01. Draft 21 moved the furniture.
An exasperated packet struggles with a wobbling stack of pages beside Document 21, above an unchanged cable marked Wire 20. The document was reorganized; the wire format stayed the same.
Seattle interop still targets Draft 20.
Source: https://blog.rvnd.me/moqmusings/2026-09-07-moq-went-to-amsterdam-and-came-back-with-a-network#draft-21-moved-the-furniture

02. The relay has an operator.
A quietly proud packet holds a clipboard and rests a spanner over its shoulder beside a relay cabinet beneath NetActuate’s announced footprint of more than 45 locations. The number describes the provider’s network footprint, not independently measured relay deployments.
A commercial service, with capacity and monitoring.
Source: https://blog.rvnd.me/moqmusings/2026-09-07-moq-went-to-amsterdam-and-came-back-with-a-network#someone-volunteered-to-run-the-relays

03. Less waiting. More television.
A delighted packet leans forward in an armchair with a remote and popcorn beside a television comparison: Zattoo’s production delay of 10 to 13 seconds versus about 2 seconds in its MoQ prototype. These are Zattoo’s reported test results; the prototype was not commercially available.
Vendor-reported tests. Not commercially available.
Source: https://blog.rvnd.me/moqmusings/2026-09-07-moq-went-to-amsterdam-and-came-back-with-a-network#zattoo-made-the-delay-much-smaller

04. One path. Two networks.
A crouching packet concentrates, tongue at the corner of its mouth, while joining two cable connectors between a relay in Oracle’s OCI network and an Oracle relay running inside Netskrt infrastructure. This illustrates the announced IBC demonstration between separately operated networks, not universal interoperability.
Separate operators, coordinating delivery.
Source: https://blog.rvnd.me/moqmusings/2026-09-07-moq-went-to-amsterdam-and-came-back-with-a-network#ibc-turned-the-relay-diagram-into-a-group-project

05. Audio travels lighter.
A determined packet squeezes its eyes shut and sits on an overstuffed case of metadata papers: more than 100 bytes of repetitive overhead become about 2 bytes in the discussed LocMAF proposal. The comparison concerns overhead, not the audio payload. The receiver reconstructs the full CMAF container.
Receiver rebuilds CMAF. Not yet adopted.
Source: https://blog.rvnd.me/moqmusings/2026-09-07-moq-went-to-amsterdam-and-came-back-with-a-network#meanwhile-audio-metadata-went-on-a-diet

06. From building it to running it.
A blueprint labelled Build leads to a relay cabinet labelled Deploy, then a weary but satisfied packet relaxes with coffee at a control desk under Operate. This is the author’s reading of the week, not a claim that the standard or ecosystem is finished.
The week’s signal: integration is becoming the work.
Source: https://blog.rvnd.me/moqmusings/2026-09-07-moq-went-to-amsterdam-and-came-back-with-a-network#the-signal-this-week
