MBONED IETF 112 2021 Nov 10 too:fucking:early Note takers: Jake Holland, Kyle Rose Status of WG items Chairs, 10 min iOS Multicast entitlement process Tommy Pauly, 15 min - slides: https://datatracker.ietf.org/meeting/112/materials/slides-112-mboned-multicast-entitlements-00.pdf - Jake: x.local names with mdns that's not dns-sd? - one-time popup to permit apps to communicate with local network, permission can be changed later in settings. Reason can be provided in popup. - Jake: any difference between ssm and asm? global vs. local scoped addresses? - nope - Jake: any changes likely on this front going forward? - nope. stayed the same for last release, we know of no other needs here. - Lenny: entitlement is not for users, right? - right, just a developer process at build time - Lenny: do they generally get approved on request? - yes, thousands approved. there were some that didn't, but most are legit, or we push them toward an api suited to their use case (e.g. bonjour) Multicast to the Browser: draft-ietf-mboned-ambi, draft-ietf-mboned-cbacc, draft-ietf-mboned-dorms, draft-ietf-mboned-mnat Jake Holland, 40 min - slides: https://datatracker.ietf.org/meeting/112/materials/slides-112-mboned-multicast-to-the-browser-update-00.pdf - for background on drafts, see IETF 111 summary: https://www.youtube.com/watch?v=3O3Zh1Hj6co&t=1044s - Sandy: Will the proposal require any changes to multicast group management protocols? - no, only transports and only from UDP up, nothing else would be affected, as far as i know. - Lenny: "There isn't enough interest" seems to be a subjective judgment, and contradicts the existence of work today, which presumes interest. - Ekr's contention is that Chromium people seem lukewarm on it. - Jake's impression is that Chromium folks want to see more interest before putting work into this. - Jake is seeing more interest from third parties (e.g., Comcast, Virgin Media) who are interested in multicast as a scalability measure. - Lenny: One of other concerns is integrity - AMBI provides integrity, and no one contended it wouldn't work. Probably belongs in msec or whatever forum winds up taking this work on. - Sec AD said "it's not using any novel crypto, and doesn't do anything tricky re: crypto agility, etc., so it probably doesn't present anything especially challenging from a security perspective", but Jake feels there's enough complexity there to warrant additional security attention. - Lenny: So AMBI would do it, if it gets more eyeballs? - Yes, though the AMBI equivalent for QUIC datagrams will likely be slightly different. BIER and AMT Deployment in RARE/GEANT Network Csaba Mate/Frederic Loui, 20 min - slides: https://datatracker.ietf.org/meeting/112/materials/slides-112-mboned-bier-amt-depolyment-in-geantrare-network-00.pdf - Jake: nice work, is there plans to run the DNS-based AMT discovery? - maybe, we need to make it easier so senders and receivers don't know the amt parts - jake: yes exacty, please look at RFC 8777, will talk more on the list. - Tim: next important thing is to get content online, looking for good ways to pull in content and make it available. right now traffic is low, just eumetstat, but vision is to make it possible for more people to stream whatever. - Lenny: How difficult to implement AMT? Spec clear? - Csaba: Wasn't difficult - Frederic: Want to make sure to invite people, this is an open lab, contact us and we can make sure you can make use of it and help you get going Offnet Sourcing Demo Lenny Giuliano, 10 min - slides: https://datatracker.ietf.org/meeting/112/materials/slides-112-mboned-off-net-sourcing-of-multicast-streams-00.pdf - demo results: choppy for some users, but people got it running - jake: might want to run this on a driad discoverable relay, will see if i can fit it in. lmk if anyone wants to see it. - Frederic: we'll look at how to run this in the p4 lab, this looks like maybe a useful source of content draft-szcl-mboned-redundant-ingress-failover Yisong Liu, 10 min - slides: https://datatracker.ietf.org/meeting/112/materials/slides-112-mboned-draft-szcl-mboned-redundant-ingress-failover-00.pdf - jake: would like to see more specifics on failover, how the need for failover is detected and signaled, at least recommendations on existing approaches. very useful space to work in, thanks for working on it.