4096d2e3 “Switch MOQT to using new varint format”, 8427f949 “Use delta encoding for Subscribe Filters of type Absolute Range”, 96c9a9c0 “Merge CLIENT_SETUP and SERVER_SETUP into a single control message”. First sustained draft-19 wire-prep work the wiki has tracked from any impl.
Implementations: moq ~12 merges + 7 OPEN cycle (June 3 ~05:00 → June 4 ~05:00 UTC, all kixelated unless noted). Highlights: PR #1620 OPEN “feat(watch): latency range with buffered playback” — reframes playback latency as a range [min, max] unifying live and buffered playback into one mechanism, unblocks voice-AI/TTS use cases (pipecat-ai/pipecat#4629) where the response is written faster than real-time. 2 new external contributors in 24h: arielmol PR #1617 OPEN SCTE-35 from MPEG-TS into a new data catalog track (first broadcast-ad-marker plumbing in moq-dev/moq); vipyne Issue #1614 + PR #1615 OPEN AudioProducer Python flush/cancel for AI voice agent interruption use cases. Together PR #1620 + #1615 + #1617 = 3 PRs adding new use-case verticals in 24h: AI-voice playback, AI-voice ingest, broadcast-ad-marker carriage. Plus watch refactor (PR #1591 Computed signals MERGED, PR #1592 inputs-vs-outputs MERGED, PR #1588 stop downloads when muted/paused/off-screen), CI release pipeline shakedown (PR 1623), moq-net per-control-stream task (PR #1621), apt keyring rename (PR #1611). moqx: paul-mondain CAT token PR #264 MERGED June 3 22:57 UTC after 33 days OPEN (+1807/−13, 20f) — first openmoq/moqx external-contrib-driven auth path now landed in main. Plus Tim Evens PR #376 + PR #377 MERGED (Debian bookworm + macOS Apple Clang 21); gmarzot PR #370 MERGED folly XLOG. afrind opens 3 new issues: #380 standard token type values; #381 auth for MOQT peers; #382 Draft-18 updates for relay behavior (priority-ranked sub-issues = draft-18 catch-up backlog scoped). mondain/moqxr: PR #14 OPEN by TilsonJoji (new external contributor) “Add SRT MPEG-TS ingest with fMP4 repackaging and MoQ object publishing” — first MoQ-side SRT ingest path in the openmoq orbit. quiche moqt: vasilvv 3 draft-19 wire-prep commits (see above). moq-rs: PR #167 + PR #171 untouched. moqtail, moq-js (PR #72 OPEN), imquic (PR #27 OPEN, updated June 3 17:14 UTC), Moqintosh, moqlivemock (PR #90 dependabot quic-go bump), Eyevinn/warp-player, Eyevinn/moqtransport, quiche_moq, Quicr/cat-token all quiet.
Interop: 177 / 54 / 122 / 0 at 2026-06-04 00:55:39 UTC — −3 pass vs June 3 (57 → 54, 32.2% → 30.5%, −1.7pp), reverses Jun 3’s +6 jump symmetrically; skip drops 1 → 0; 17-day cadence streak (new longest the wiki has tracked). Version breakdown still 0 at target · 0 ahead · 177 behind — impl-registration drift unchanged. Slack signal June 2 09:18 UTC by Yu You (Nokia): “we are in the process to host a remote relay for the interop. we have the support of v17 and v18 (basic ones with current interop test cases) for testing” — first announced impl-side v18 endpoint, would close the 0-at-target gap if registered. London hackathon 5 days away.

draft-ietf-moq-cmsf-01 PUBLISHED — 2nd WG draft in 2 days

Will Law’s CMSF -01 lands on Datatracker June 3, 187 days after -00 (Dec 2025). The 2-day MSF→CMSF publication cluster aligns the two streaming-format specs for the London Day-2 35-min MSF/CMSF Will Law slot.

Mailing-list announcements June 3 by wilaw:

“MSF updates” — companion narrative to MSF -01 (the 134-day-saga draft that landed June 2)
“CMSF updates” — companion narrative to CMSF -01

First wilaw-authored per-spec mailing-list update tracked by the wiki — historically he announces draft revisions via Slack rather than email. The June 3 list posts pair CMSF -01 with MSF -01 as a coordinated London-cycle update.

LOCMAF Slack thread — initData carriage as 3-way design space

Under Tobbe’s LOCMAF -00 announcement on #moq June 3 08:51 CEST, 5 substantive replies through 22:00 CEST form the first cross-author engagement on the new individual draft:

When	Author	Substance
08:51 CEST	Torbjörn Einarsson	Announcement of LOCMAF -00; cites msf draft-01 `initData` references + catalog compression as enablers for the reduced-complexity v0.2 design
18:13 CEST	afrind	”Have we considered init data as a track property?“
19:14 CEST	Tobbe	Replies: msf -01 introduces `initData` type currently only `"inline"`, but extensible to a track reference; links moq-msf Issue #153
20:06 CEST	afrind	”InitData as a separate subgroup?“
21:37 CEST	wilaw	”As Torbjorn mentioned, we have added the init structure in MSF as an extensible scheme. if the init is not going to change for the life of the track, then we could add indeed add it as a defined track property. Adding it as a subgroup would be wasteful, as you would have to continually re-send it. Currently we do not allow track init properties to change once publish has begun.”
22:00 CEST	afrind	”Only Publish the init subgroup in groups where it changed.”

The thread surfaces 3-way design space for initData carriage:

Catalog-referenced (msf -01’s inline type extended to a track reference) — Tobbe’s interpretation
Track property (afrind suggestion, wilaw seconds for steady-state immutability)
Per-group subgroup with only-when-changed delivery (afrind suggestion, key-rotation case)

Significance: First substantive cross-author engagement on LOCMAF, counter-signal to the “individual drafts sit with no engagement” pattern. Promotes the stale moq-msf Issue #153 (“initTrack does not work”) to active design engagement. Likely surfaces at the London Day-2 35-min MSF/CMSF Will Law slot as a structured design discussion of init carriage that also affects CMSF (full CMAF chunks) and LOC (self-initializing per-object).

sharmafb 5-item draft-18 editorial sprint

sharmafb (Cisco; not to be confused with Suhas Nandakumar) files 5 contributions in 18 minutes June 3 21:09-21:27 UTC, all draft-18 errata-style:

Item	Subject
PR #1645	”Change PUBLISH_OK = 0x1E in table” — PUBLISH_OK message type was removed but its table row remained
Issue #1646	”Application-specific property codepoints conflict” — §2.5 reserves 0x38-0x3F + 0x3800-0x3FFF for application use, but §15.8 reserves 0x78-0x7F. Direct codepoint collision
Issue #1647	”TIMESTAMP and SUBGROUP_DELIVERY_TIMEOUT have same value” — §15.8 codepoint collision
PR #1648	”Soften one mandatory response for a REQUEST_UPDATE” — §10.9.1 already allows coalesced failed updates to produce only one REQUEST_ERROR
PR #1649	”Modify text about PUBLISH_DONE on a control stream” — companion to Tim Evens’s Issue #1641

Combined with Issue #1644 by kixelated June 3 15:55 UTC “Duplicate Track Properties on repeated SUBSCRIBE_OK / FETCH_OK” (the moq-transport-side surfacing of the same gap that moq-lite-05 PR #1609 already addresses with TRACK_INFO Track Stream 0x6), the draft-18 errata + design issue load now stands at 14 active items (6 normative PRs OPEN + 8 design issues OPEN) heading into London Day-1 0900-1045 “MOQT Issues” 180-min slot — infeasible to ship all 6 PRs on the floor; chairs will likely defer pure-editorial items (PUBLISH_OK code change #1645 + PUBLISH_DONE wording #1649) to editor discretion, leaving DTS + EXPIRES + Fill as the substantive merges-on-floor candidates.

google/quiche moqt starts draft-19 wire prep in earnest

3 vasilvv commits June 3 (03:08 — 10:58 UTC) all directly altering MOQT wire shape:

SHA	When	Subject
`4096d2e3`	03:08 UTC	”Switch MOQT to using new varint format” — draft-19 wire bump
`8427f949`	03:22 UTC	”Use delta encoding for Subscribe Filters of type Absolute Range” — wire compaction
`96c9a9c0`	10:58 UTC	”Merge CLIENT_SETUP and SERVER_SETUP into a single control message” — handshake simplification

Plus martinduke 52de014c June 2 22:45 UTC (MoqtResponseCallback MessageParameters in REQUEST_OK) + c25d5258 June 3 15:30 UTC (ASAN fix).

First sustained draft-19 wire-prep work the wiki has tracked from any impl — up until now quiche moqt commits since May 14 have been refactor work (stream-class factoring + subscription publisher extraction). With vasilvv pivoting to wire-format changes, the implementation track for draft-19 is now open — moq-transport spec text changes for these will likely surface at London.

openmoq/moqx — PR #264 paul-mondain CAT token MERGED after 33 days OPEN

paul-mondain PR #264 MERGED June 3 22:57 UTC — the largest single moqx merge so far (+1807/−13, 20 files). Adds opt-in CAT-style token authentication with per-service auth config block, internal signed CWT/HMAC token verifier for exp/moqt/moqt-reval claims, AUTHORIZATION_TOKEN credential validation, and authorization checks on PUBLISH_NAMESPACE/PUBLISH/SUBSCRIBE_NAMESPACE/SUBSCRIBE/FETCH/TRACK_STATUS. Preserves existing relay peering token behavior.

Cross-impl auth status June 4: 3 cross-impl auth-relay-path PRs were open in parallel — openmoq/moqx PR #264 (paul-mondain CAT, OPEN since May 1) + cloudflare/moq-rs PR #169 (englishm-cloudflare AuthHook design) + cloudflare/moq-rs PR #171 (suhasHere C4M implementation). PR #264 lands first; PRs 171 still OPEN and untouched June 3-4. Combined with PR #286 (paul-mondain Catapult submodule wiring, stacked on #264) still OPEN, the openmoq/moqx auth story still has one more stack-on-top to land before complete.

afrind opens 3 new issues June 3 23:09-23:43 UTC:

Issue #380 “auth: use standard token type values” — “Now we support cat4moq, but allow the operator to configure the token type. We should use whatever the spec defines.” (filed ~10 min after PR #264 merge)
Issue #381 “auth: how does it work for MOQT peers?” — relay-peer auth bypass / inter-relay token presentation question
Issue #382 “Draft-18 updates for relay behavior” — priority-ranked sub-issues including NAMESPACE_TOO_LARGE error+reset enforcement, SUBSCRIBE_NAMESPACE stream closure semantics, plus “see what’s coming from upstream into moxygen/relay/MoQRelay and moxygen/relay/MoQCache — we likely need to port these over”. First openmoq/moqx draft-18 catch-up backlog scoped as a tracking issue.

TilsonJoji emerges at mondain/moqxr — SRT MPEG-TS ingest

mondain/moqxr PR #14 OPEN by TilsonJoji (new external contributor; OPEN June 1, last updated June 3 17:51 UTC) “Add SRT MPEG-TS ingest with fMP4 repackaging and MoQ object publishing” — first MoQ-side SRT ingest path in the openmoq orbit. Parallel to moq-dev/moq PR #1587 (MPEG-TS over MoQ via mpeg2ts crate) + mondain/moq2ts (MSFTS demonstrator). The mondain ecosystem now spans 3 MPEG-TS-adjacent paths: moqxr (SRT MPEG-TS ingest), moq2ts (MSFTS demonstrator), and the moqxr-as-SDK link target.

moq-dev/moq voice-AI / TTS use case lands at the player

PR #1620 OPEN by kixelated “feat(watch): latency range with buffered playback” (+726/−75, 14 files) reframes playback latency in @moq/watch as a range [min, max]. The PR body opens:

Reframes playback latency in @moq/watch as a range [latency-min, latency-max], unifying live (minimize-latency) and buffered playback into a single mechanism. This unblocks voice-AI / TTS use cases (e.g. pipecat-ai/pipecat#4629) where a response is written faster than real-time with future timestamps and should be buffered and played at the encoded pace, rather than the player aggressively minimizing latency and skipping ahead.

Latency was never really a single point — it’s a range, and “minimize latency” is just the degenerate case where the range collapses (max == min).

The sync reference (wall-clock anchor) is re-anchored (skipped forward) only when keeping it would push latency past the bounds. Pairs with PR #1615 by vipyne (new external contributor) addressing the ingest side: AudioProducer (Python) flush/cancel primitive for real-time interruption (AI voice agents). Together with PR #1617 by arielmol (new external contributor) “Ingest SCTE-35 from MPEG-TS into a data catalog track” (broadcast-ad-marker plumbing), moq-dev/moq adds 3 new use-case verticals in 24h: AI-voice playback (PR #1620), AI-voice ingest (PR #1615), broadcast-ad-marker carriage (PR #1617).

Yu You (Nokia) announces v17+v18 remote relay for interop

Yu You (Nokia) Slack #moq June 2 09:18 CEST: “we are in the process to host a remote relay for the interop. we have the support of v17 and v18 (basic ones with current interop test cases) for testing.” — first impl-side v18 endpoint announcement as pre-London infrastructure addition. If registered with the interop runner, this would be the first impl-registered v18 entry, finally closing the runner’s 0-at-target gap. Pairs with Yu You’s earlier May 29 mailing-list YES vote on DTS adoption (with measured Nokia implementation benefit numbers) as continued Nokia engagement across spec consensus + production deployment.

Activity (June 2 06:00 UTC → June 3 06:00 UTC) — draft-ietf-moq-msf-01 PUBLISHED on Datatracker June 2 (Day +134 saga closes); draft-einarsson-moq-locmaf-00 NEW individual draft by Tobbe + Hugo Björs; afrind opens transport PR #1642 replacing Joining Fetch with Subscription Fill; Tim Evens (Cisco) emerges as new cross-repo contributor (transport Issue #1641 + openmoq/moqx PRs 377); moq-dev/moq ~14 merges incl /health endpoint + moq-lite-05 wire-feature triple + MPEG-TS fMP4 close-out; interop 177/57/119/1 new May+June high 32.2% (+6 pass)

TL;DR:

draft-ietf-moq-msf-01 PUBLISHED June 2 on Datatracker by Will Law (Akamai) + Suhas Nandakumar (Cisco) — first MSF revision in 134 days (since -00 Jan 19 2026); ends the 5-day slippage saga of wilaw’s May 27 Slack “Friday” pledge. Final pre-submission clean-up was moq-wg/msf PR #176 by wilaw (June 2 11:18 UTC, −7/1f, “Remove unused RFC references and SCTE35 entry”). Companion moq-wg/cmsf PR #21 (+69/−20, MERGED 13:54 UTC, closes #20) updates CMSF’s MSF reference from -00 → -01.
draft-einarsson-moq-locmaf-00 NEW individual draft submitted June 2 by Torbjörn Einarsson (Eyevinn) + Hugo Björs (KTH) — “Low Overhead CMAF for Media over QUIC (LOCMAF)” — compact wire format carrying CMAF chunk metadata as tagged fields while preserving sample data unchanged; receiver reconstructs functionally equivalent CMAF chunks for MSE/EME playback. First IETF artifact from the wiki maintainer. Slots between LOC (strips fMP4 overhead but loses MSE/EME) and CMSF (keeps CMAF semantics at full chunk size) — compact-fMP4 carrier the player reconstructs to a CMAF chunk in browser memory before handing to MSE.
moq-transport PR #1642 OPEN June 2 17:01 UTC by afrind “Replace Joining Fetch with Subscription Fill, add Current Group” (+188/−148, 1f) — structural redesign: Joining Fetch removed entirely; new subscription filter types AbsoluteStartFill (0x5), AbsoluteRangeFill (0x6), CurrentGroup (0x7), RelativeStartFill (0x8); new FillDescending (0x3) group order; new Current Group Delivery + Fill Semantics sections; publisher opens single fill fetch stream (FETCH_HEADER reusing subscription Request ID) for past objects while delivering current/future via subscribe subgroups/datagrams. Replaces ianswett’s PR #1627 (OPEN since May 3, ianswett auto-CLOSES it June 3 00:16 UTC, 30 days OPEN). Plus 2 new draft-18 design issues — Issue #1641 by Tim Evens (Cisco) on PUBLISH_DONE control-stream-vs-request-stream wording (afrind reply “Request streams are considered ‘control’ streams in most of the text, but we can probably crisp this up”); Issue #1643 by afrind on bidi-stream FIN semantics. London Day-1 0900-1045 MOQT Issues 180-min slot now has 3 normative-text deliverables (DTS PR #1638 + EXPIRES PR #1640 + Fill PR #1642) + 4 design-issue threads — will require triage.
Implementations: moq another ~14-merge cycle (June 2 20:00 UTC → June 3 04:00 UTC, all kixelated): PR #1604 /health load-shedding endpoint (+905/−5, 11f, first explicit production-CDN feature); moq-lite-05 wire-feature triple (PR #1595 MERGED Frame Start in FETCH + PR #1601 OPEN TrackConsumer::fetch + PR #1609 OPEN TRACK_INFO Track Stream 0x6); MPEG-TS fMP4 export close-out (PR #1590 out-of-band avc1/hvc1 + PR #1593 AAC esds in fMP4); API ergonomics (PR #1606 ergonomic subscribe(None)); release + dependency wave (PR #1607 0.12.7 + PR 1603). moq-lite-05 now has 5 wire features in 7 days (deflate + AnnounceOk + Frame Start + TrackConsumer::fetch + TRACK_INFO Track Stream) — the most active wire-protocol evolution venue in the ecosystem. moqx: Tim Evens (Cisco) emerges as new external contributor — PR #376 OPEN Debian bookworm + CPM (+92/−31, 5f); PR #377 OPEN macOS Apple Clang 21 fmt workaround (+21/−6, 2f). Same TimEvens filing transport Issue #1641 — cross-repo emergence Day +1 of activity. 3rd Cisco contributor reaching mainline moq-wg artifacts alongside Suhas Nandakumar + Mo Zanaty. moq-rs: PR #167 minor refresh June 3 05:37 UTC; 171 untouched. mondain, moqtail, moq-js (PR #72 OPEN), imquic (PR #27 OPEN), quiche moqt, quiche_moq, Moqintosh, moqlivemock, Eyevinn/warp-player, Eyevinn/moqtransport all quiet.
Interop: 177 / 57 / 119 / 1 at 2026-06-03 00:56:06 UTC — +6 pass vs June 2 (51 → 57, 28.8% → 32.2%, +3.4pp), first 32%+ pass rate since cadence recovery May 19, largest single-day pass-count jump since the May 19 cadence recovery; 16-day cadence (new longest streak the wiki has tracked). Version breakdown still 0 at target · 0 ahead · 177 behind — impl-registration drift unchanged; the +6 jump comes from better cross-version compatibility behavior not registration-side closure. London hackathon 6 days away.

draft-ietf-moq-msf-01 PUBLISHED — 134-day saga closes

Will Law’s May 27 Slack pledge “I plan to release a new draft this Friday ahead of the London interop” (May 29 UTC) finally lands on Datatracker June 2 after 4 days of slippage:

Date	Status	Notes
2026-05-27 12:30 UTC	Pledge	wilaw Slack: “this Friday” (May 29)
2026-05-29 EOD	Day +130, slip 1	No submission
2026-05-30 EOD	Day +131, slip 2	No submission
2026-05-31 EOD	Day +132, slip 3	No submission
2026-06-01 EOD	Day +133, slip 4	No submission
2026-06-02 ~13:54 UTC	Day +134, submitted	-01 lands; CMSF reference bump merged shortly after

The 5-day slippage was real but contained: only ~3 trailing editorial PRs landed between announced ETA (May 29) and actual submission (June 2): PR #173 normative refs (May 29), PR #174 timestamp rounding (May 29), PR #171 parent namespace (May 30), PR 175 (June 1 burst), PR #176 RFC refs clean-up (June 2). The longer cycle absorbed Issue #163 version-string fix (PR #175 June 1) + RFC reference clean-up (PR #176 June 2) that would otherwise have needed a -02.

MSF -01 cadence: 134 days from -00 (Jan 19) to -01 (June 2). Sets MSF’s publication tempo as ~4 months per revision — slower than moq-transport (-17 → -18 in 49 days) but consistent with MSF being a packaging-format spec rather than a wire-protocol spec.

18 wilaw events May 24-June 2 (largest single-contributor MSF push since draft adopted) all absorbed into -01:

PR #157 (suhasHere group numbering)
PR #159 (suhasHere catalog compression)
PR #165 (wilaw bitrate properties + mandatory audio/video fields)
PR #166 (wilaw typed-object initDataList; Tobbe’s design adopted)
PR #167 (wilaw targetBuffer; closes kixelated’s Issue #150)
PR #168 (wilaw catalog object specs + numbering)
PR #171 (wilaw parent namespace for clone tracks)
PR #173 (wilaw normative reference update)
PR #174 (wilaw timestamp rounding)
PR #175 (wilaw version-string)
PR #176 (wilaw RFC refs + SCTE35 removal)

draft-einarsson-moq-locmaf-00 — First IETF artifact from the wiki maintainer

Torbjörn Einarsson (Eyevinn Technology) + Hugo Björs (KTH) submit draft-einarsson-moq-locmaf-00 “Low Overhead CMAF for Media over QUIC (LOCMAF)” June 2 on Datatracker. Per abstract: “LOCMAF defines a compact wire format that enables streaming low-latency CMAF media over MoQ Transport with significantly reduced per-object overhead. The format carries CMAF chunk metadata as tagged fields while preserving sample data unchanged, with the receiver reconstructing functionally equivalent CMAF chunks suitable for MSE/EME playback pipelines.”

Design position: slots between LOC (lean container, no MSE/EME compat) and CMSF (full CMAF chunks). LOCMAF’s distinguishing choice: carry CMAF chunk metadata as tagged fields, preserve sample data unchanged, reconstruct CMAF chunk receiver-side — transparent to MSE/EME consumers (the player still sees standard CMAF chunks) while removing redundant box headers from the wire.

Motivation: Eyevinn’s existing CMAF-based pipeline (HLS/DASH origins + low-latency CMAF chunked encoding) needs a MoQ wire shape that doesn’t force re-architecting the player-side MSE/EME glue. LOCMAF preserves that. Also resolves Tobbe’s May 23 moq-wg/msf Issue #153 Point 4 (“AVC3 doesn’t resolve the mid-stream-change question because Safari (notably for FairPlay DRM) requires avc1/hvc1 sample entries”) by allowing parameter sets out-of-band as tagged fields.

See the new wiki page for moq-locmaf for details.

afrind PR #1642 — Joining Fetch removed, Subscription Fill introduced

afrind’s PR #1642 (+188/−148, 1f) structurally redesigns MOQT’s past-object retrieval:

Removed:

Joining Fetch types
Joining Fetches section
Joining Fetch Range Calculation
Joining Location
INVALID_JOINING_REQUEST_ID error code

Added (new subscription filter types):

AbsoluteStartFill (0x5)
AbsoluteRangeFill (0x6)
CurrentGroup (0x7)
RelativeStartFill (0x8)
FillDescending (0x3) group order (descending for fill, ascending for subscribe)

Added (new sections):

Current Group Delivery — publisher subgroup ordering + relay object accounting requirements, applies to any filter covering current group
Fill Semantics — defines fill fetch stream behavior

Modified:

Standalone Fetch fields inlined directly into FETCH wire format
LARGEST_OBJECT serves as fill boundary in SUBSCRIBE_OK
FILL_TIMEOUT extended to apply in SUBSCRIBE messages with fill filter types
REQUEST_UPDATE with a fill filter type opens a new fill fetch stream using the REQUEST_UPDATE’s Request ID
Fill filter types + CurrentGroup disallowed in PUBLISH_OK (stale LARGEST_OBJECT); recommends forward=0 + REQUEST_UPDATE pattern instead

Publisher behavior: opens a single fill fetch stream (FETCH_HEADER reusing the subscription’s Request ID) for past objects while delivering current/future via subscribe subgroups/datagrams.

Replaces ianswett’s PR #1627 “SUBSCRIBE with Joining Fetch” OPEN May 3 → CLOSED June 3 00:16 UTC (30 days OPEN). The architectural simplification (one filter dimension instead of two request types) and FillDescending group order directly answer ianswett’s Issue #1614 “(JOINING) FETCH + SUBSCRIBE prioritization”.

Tim Evens (Cisco) emerges as new cross-repo contributor

Tim Evens files 3 contributions across 2 repos within ~18h June 2 06:19 UTC → June 3 05:09 UTC:

Repo	PR/Issue	When	Subject
moq-wg/moq-transport	Issue #1641	June 2 06:19 UTC	”Publish DONE, control stream or request stream?” (draft-18 wording inconsistency)
openmoq/moqx	PR #376	June 2 23:42 UTC	”Fix Debian bookworm builds and switch deps to CPM” (+92/−31, 5f)
openmoq/moqx	PR #377	June 3 05:09 UTC	”fix macOS deps build with Apple Clang 21+ and tidy sanitizer flags” (+21/−6, 2f)

Cisco footprint across MoQ now spans 3 named individuals + 2 implementations + 4 specs:

Contributor	Specs	Implementations
Suhas Nandakumar	cmsf, msf, secure-objects, privacy-pass	cloudflare/moq-rs (PR #167 + #171)
Mo Zanaty	loc (draft author), agenda	—
Tim Evens	moq-transport draft-18 design (#1641)	openmoq/moqx (#376, #377)

Structurally the most distributed corporate presence in the MoQ ecosystem. Tim Evens’s emergence in the same week as the multi-thread sprint stack (PR 361-365 OPEN) + Suhas’s C4M AuthHook (PR #171 OPEN) means multi-author multi-corporation collaboration on openmoq/moqx is now visible in main-branch PR list — distinct from the kixelated-solo pattern at moq-dev/moq.

moq-dev/moq operational milestone — `/health` endpoint

PR #1604 (+905/−5, 11f) adds an unauthenticated GET /health to moq-relay’s existing web server:

Returns: 200 ok when every configured threshold passes; 503 overloaded + one plain-text line per breached threshold when any threshold breaches
No thresholds configured: pure liveness probe
Metrics: cross-platform sysinfo crate; load-average flag Unix-only
Config under [web.health] with matching --web-health-* flags + MOQ_WEB_HEALTH_* env vars:
- cpu — global CPU usage (75 or 75%)
- ram — memory usage (80% of total or 32GB/32GiB absolute used)
- rx / tx — aggregate throughput (unit required; lowercase b = bits, uppercase B = bytes)

First explicit production-CDN load-shedding feature in any tracked MoQ implementation. Pairs with PR #1571 May 31 (externalized --cluster-connect-api peer list) + PR #1574 May 31 (per-auth-root presence-based billing). Together: moq-dev/moq is now structurally a production CDN substrate with load-balancer integration, peer-list externalization, presence billing, and the moq-pro pattern consolidated.

moq-lite-05 wire-feature triple — pace continues

Counting from PR #1518 (Lite05Wip variant May 27) the wire-feature additions in 7 days are:

PR #1531 May 28 — Compression codec field in SUBSCRIBE_OK (deflate compression)
PR #1573 June 1 — AnnounceOk message (responder origin + initial active count)
PR #1595 June 3 — Frame Start (i) field in FETCH (resume mid-group)
PR #1601 OPEN June 3 — TrackConsumer::fetch first-class single-group FETCH
PR #1609 OPEN June 3 — TRACK_INFO Track Stream 0x6 (immutable props off SUBSCRIBE_OK)

All gated on Lite05Wip so default ALPN/Versions don’t advertise — opt-in for testing without disrupting Lite04 default negotiation.

moq-lite-05 is structurally the most active wire-protocol evolution venue in the MoQ ecosystem — 5 wire features in 7 days outpaces moq-transport’s draft-18 → draft-19 prep both in feature count and merge cadence.

Activity (June 1 06:00 UTC → June 2 06:00 UTC) — wilaw moq-transport PR #1638 DTS OPEN as PR-against-transport-18 (consensus-decided design lands as base-spec PR per Cullen’s preference); 3 MSF PRs MERGED (compression #159, bitrate #165, string-version #175) while MSF -01 still NOT submitted Day +134; moq-dev/moq 14 merges + 6 OPEN including MPEG-TS bridge PR #1587 (matches MSFTS/moq2ts) + moq-lite-05 AnnounceOk MERGED + per-track timescale (long-pending #1439) MERGED + qmux version mapping (long-pending #1513) MERGED + subscribe_track async (long-pending #1540) MERGED; cloudflare/moq-rs PR #171 AuthHook + C4M implementation OPEN by suhasHere (+2090/-15, implementation of design proposal #169); mondain/moq2ts MSFTS demonstrator publicly announced on Slack with cross-platform CI build workflow; Martin Duke asks moq if anyone wants draft-16 interop at London, Alan Frindell + Mike English confirm 14/16/18 multi-version support; google/quiche 2 commits (FETCH stream + RemoteTrack refactor); openmoq/moqx afrind multi-thread relay sprint continues with 5 OPEN PRs; interop 177/51/125/1 first draft-18 target run (-3 pass, 0 at target · 0 ahead · 177 behind reflects impl-registration vs draft-version drift)

TL;DR:

moq-transport PR #1638 OPEN June 1 11:06 UTC by wilaw “Add Dynamic Track Switching (DTS)” (+147/−0, 1 file, fixes #259) — DTS lands as a PR-against-moq-transport-18 base spec, the integration path Cullen Jennings May 27 explicitly preferred (“I disagree it should be done as a separate draft. I think it should be worked as a PR that can be discussed in context”) against Magnus Westerlund’s separate-Internet-Draft proposal. The consensus-decided design from the May 26-Jun 4 mailing-list call (Will Law / Gwendal Simon / Yu You-Nokia / Ali Begen = 4 YES votes) now has concrete normative text staged for merge, adds a new SWITCHING-SET-ASSIGNMENT parameter, details subscriber operations + relay behavior + bandwidth allocation algorithm. 8 days before London Day-1 0900-1045 “MOQT Issues” slot, the DTS thread shifts from list-debate to PR-review.
MSF 3 PRs MERGED June 1 (the trailing editorial work for -01 finally completes): PR #159 MERGED 09:42 UTC “Add catalog compression support via Track/Object Properties” by suhasHere (+114/−4, fixes #144 “Compression for the catalog” by vasilvv that was 21 days OPEN); PR #165 MERGED 09:33 UTC “Update bitrate and related properties in draft” by wilaw (+51/−25 — adds Maximum GOP Duration / Maximum Group Duration / Average Bitrate track properties, makes sample_rate+channels required for audio, codec+width+height required for video, redefines bitrate as maximum); PR #175 MERGED 13:28 UTC “Change version type from Number to String in MSF” by wilaw (+19/−16, fixes #163 “Version should carry draft info for interop until released”). With these merged, the last remaining MSF PR from wilaw’s May 24-27 sprint (PR #165) finally lands; only PR #156 (suhasHere “MOQT Object to Stream mapping implementation-specific”) + PR #169 (wilaw MOQT mapping clarification) remain OPEN. MSF -01 STILL NOT submitted to Datatracker — Day +134 since -00 (Jan 19 2026). 5th consecutive day of slippage on wilaw’s May 27 Slack “Friday” pledge.
moq — kixelated 14-merge + 6-OPEN burst June 1-2 (clearing pre-London backlog including 3 long-pending PRs):
- PR #1587 OPEN Jun 2 02:25 UTC “moq-mux: add MPEG-TS (transport stream) import and export” (+1548/−2, 15f) — bridges MoQ ↔ MPEG-2 TS via new container/ts demux/mux, enables ffmpeg -f mpegts - | moq-cli publish and moq-cli subscribe --format ts | ffplay - without transcode. Single mpeg2ts crate handles both directions, codecs H.264/H.265/AAC. Aligns with Paul Gregoire’s MSFTS draft + mondain/moq2ts C++ demonstrator announced same day. CLAUDE.md update: “prefer maintained third-party crate over hand-rolling non-core functionality” — first explicit codebase architectural note about scope discipline.
- PR #1573 MERGED Jun 1 16:09 UTC moq-lite-05 AnnounceOk (was OPEN in June 1 entry; merged ~13h after open into dev branch).
- PR #1540 MERGED Jun 1 15:34 UTC “moq-net: make subscribe_track async, blocking on SUBSCRIBE_OK” (+1221/−632, 55f) — long-pending since May 29 (4 days OPEN); reshapes subscribe_track so subscription resolves once publisher confirms via SUBSCRIBE_OK, separates Subscription (subscriber wire params) from Track (publisher immutable properties), concurrent subscribers coalesce. Picks up the idea started in PR #1439 MERGED 19:13 UTC “Add per-track timescale and frame timestamps to moq-lite” — even longer-pending PR that finally lands.
- PR #1513 MERGED 19:03 UTC “moq-net: map MoQ versions to required qmux versions” — concrete code for the WG-decided qmux pinning table (moq-transport-18 must ride on qmux-01, 14-17 on qmux-00, moq-lite unconstrained). Long-pending since May 28.
- PR #1581 MERGED Jun 2 02:25 UTC “relay: unified —auth-api” (+576/−24, 6f) — one HTTP call returns key + public + alias, supports MoQ CDN dashboard (moq-dev/moq-pro#47) giving each project a stable id plus an editable vanity path. Continues moq-pro pattern: external API resolves demo → x7k2qp server-side; relay stays topology-agnostic.
- Plus PR #1585 moq-net per-track cache age in SUBSCRIBE_OK + reconcile draft field order + port timescale to js/net; PR #1586 libmoq moq_error() exposed; PR #1579 don’t advertise illegal qmux-00.moqt-18 pair; PR #1580 WebSocket keep-alive on client path; PR #1582 re-export Hang from @moq/publish/@moq/watch; PR #1576 split TrackConsumer into track handle + TrackSubscriber; PR #1577 shrink moq-ffi + libmoq staticlibs with LTO; PR #1584 prefer esm.sh over jsDelivr for no-build CDN usage; PR #1589 moq-boy exit non-zero on reconnect give-up.
- OPEN follow-ons: PR #1590 out-of-band avc1/hvc1 in MPEG-TS export; PR #1588 decouple download gating from Renderer; PR #1591 Computed derived signals; PR #1592 @moq/watch component signals inputs vs outputs; PR #1583 OPEN by external contributor nuts-rice “js: add Opus dtx for voice WIP”.
moq-rs PR #171 OPEN June 1 18:40 UTC by suhasHere “Add pluggable AuthHook trait and C4M token authentication” (+2090/−15, 30 files) — implementation of PR #169 design proposal from May 28; adds 2 new crates (moq-auth AuthHook trait + AllowAll/KeyValue/Logging hooks + auth types; moq-auth-cat C4M hook with signature verification, claims validation, MOQT scope matching), parses AUTHORIZATION TOKEN from CLIENT_SETUP, calls on_setup at session establishment (fail-closed), on_request before Publish/Subscribe/TrackStatus, CLI flags --auth-shared-secret + --auth-cat-public-key. First concrete cross-impl auth implementation engagement after May 30 thibmeu 8-comment review burst on PR #169 — Suhas advances from trait proposal to working implementation within ~4 days. PR #169 itself remains OPEN; PR #170 (Manish draft-16 rewrite) untouched Day +3.
Slack moq — London draft-version coordination breaks May 28-Jun 1 7-day silence:
- Paul Gregoire June 1 19:43 UTC: “Created a demonstrator for the msfts draft https://github.com/mondain/moq2ts — only works with one relay that I know of at this time, but no one should be surprised by that” — first publicly-announced reference implementation of draft-gregoire-moq-msfts-00 (the MSFTS individual draft Paul co-authored with Gwendal Simon, submitted May 6).
- Martin Duke June 1 22:43 UTC: “Is anyone coming to London interested in draft-16 interop? There is no way we’re going to be done with draft-18 migration, but if everyone else will be we might as well break -16.” — reactions: gb (1), face_with_monocle (1). First London-week multi-version-strategy question.
- Alan Frindell June 2 01:50 UTC: “moxygen will have support for 14 and 16, and hopefully enough 18 to get some interop” — 3-version simultaneous support (14, 16, 18) for openmoq/moqx-stack.
- Mike English June 2 02:28 UTC: “moq-rs will also have instances of 14, 16, and a small start on 18 for people to test against.” + “I’m also feeling like 18+ might be where we finally start on proper multi-version support” — first explicit framing of draft-18 as a multi-version-support inflection point for cloudflare/moq-rs.
- Carry-forward: 4 impl-team responses to Martin Duke’s call within 6 hours = London 2026 will exercise (14, 16, 17, 18) × N-impls cross-version matrix manually, beyond the automated runner; this is the first explicit pre-London coordination on how many drafts the hackathon should actually try to interop against, vs. unilateral “everyone’s on 18” assumption.
Implementations: see TL;DR above for moq (14 merges + 6 OPEN), moq-rs (PR #171). moq2ts (created May 21, language: C++) — surge June 1-2 (~9 commits) on cross-platform CI build workflow (Linux/macOS/Windows via GitHub Actions, AppImage + macdeployqt + windeployqt packaging), ffmpeg 8 + Qt qsizetype build fixes, native AVCaptureSession publish on macOS (libav/avfoundation hangs); links against prebuilt moqxr publisher SDK (pinned MOQXR_VERSION=v0.3.2) for relay-capable builds. moqxr 3 commits June 1 by Paul Gregoire packaging the publisher static library + bundling picoquic/picotls dependencies for the SDK consumer path (enables moq2ts above to link standalone). moqx afrind multi-thread relay sprint continues June 1 — PR #361 relay_thread config + allow > 1 thread, PR #362 isolate relay state on dedicated executor for multiple I/O threads, PR #363 MultiThread relay test mode, PR #364 cache as passive subscriber of primary forwarder, PR #365 per-thread local forwarder data path (use_local_forwarders) — 5-PR stack OPEN simultaneously for the threads > 1 production-deployment milestone. Plus PR #371 bpf reuseport refinement MERGED, PR #370 by gmarzot folly XLOG init OPEN. quiche moqt breaks Day +5 silence — 2 commits June 1 by martinduke: 3b9d5450 17:28 UTC “Move parameter handling from session to RemoteTrack” + 0b92a8b4 17:34 UTC “Get rid of MoqtUpstreamFetch::LocationIsValid because FETCH streams are diff-encoded; Malformed Tracks are no longer possible to encode in a FETCH stream”. moqtail, moq-js (PR #72 still OPEN), imquic (PR #27 still OPEN, last update Jun 1 13:15), quiche_moq, Moqintosh, moqlivemock, Eyevinn/warp-player, Eyevinn/moqtransport all quiet.
Interop: 177 / 51 / 125 / 1 at 2026-06-02 00:50:22 UTC — first run at new draft-18 target following PR #68 merge June 1 05:21 UTC. −3 pass vs June 1 (54 → 51, 30.5% → 28.8%, −1.7pp); 15 consecutive days of daily reports (May 19-Jun 2), longest cadence streak the wiki has tracked. Version breakdown shifts to 0 at target · 0 ahead · 177 behind — NO impl is registered with the runner as advertising draft-18 even though moq, moqxr, and imquic have draft-18 code in main. Drift between runner-registered-version and impl-main-branch-version is now the visible structural gap (Martin Duke’s Slack question 4h later is the implementer-side response). London hackathon 7 days away.

moq-transport PR #1638 — Dynamic Track Switching lands as base-spec PR

Will Law opens DTS PR against moq-transport-18 June 1 11:06 UTC

PR #1638 “Add Dynamic Track Switching (DTS)” by wilaw — +147 / −0, 1 file, fixes Issue #259:

Adds the Dynamic Track Switching (DTS) feature, detailing subscriber operations, relay behavior, and bandwidth allocation algorithm. Adds a new SWITCHING-SET-ASSIGNMENT parameter.

Context — consensus call disposition:

May 26: Will Law explicit YES on adoption + integration into moq-transport-18 (mailing list).
May 27: Cullen Jennings explicit YES, prefers PR-against-moq-transport-18 over separate companion RFC: “I disagree it should be done as a separate draft. I think it should be worked as a PR that can be discussed in context”.
May 27: Gwendal Simon (SWITCH co-author) 4-point structured rebuttal of Martin Duke’s “Thoughts on SWITCH” framing — down-switch is 4 messages not 3 with break-before-make hard-freeze; up-switch N+k group selection needs relay-side info; SWITCH is additive not replacive; 4 implementations exist.
May 28: Gwendal Simon YES on adoption + YES on integration with DTS reservations.
May 29: Yu You (Nokia) YES + YES with measured benefit numbers (“my team has already implemented both”).
May 29: Ali Begen (MOQtail co-author) YES + YES with “if necessary” framing on separate-RFC fallback (“my team has already implemented both in MOQtail and plans to share implementation results”).
June 4: consensus call formal close.

Tally heading into close: 4 YES + 0 NO with 3 of 4 SWITCH co-authors + Nokia implementation report = consensus call substantively decided 14 days before close. PR #1638’s structural choice (PR-against-transport-18 base spec instead of separate companion Internet-Draft) matches Cullen’s preference and the implementer pattern (MOQtail + Nokia both implemented DTS as integrated transport feature, not as separate spec) rather than Magnus Westerlund’s separate-Internet-Draft proposal that would have given the WG more rev-cycle flexibility.

Wider impact:

London Day-1 0900-1045 “MOQT Issues” 180-min slot now structurally anchored around: bidi-stream-credit (Issue #1637 martinduke May 29) + request-blocking design (Issue #1519 vasilvv since Mar 2) + DTS PR #1638 (wilaw June 1) + EXPIRES extension (Issue #1639 + PR #1640 martinduke June 1 19:50/22:14 UTC).
DTS PR #1638 + EXPIRES PR #1640 are the two normative-text deliverables for London Day-1; the others are issue-level discussion.

moq-rs PR #171 — AuthHook trait + C4M implementation lands

Suhas Nandakumar opens implementation PR June 1 18:40 UTC

PR #171 “Add pluggable AuthHook trait and C4M token authentication” by suhasHere — +2090 / −15, 30 files:

This PR is an implementation of #169. Adds intra-scope authorization to the relay via a pluggable AuthHook trait, with a C4M (CAT for MoQ) implementation using the cat-token crate.

New crates:

moq-auth — AuthHook trait, AllowAll/KeyValue/Logging hooks, auth types
moq-auth-cat — C4M hook: signature verification, claims validation, MOQT scope matching

Relay integration:

Parses AUTHORIZATION TOKEN from CLIENT_SETUP (USE_VALUE wire format)
Calls on_setup at session establishment (fail-closed)
Calls on_request before Publish/Subscribe/TrackStatus operations
CLI: --auth-shared-secret (token type 0) and --auth-cat-public-key (C4M, requires --features auth-cat)

Why this matters:

First concrete cross-impl auth implementation engagement after May 30 thibmeu 8-comment review burst on PR #169.
The PR text includes end-to-end test plans for both shared-secret and C4M auth — Suhas demonstrates working code rather than gating on PR #169 design-discussion closure.
The thibmeu review on PR #169 critiqued the trait surface for accommodating Privacy Pass challenge-reply (issuer-aware) — PR #171’s implementation does not yet address that pivot; Suhas chose to ship the verify-callback shape and let Privacy Pass adapt or extend later. Design vs implementation cleavage now visible: PR #169 stays open as design-discussion forum, PR #171 stays open as concrete implementation to test against.
Cross-impl significance: openmoq/moqx has been driving CAT token auth via PR #264 (paul-mondain) + PR #286 (Catapult submodule) since May 1; cloudflare/moq-rs PR #171 is the second relay implementation of CAT auth under the C4M framing, structurally parallels rather than reuses openmoq/moqx’s pattern.

Carry-forward: London Day-1 PRIVACY_PASS slot (paired with moq-privacy-pass draft-02 advance) now has two concrete relay implementations (openmoq/moqx CAT + cloudflare/moq-rs C4M) plus thibmeu’s pending Privacy Pass critique on PR #169 — three vectors of auth-implementation experience to inform draft-19 design.

mondain/moq2ts — MSFTS demonstrator announced

Paul Gregoire publicly announces moq2ts on moq June 1 19:43 UTC

mondain/moq2ts (C++, created May 21 2026, “MOQ + M2TS”) — first publicly-announced reference implementation of draft-gregoire-moq-msfts-00:

Created a demonstrator for the msfts draft https://github.com/mondain/moq2ts — only works with one relay that I know of at this time, but no one should be surprised by that 😉

Surge May 31-Jun 2 (~9 commits):

Cross-platform CI build workflow (8306eee9 June 1 19:50 UTC) — Linux/macOS/Windows via GitHub Actions, AppImage via linuxdeploy + macdeployqt + windeployqt+ldd packaging.
Real relay-capable binaries by linking moqxr SDK (5721c914 June 1 22:34 UTC) — links against prebuilt openmoq publisher SDK pinned MOQXR_VERSION=v0.3.2, Linux wraps archives in --start-group to resolve circular picohttp/picoquic deps.
ffmpeg 8 + Qt qsizetype build fixes (7c143817 22:56 UTC, aef64cd0 23:29 UTC) — avio_alloc_context const-buffer change for ffmpeg 8 + cross-version compat via LIBAVFORMAT_VERSION_MAJOR check.
Native AVCaptureSession publish on macOS (8f281a43 June 2 00:29 UTC) — libavdevice/avfoundation hangs in avformat_open_input because it does not pump AVCaptureSession run loop from worker thread; uses native AVCaptureSession directly for preview + publish paths (uyvy422→BGRA in delegate, audio as packed/planar PCM).
README/DEVELOPER refresh for current MSF + MSFTS catalog (06a17f7a + 4e377e17 June 1 19:31-19:34 UTC) — documents mediatimeline MSF type for the timeline side-track + compact [mediaTimeMs,[groupId,objectId],wallclockMs] record array.

Synchronicity: same June 2 02:25 UTC window, moq PR #1587 opens MPEG-TS import/export support for moq-mux via single mpeg2ts crate — both moq2ts (C++/MSFTS draft-specific) and moq-dev/moq PR #1587 (Rust/codec-agnostic) advance MPEG-TS↔MoQ bridging within the same 24-hour window. First concrete cross-impl MPEG-TS interop substrate.

Carry-forward: with mondain/moqxr SDK packaging (a8ad5f9d/deea689f/d43d9804 June 1) + moq2ts CI builds linking against it + moq-dev/moq PR #1587 import/export, MSFTS as a streaming-format becomes the second concrete MoQ-side MPEG-2 TS path (after moqlivemock catalog work) — directly relevant to London Day-2 35-min MSF/CMSF/MSFTS slot.

Slack moq — Martin Duke draft-16-or-18 London poll

Martin Duke poses draft-version question June 1 22:43 UTC

Martin Duke Slack moq June 1 22:43 UTC:

Hi, is anyone coming to London interested in draft-16 interop? There is no way we’re going to be done with draft-18 migration, but if everyone else will be we might as well break -16.

Reactions: 🇬🇧 (1), 🧐 (1).

Responses (within ~6h):

Alan Frindell June 2 01:50 UTC: “moxygen will have support for 14 and 16, and hopefully enough 18 to get some interop”.
Mike English June 2 02:28 UTC: “moq-rs will also have instances of 14, 16, and a small start on 18 for people to test against.” + 02:29 UTC: “I’m also feeling like 18+ might be where we finally start on proper multi-version support”.

Cumulative London-week version inventory (from these 3 responses):

Impl	-14	-16	-17	-18
moxygen (openmoq/moqx)	✓	✓	—	partial
moq-rs (cloudflare)	✓	✓	—	partial
moq-dev/moq	—	✓	✓	✓ (main)
moqxr (mondain)	—	—	—	✓ (main)
imquic (meetecho)	—	—	✓	partial
moqtail	✓	—	—	—

Why this matters:

Draft-18 was published May 12 — 21 days before London but no impl has full draft-18 conformance as of June 2; the partials are sufficient for some interop only.
Martin Duke’s question reframes the matrix: not “everyone migrates to 18 by London” but “how many drafts do we actually run cross-impl interop against at London”. The default expectation appears to be 14 + 16 + partial 18 = 3-version matrix during London week.
Mike English’s “multi-version support” comment is a structural signal: until now, MoQ impl strategy has been “track latest draft, migration is best-effort per-impl”; Mike is signaling that draft-18 may be the version where multi-version support becomes a deliberate impl feature, not just a side effect of slow migration. This would echo HTTP/2 multi-version support patterns where impls keep older draft codepaths intentionally to support stragglers.

Carry-forward: the interop runner is automated-version-locked (one target per matrix run); London Day-week interop is manual and operator-driven. Mike’s PR #68 draft-18 target bump landed June 1, but the matrix is now structurally ahead of every impl (0 at target · 0 ahead · 177 behind) — which gives the manual London interop work latitude to exercise 14/16/18 cross-version freely without contradicting the automated matrix.

Activity (May 31 06:00 UTC → June 1 06:00 UTC) — interop runner PR #68 MERGED draft-18 target bump (the matrix-shape lever the wiki has been tracking 14 days); MoQ Monthly #2 published ending Day-+31 silence with London framing + Dan Rayburn town hall + CacheFly/Red5 CDN beta; moq-dev/moq cluster-mesh infrastructure burst (5 merges + 1 OPEN in 12h, externalizes peer-list to operator-owned endpoint + deterministic FNV-1a route tie-break + per-auth-root billing); openmoq/moqx afrind PR #359 fixes relay shutdown hang; mondain/moqxr 2 cancellable-flush commits; moq-wg + mailing list + Slack all quiet; interop 177/54/122/0 (+2 pass, 30.5%, 14-day cadence longest streak, 5-day monotonic uptick to new May+June high, first 30%+ pass rate)

TL;DR:

englishm-cloudflare merges interop runner PR #68 “Update interop target to draft-18” June 1 05:21:25 UTC — 14 days after OPEN (May 18 21:01 UTC). First matrix-shape change since PR #71 (moqx docker adapter, May 25) and first interop-target version bump since matrix inception. The June 1 00:49 UTC report still shows draft-16 as the target (PR merged ~4.5h after report cut), so the June 2 report will be the first run targeting draft-18. Three implementations (moq-dev/moq, mondain/moqxr, meetecho/imquic) have been on draft-18 main for 2-3 weeks; the matrix is finally catching up. London hackathon 8 days away — only structural matrix-shape lever the wiki had been tracking lands precisely in the pre-London window.
MoQ Monthly #2 published May 31 by Mike English — “A new draft, an industry town hall, and London next week”. Ends Day-+31 silence since #1 (Apr 30); cadence reasserted just before London. Coverage: Draft-18 publication May 12 (unified URI schemes, separated subscription messages, reserved namespaces for extensions); London interim June 9-12 at Cloudflare’s office; WGLC targeted after July Vienna plenary, IESG publication Dec 2026; Dan Rayburn town hall May 12 (commercial-interest signal); Streaming Tech Sweden May 21 featured Vindral + SVT production implementations; Luke Curley’s “MoQ Boy” Game Boy emulator demo; CDN signal: Fastly positioning Apr 15, CacheFly + Red5 announce MoQ beta for summer 2026, Cloudflare + Akamai existing initiatives; 3 ACM Multimedia Systems 2026 papers from Özyeğin University researchers; new MPEG-2 TS + neural video codec drafts; Jan Ozer comprehensive overview at Streaming Learning Center; upcoming FOKUS Media Web Symposium (June 16-17, Berlin), CommCon 2026 (June 9-11, Düsseldorf), RTC.ON 2026 (Sep 16-18, Kraków).
moq-dev/moq — kixelated cluster-mesh infrastructure burst May 31 17:51 UTC → June 1 03:20 UTC. 5 merges + 1 OPEN in ~12 hours: PR #1569 MERGED 17:51 UTC “relay: dedup mesh dials with a URL-order tiebreaker” (+30/−2) — gossip-discovered peers only dialed by the lexicographically-smaller node, eliminates redundant outbound dial per pair. PR #1570 MERGED 19:26 UTC “moq-net: deterministic route tie-break for equal-length paths” (+83/−18) — FNV-1a (hop_count, hash) lex compare so every node converges on the same winner across rolling deploys (DefaultHasher explicitly not stable across Rust versions). PR #1571 MERGED June 1 00:22 UTC “moq-relay: add --cluster-connect-api and split cluster identity from gossip” (+874/−150, 10f) — externalizes peer-list source to an http(s) URL or local file returning bare JSON array of hostnames, polled with Cache-Control semantics + ETag/Last-Modified conditional revalidation + fail-static on error. Re-adds --cluster-node <self-url> for identity; --cluster-mesh becomes boolean gossip toggle (breaking change to recent flag from PR #1504). Relay stays topology-agnostic: all routing decisions live in whatever service answers the endpoint (the /cluster/connect endpoint itself lives in moq-pro). PR #1572 MERGED 01:11 UTC “simplify cluster-connect-api polling onto the HTTP cache” (+38/−93) — rebases #1571 onto dev, delegates freshness to existing http-cache-reqwest middleware instead of hand-parsing Cache-Control (net −58 LOC). PR #1574 MERGED 03:20 UTC “moq-relay: count connected sessions per auth root for billing” (+355/−43, 3f) — new sessions.json (external) + internal/sessions.json (internal) stats tracks mapping auth root → {sessions, sessions_closed} for presence-based billing regardless of data flow (idle authenticated session billable, derived from (broadcast, session) atomics is not). PR #1573 OPEN 02:47 UTC “moq-lite-05: add AnnounceOk message (responder origin + initial active count)” (+493/−36, 16f) targeting dev — second concrete moq-lite-05 wire feature after PR #1531 deflate compression May 28: publisher sends AnnounceOk once after reading AnnounceInterest, reports responder’s origin id once (replaces per-Announce trailing-hop redundancy) + active: N count followed by N initial Announce::Active (discrete initial-set boundary, successor to AnnounceInit); enables connect() to block until initial set has landed via new SyncLatch, closes startup race where synchronous get_broadcast() post-connect could miss broadcasts live-but-not-yet-gossiped. JS mirror lands the same wire change.
Implementations: moq see above (5 merges + 1 OPEN; PR #1575 chore-release closes #1557; dependabot 1568 docker/setup-buildx/flake-checker bumps merged). moqx — afrind PR #359 MERGED May 31 22:08 UTC “relay: fix shutdown hang when sessions hold lingering server refs” (+31/−0, 5f) — main() now owns IOThreadPoolExecutor via unique_ptr, servers hold raw pointer; explicit stop + clear before joining IO pool to prevent QuicServer::shutdown posting to dead worker EVBs (would hang shutdown until 10s watchdog fired). MoqxRelayServer::stop() + MoqxPicoRelayServer::stop() made idempotent via context_ sentinel. Plus 2 sync-bot moxygen merges (PR #357 b3fc363 10:35 UTC + PR #358 e282ba6 14:46 UTC). moqxr 2 commits June 1 by Paul Gregoire: ced45c85 03:38 UTC “Make publish_live_objects cancellable via close() stop flag” + c18924cc 03:44 UTC “Bound the publish_live_objects graceful flush on stop” — incremental shutdown/cancellation polish. moq-rs quiet Day +2 since 8-thibmeu review burst May 30; PR #169 AuthHook and PR #170 (Manish draft-16 rewrite) both untouched. quiche moqt Day +2 silent. moqtail, moq-js (PR #72 still OPEN), imquic (PR #27 still OPEN), quiche_moq, Moqintosh, moqlivemock, Eyevinn/warp-player, Eyevinn/moqtransport all quiet.
Interop: 177 / 54 / 122 / 0 at 2026-06-01 00:49:16 UTC — +2 pass vs May 31 (52 → 54, 29.4% → 30.5%, +1.1pp). First 30%+ pass rate since cadence recovery May 19. 14 consecutive days of daily reports (May 19-Jun 1), longest cadence streak the wiki has tracked. Rolling 5-day band 50-54, trajectory 49 → 50 → 52 → 54 = 5-day monotonic uptick continues (counting May 28→Jun 1: 46 → 49 → 50 → 52 → 54, +8 pass over 5 days). Target at report time still draft-16 but PR #68 MERGED ~4.5h after report cut — June 2 will be the first draft-18 run. London hackathon 8 days away.

Interop runner — PR #68 MERGED draft-18 target bump

Mike English (englishm-cloudflare) merges PR #68 June 1 05:21:25 UTC

PR #68 “Update interop target to draft-18” — 14 days OPEN → MERGED:

OPENED: May 18 21:01 UTC (Slack announcement same day: “Working through some CI issues, and then I’ll be bumping the interop target in the automated interop test runner to draft-18, too”).
MERGED: June 1 05:21:25 UTC by englishm-cloudflare (Mike English, Cloudflare/interop runner maintainer).
Diff: 1 file, +1/−1 (the target version string in the config).
Single commit: f6db6cec May 18 20:57 UTC — i.e. the actual change has been ready for 14 days; the merge was gated on CI infrastructure stabilization (May 19 PR #69 per-test timeout fix → matrix cadence resumes; May 25 PR #71 moqx docker adapter; subsequent 13 daily reports).

Why this matters:

First matrix-shape change since PR #71 May 25 (moqx docker adapter): 7 days of “matrix shape unchanged at 177 tests” since the last lever moved.
First interop-target version bump since matrix inception — the matrix has tracked draft-14, draft-15, draft-16 over the past months as implementations updated; this is the first time the matrix’s target version steps forward.
Catches up to three implementations on draft-18 main: moq (PR #1418 May 18), moqxr (8 commits May 19-20), imquic (May 20 09:25 UTC). The structural gap (matrix-on-draft-16 vs three impls-on-draft-18) closes precisely 8 days before London.
Expected June 2 matrix impact: Version breakdown of 177 tests was 97 at target (draft-16) · 8 ahead (draft-17) · 72 behind (draft-14) under the draft-16 target. Under a draft-18 target, the “at target” / “ahead” / “behind” decomposition will shift: implementations on draft-18 are now “at target” (previously “ahead”), draft-17 becomes “behind” (previously “ahead”), draft-16 becomes “behind” (previously “at target”). Whether this nets pass-count up or down depends on whether the matrix harness counts cross-version pairings as expected-fail or genuine-fail.

Carry-forward: with 14 consecutive days of daily cadence and a +8 pass count over 5 days under the old target, the matrix has clearly stabilized; the draft-18 bump introduces a one-time shape change whose impact will only show on June 2. No further structural levers remain in the matrix before London — only individual-impl draft-18-conformance fixes can move pass counts now.

Carry-forward (London Day-2 slot): Mike’s May 27 Slack request for streaming-format-level automated interop (MSF/CMSF/LOC matrix beyond wire-protocol matrix) is now the next structural feature request on the runner, downstream of the draft-18 target landing — Mike signaled interest, no implementation has started.

Mike English publishes MoQ Monthly #2 May 31

MoQ Monthly #2 — “A new draft, an industry town hall, and London next week”:

Cadence: #0 (Mar 3) → #1 (Apr 30, Day +58) → #2 (May 31, Day +31) — cadence shortened from 58 days to 31 days, closing toward true monthly cadence.
Timing relative to London: published 8 days before London hackathon (June 9), 9 days before interim Day-1.

Coverage:

Draft-18 publication May 12 — unified URI schemes, separated subscription messages, reserved namespaces for extensions. WGLC after July Vienna plenary; IESG publication December 2026.
London interim June 9-12 at Cloudflare’s office.
Industry signal:
- Dan Rayburn town hall May 12 (commercial-interest indicator outside IETF).
- Streaming Tech Sweden May 21 featured production implementations by Vindral and SVT.
- Luke Curley’s “MoQ Boy” Game Boy emulator demo (low-latency capabilities + resource management showcase).
CDN ecosystem growth:
- Fastly published positioning April 15.
- CacheFly + Red5 announced MoQ beta for summer 2026.
- Cloudflare + Akamai existing initiatives.
Research & standards:
- 3 ACM Multimedia Systems 2026 papers from Özyeğin University researchers (Ali Begen’s group).
- New layering drafts for MPEG-2 Transport Stream (moq-msfts) and neural video codecs (moq-nmsf).
- Jan Ozer comprehensive overview at Streaming Learning Center.
Upcoming events:
- FOKUS Media Web Symposium (June 16-17, Berlin) — right after London week.
- CommCon 2026 (June 9-11, Düsseldorf) — overlaps London Day-1/2.
- RTC.ON 2026 (Sep 16-18, Kraków).

Significance for wiki: First MoQ Monthly issue covering the wiki-tracked May 2026 window in retrospective form — gives third-party validation of the activity the wiki recorded in 2026-05 discussions (the Dan Rayburn town hall, Luke’s MoQ Boy demo, the SVT/Vindral Streaming Tech Sweden coverage). The CacheFly + Red5 MoQ beta is new ecosystem signal the wiki had not yet captured.

Carry-forward: MoQ Monthly’s revived cadence positions it as a per-London-cycle external news source — #3 (June 30?) would cover the London interim outcomes in retrospective.

moq-dev/moq — cluster-mesh infrastructure burst May 31 → June 1

5 merges + 1 OPEN in ~12 hours by kixelated. Theme: externalize cluster topology to an operator-owned endpoint + deterministic route tie-break + per-auth-root billing + moq-lite-05 AnnounceOk wire feature.

Mesh-dial dedup + deterministic route tie-break

PR #1569 MERGED May 31 17:51 UTC “relay: dedup mesh dials with a URL-order tiebreaker” (+30/−2). Gossip-discovered peers dialed only if URL sorts after self (peer > self_url). Lexicographically-smaller node is client, larger is server. Inbound connection still arrives for the skipped side. Bidirectional cluster session means one connection suffices; eliminates redundant outbound dial per pair. Replaces old peer == self_url self-skip. Scope: gossip only — explicit --cluster-connect always dials, passive-rendezvous via static path.
PR #1570 MERGED 19:26 UTC “moq-net: deterministic route tie-break for equal-length paths” (+83/−18). When two announcements compete for the same broadcast path with equal hop counts, the old code kept whichever arrived most recently (arrival order differs per node → relays in a cluster could pick different routes and flap). New route_key(name, hops) -> (usize, u64) returns (hop_count, FNV-1a hash over name + hop chain); lex compare, lowest hash wins on tie. FNV-1a instead of DefaultHasher because std’s output is explicitly not stable across Rust versions — during rolling deploys, mismatched binaries must still agree on a route. Mixing the broadcast name in spreads equal-length routes across upstreams instead of funneling onto one. On exact key tie, incumbent stays (no churn).

Cluster-connect-api: externalize peer-list to operator endpoint

PR #1571 MERGED June 1 00:22 UTC “moq-relay: add --cluster-connect-api and split cluster identity from gossip” (+874/−150, 10f). New --cluster-connect-api flag accepts an http(s) URL or local file path returning bare JSON ["a.pop.example", "b.pop.example"]. Reconciles dials at runtime: new peers dialed, dropped peers aborted. Composes with static --cluster-connect (never reconciled away) and gossip. HTTP polled with Cache-Control semantics (max-age + stale-while-revalidate); conditional revalidation (ETag / Last-Modified); fail-static on error. File re-read when mtime changes. The relay’s --cluster-node value is sent as ?node=, the cluster mTLS client cert identifies the caller, so the endpoint can return per-node peer lists.
- Re-adds --cluster-node <url> as relay’s own identity (used for ?node= query parameter and as the address gossip advertises).
- --cluster-mesh becomes boolean gossip toggle (was URL). Enabling without --cluster-node errors at startup. Breaking change to a recently-added flag (originally added in PR #1504 May 24).
- The /cluster/connect endpoint itself (proprietary routing over node inventory) lives in moq-pro. moq-dev/moq stays topology-agnostic; all routing decisions move out to whatever service answers the endpoint.
PR #1572 MERGED June 1 01:11 UTC “simplify cluster-connect-api polling onto the HTTP cache” (+38/−93). PR #1571 hand-rolled stale-while-revalidate parsing; #1572 rebases onto dev and delegates to existing http-cache-reqwest middleware. Net −58 LOC. PR body documents the crate landscape check (http-cache 0.21.0 = RFC 7234 only; http-cache 1.0.0 alpha-only; moka dropped background threads in 0.12). Re-checks endpoint on fixed 30s cadence; cached list served with no network round-trip while fresh; conditional GET when stale; serves stale on revalidation failure.

Per-auth-root billing

PR #1574 MERGED 03:20 UTC “moq-relay: count connected sessions per auth root for billing” (+355/−43, 3f). Presence-based billing needs to know how many sessions are connected to a node keyed by auth root. Existing per-broadcast stats only tracked subscriptions + per-(broadcast, session) sentinel; node-level session count not derivable. Two new stats tracks: sessions.json (external) + internal/sessions.json (internal), each a JSON object mapping auth root → {sessions, sessions_closed}. Counts presence regardless of data flow — guard created in relay’s connection handler where token.root is known, held for whole connection. RAII SessionStats guard via new StatsHandle::session(root). Example: {"acme": {"sessions": 3, "sessions_closed": 1}, "globex": {"sessions": 1, "sessions_closed": 0}}.

moq-lite-05 AnnounceOk wire feature

PR #1573 OPEN June 1 02:47 UTC “moq-lite-05: add AnnounceOk message (responder origin + initial active count)” (+493/−36, 16f) targeting dev. New AnnounceOk message on the announce stream, sent once by publisher right after reading AnnounceInterest and before any Announce. Two purposes:
1. Reports responder’s origin id once instead of stamping it onto trailing hop of every Announce. In Lite05, node no longer stamps its own origin; receiver stamps the remote sender’s origin (from AnnounceOk) on receipt. Stored hop chain byte-identical to Lite04 (loop detection / shortest-path selection unchanged).
2. Reports active: N count of currently-active broadcasts, followed by exactly N initial Announce::Active. Gives announce stream a discrete initial-set boundary (the successor to AnnounceInit).
Enables connect-blocking: lets connect() block until initial set has landed via new SyncLatch (fires once every announce-prefix stream has its initial set; generalized across AnnounceInit Lite01/02 + AnnounceOk + N Lite05; Lite03/04 have no boundary and resolve immediately). Closes startup race where synchronous get_broadcast() post-connect could miss broadcasts live-but-not-yet-gossiped.
Opt-in: Lite05Wip is not advertised over ALPN or default version set; nothing negotiates by default. Second concrete moq-lite-05 wire feature after PR #1531 deflate compression (May 28).
JS mirror: @moq/net mirrors the wire, but does NOT add connect-blocking — JS Connection is pull-based (announced() opens stream lazily, consume(path) subscribes directly without consulting announcements), so synchronous get_broadcast() race doesn’t exist there.
IETF draft / spec being updated separately.

Pattern: PR #1573 demonstrates the PR #1518 Lite05Wip unadvertised version variant working exactly as intended: features can land gated without wire exposure (no peer negotiates Lite05Wip yet because it’s omitted from ALPNS and Versions::all()). Same gating model as PR #1531 deflate.

openmoq/moqx — afrind PR #359 relay shutdown hang fix

PR #359 MERGED May 31 22:08 UTC by afrind “relay: fix shutdown hang when sessions hold lingering server refs” (+31/−0, 5f):

“main() now owns the IOThreadPoolExecutor via unique_ptr (servers hold a raw pointer) and explicitly stops + clears servers before joining the IO pool. Previously, lingering shared_ptr refs from in-flight sessions or coroutines could delay ~MoqxRelayServer past the executor’s destruction, leaving QuicServer::shutdown to post to dead worker EVBs and hang shutdown until the 10s watchdog fired.”

MoqxRelayServer::stop() + MoqxPicoRelayServer::stop() made idempotent using context_ as sentinel — ~MoqxRelayServer no-op once main has already stopped the server.
Follows the May 28 PR #351 “relay: IOThreadPoolExecutor owned exclusively by main” (multi-thread sprint Day 2) — completes the ownership model with explicit shutdown ordering.
Plus 2 automated moxygen submodule sync merges: PR #357 b3fc363 MERGED May 31 10:35 UTC + PR #358 e282ba6 MERGED 14:46 UTC.

Carry-forward: afrind’s multi-thread sprint Days 1-3 was May 27-29 (22 events). May 30-31 was quieter (2 sync-bot merges May 30 + this PR #359 + 2 sync-bot merges May 31). With London hackathon 8 days away, PR #331 (relay_thread config + allow > 1 thread) remains the only PR-merge-away gate from threads > 1 being usable in production.

mondain/moqxr — Paul Gregoire cancellable graceful flush

2 commits June 1 by Paul Gregoire:

ced45c85 June 1 03:38:17 UTC “Make publish_live_objects cancellable via close() stop flag”.
c18924cc June 1 03:44:35 UTC “Bound the publish_live_objects graceful flush on stop”.

Continues the May 30 21b791a8 “Fix live object catalog ordering and stop wakeup” (+53/−7) shutdown-polish theme. Incremental hardening of mondain/moqxr’s live-object publishing path for clean shutdown.

Mailing list — weekly digest only, weekend silence

Single message May 31 03:01:31 UTC: Weekly github digest (Media Over QUIC Activity Summary) from Repository Activity Summary Bot. Auto-summary of moq-wg repo activity:

moq-transport: 1 new issue (Issue #1637 “What does MOQT do without bidi stream credit?” + 4 ianswett/martinduke comments; 2 other issues received commentary).
moq-wg/msf: 1 new issue (“Reference the latest version of LOC”) + 17 new comments across 10 existing issues + 8 issues closed + 9 new PRs (timestamp rounding, buffer properties, init data, bitrate, compression).

No substantive on-list MoQ discussion May 31 or June 1. Weekend mailing-list silence persists into a 3rd day (May 30 / 31 / Jun 1). Ali Begen’s May 29 YES vote on SWITCH/DTS remains the latest substantive thread.

IETF Datatracker — no new revisions

No new drafts published May 31 or June 1. WG document state unchanged from May 31:

draft-ietf-moq-transport-18 (Day +20 since May 12)
draft-ietf-moq-msf-00 (Day +133 since Jan 19; wilaw’s May 27 Slack “Friday” pledge for -01 still unfulfilled into 4th consecutive day; editorial work on main complete — PR #166/167/168/171/173/174 all MERGED — only xml2rfc submission remains)
draft-ietf-moq-loc-02
draft-ietf-moq-secure-objects-00 (PR #88 still OPEN staging test vectors for -01)
draft-ietf-moq-privacy-pass-auth-02
draft-ietf-moq-cmsf-00 (Issue #122 “initial text on zapping” Gwendal-vs-Suhas editorial debate from May 30 unresolved)

Individual drafts also unchanged.

Slack — all quiet

No new messages May 31 or June 1 across #moq, #moq-rs, #moq-js, #libquicr, #moq-interop-runner. The afrind self-fetch thread (May 28) and Lorenzo moq-mi/LOC thread (May 27) remain the latest substantive Slack discussion.

MOQ Wiki

Explorer

Discussions - June 2026

London interim agenda PUBLISHED — Day-1 is MOQT-blocker time, DTS/SWITCH demoted

DTS consensus readout — “extension WG document”, and the Cullen reversal

Tobbe ships MSF/CMSF catalog validator (CUE) — finds draft-01 bugs

Kota Yatagai debuts Moqtopus; interop spreadsheet → automated runner

thibmeu opens privacy-pass design issues

moq-transport: EXPIRES + sharmafb editorial PRs merge; 3 new design issues

draft-ietf-moq-cmsf-01 PUBLISHED — 2nd WG draft in 2 days

LOCMAF Slack thread — initData carriage as 3-way design space

sharmafb 5-item draft-18 editorial sprint

google/quiche moqt starts draft-19 wire prep in earnest

openmoq/moqx — PR #264 paul-mondain CAT token MERGED after 33 days OPEN

TilsonJoji emerges at mondain/moqxr — SRT MPEG-TS ingest

moq-dev/moq voice-AI / TTS use case lands at the player

Yu You (Nokia) announces v17+v18 remote relay for interop

draft-ietf-moq-msf-01 PUBLISHED — 134-day saga closes

draft-einarsson-moq-locmaf-00 — First IETF artifact from the wiki maintainer

afrind PR #1642 — Joining Fetch removed, Subscription Fill introduced

Tim Evens (Cisco) emerges as new cross-repo contributor

moq-dev/moq operational milestone — /health endpoint

moq-lite-05 wire-feature triple — pace continues

moq-transport PR #1638 — Dynamic Track Switching lands as base-spec PR

Will Law opens DTS PR against moq-transport-18 June 1 11:06 UTC

moq-rs PR #171 — AuthHook trait + C4M implementation lands

Suhas Nandakumar opens implementation PR June 1 18:40 UTC

mondain/moq2ts — MSFTS demonstrator announced

Paul Gregoire publicly announces moq2ts on moq June 1 19:43 UTC

Slack moq — Martin Duke draft-16-or-18 London poll

Martin Duke poses draft-version question June 1 22:43 UTC

Interop runner — PR #68 MERGED draft-18 target bump

Mike English (englishm-cloudflare) merges PR #68 June 1 05:21:25 UTC

MoQ Monthly #2 — newsletter cadence reasserts before London

Mike English publishes MoQ Monthly #2 May 31

moq-dev/moq — cluster-mesh infrastructure burst May 31 → June 1

Mesh-dial dedup + deterministic route tie-break

Cluster-connect-api: externalize peer-list to operator endpoint

Per-auth-root billing

moq-lite-05 AnnounceOk wire feature

openmoq/moqx — afrind PR #359 relay shutdown hang fix

mondain/moqxr — Paul Gregoire cancellable graceful flush

Mailing list — weekly digest only, weekend silence

IETF Datatracker — no new revisions

Slack — all quiet

tobbee/moq-llm-wiki — no open issues

Graph View

Table of Contents

Backlinks

moq-dev/moq operational milestone — `/health` endpoint