[pim] crosspost to pim: 6man draft: sub-link-scope-multica st

David 'equinox' Lamparter <[email protected]> Sun, 15 Mar 2026 06:51:40 +0100
Newsgroups gmane.ietf.pim,gmane.ietf.ipv6
Message-ID <[email protected]>
Hi all,


I'm not presenting it in any session this time around (constraints on my end), but still wanted to point out the existence of
https://datatracker.ietf.org/doc/draft-equinox-6man-sub-link-scope-multicast/
to the pim WG.

It essentially adds a new IPv6 multicast scope between host(1) and link(2), to expose limited / "specific" / "preexisting recipient definition" multicast capabilities from the lower layer (specifically, the 01:80:C2 groups in Ethernet).  In that regard it updates RFC 2464 to add a new IPv6 -> Ethernet group address mapping rule.  Right now, that's the only L2 tech exposed, there's room in the addressing for other link layers to be added/mapped by future documents.

An example says more than <n> words, so:
* ff22:/16 is the overall prefix (sadly, ff01½: isn't accepted by any IPv6 stack.  ff22 is a before-this-draft explicitly forbidden flag combination, P=1 T=0)
* ff22:0e00::/24 is the prefix where the new Ethernet specific mapping rule applies
* ff22:0e00:000e::/96 maps to 01:80:c2:00:00:0e ("never forwarded by a bridge" group) - the group ID/last 32 bits do *not* influence the Ethernet group address.

The purpose of the entire thing is allowing multicast discovery protocols on top of IPv6 to send to "only things on the other end of the Ethernet cable" or "only things this side of the next STP switch" or "only things on this wifi AP" [that last one is tricky].

Of course, being a scope between host and link, it cannot be routed and isn't really relevant for PIM; I do still need to figure out what's best re. MLD though. (It says "MLD applicability - to be decided" right now.)

The draft is currently in WG adoption call in 6man (technically it ended 2 days ago, I've grafted this mail onto the adoption thread);  I'm hoping to present it @ both pim & 6man in Vienna.  So you now have ample notice to find out if your favorite airline allows checking pitchforks ;)

(It's technically modifying a tiny piece of the IPv6 protocol itself, hence 6man.)

Cheers,


-equi
(David)

_______________________________________________
pim mailing list -- [email protected]
To unsubscribe send an email to [email protected]