MLS @ IETF 117

Additional credentials

• Raphael: Multi-cred useful for PCS, since you can indirect an
  HSM-stored key
• DKGillmor: Risk of infinite recursion?
• RLBarnes: We should forbid multi-in-multi
• Show of hands indicates consensus to adopt: unanimous support (24/0)

X25519 + Kyber

• RR: Need it on principle. Support adoption.

• RLB: Support this work. The only thing to note is that the symmetric
  bits are 128 not 256.

  • RR: Had same question. Bas told him he was wrong.
  • RLB: Should get some people to look at this.

• BHale: Support draft.

• Show of hands indicates consensus to adopt: unanimous support
  (>20/0)

Self-Remove

• RR: Always trade-offs, limited damage if this is wrong, support
  adoption in extensions draft

• JLHall: Seems like client UX could hide messages after leave?

  • RR: Client wants to verify state gets cleaned up

• RM to send a PR adding to extensions draft

MLS Extensions

• (No discussion, just an update)

Guardianship for MLS

• DKG: Does this reveal connection between guardian & edge?

  • XTian: Depends on the mode

• RLB: What are the protocol impacts?

  • XT: We didn't look at the specifics of the commit messages.
  • BH: for most of these nothing would need to change. The couple
    at the end depend on whether multi-sigs are allowed.

• ST: What's the next step?

• BH: We can write an I-D.
• XT: Eprint coming soon