Minutes IETF122: mimi: Wed 02:30
minutes-122-mimi-202503190230-00
| Meeting Minutes | More Instant Messaging Interoperability (mimi) WG | |
|---|---|---|
| Date and time | 2025-03-19 02:30 | |
| Title | Minutes IETF122: mimi: Wed 02:30 | |
| State | Active | |
| Other versions | markdown | |
| Last updated | 2025-03-19 |
minutes-122-mimi-202503190230-00
IETF 122 - MIMI WG Meeting Notes
Date: March 19, 2025
Adoption Call for draft-interop-mimi-discovery-requirements
- Richard Barnes: Minor comments but supports adoption.
- Eric Rescorla: Reviewed the discovery requirements document and
supports adoption. Acknowledges it’s not perfect but considers it a
good start. - Alissa Cooper: Conclusion: The document is considered
adopted!
Result: Adopted!
Richard Barnes - Presentation on draft-ietf-mimi-arch
- Review on the topic.
Rohan Mahy - Presentation on MIMI Protocol
- Three PRs to review:
Download Proxy PR
- Raphael: Too many options are being provided.
- Raphael: Suggested introducing an authentication mechanism to
enforce quotas. - Marvin: Opposes recommending OHTTP.
- Alissa: Agrees that there are too many options; confused about
the recommendations. - Daniel: Shares Alissa’s concerns and suggests clearly defining
that the hub is "in the hook." - Rohan: Requests that reviews be submitted to the mailing list or
in the issue tracker.
Review Issues
- Call for review: Issues #26 and #25 to be closed.
- Issue #38 ("Knocks") – Deferred for later discussion.
- Raphael: Notes that there was some consensus from last time,
and these decisions need to be included.
- Raphael: Notes that there was some consensus from last time,
Issue #23: Arbitrary State
- The new MLS AppDataUpdate can support this.
Raphael Presenting: Pseudonymous Credentials
Working Items
- MMRs require a well-defined concept of a "Connection," which is
not yet fully specified in MIMI. - Synchronizing decryption keys via TreeWrap.
- KeyPackage for MMRs publication is not yet specified but should
be straightforward. - Pre-authorization remains an open question.
Discussion Points
- Rohan: For the external commit case, can external joiners obtain
real identities? - Raphael: External commits are generally possible, but for
knocks, they are not. - Rohan: The concern is whether it's acceptable for someone to
join via external commit without receiving all information. Wants to
see if there’s a strong argument against this. - Raphael: Agrees; it makes sense that someone can join a room
without immediately seeing all members. Proposes improving this by
finding a better way to share keying material.
Rohan: Continue on the Presentation
- Can this be ready for WG Last Call by Madrid?
- Alissa: More people need to implement this!
- Richard: Agrees.
- Rohan: Some code exists, but more implementation is needed.
- Raphael: Implementation of some parts of the protocol is
underway, but the draft itself has yet to be implemented. - Richard: Acknowledges progress but suggests involving
students or hackathon participants to assist with
implementation, similar to MLS efforts. - Timo: Asks about peering relations within the group.
- Rohan: Refers to the main architecture document for
details.
- Rohan: Refers to the main architecture document for
Rohan: MIMI Contents Presentation
- Changes since -05.
- Issues that need to be fixed—if there are no objections, they will
be closed (can be reopened later). - Richard: It's not acceptable to close issues without considering
time zones. - Daniel: Nested attachments are not interoperable.
- Timo: Multipart attachments seem excessive.
- Marvin: Some features can be removed, but we generally need this
functionality.
Next Steps:
- Immediately start a WG Last Call.
- Close the open issues.
- Note: Delivery reports need more review—should they be moved to
another draft?
Discussion:
- Alissa: Let's have an implementation first.
- Richard: Agreed.
- Timo (in chat): I'd like more time on MIMI-Content. When
implementing, I found additional small issues with Markdown.
Rohan: Updates on Room Policies
- Rohan: There is an implementation in progress, expected to be
open-sourced this summer. - Timo: Will implement room policies soon.
Rohan: MIMI Identity
- Do we need this draft? Currently more of a survey of the problem
space.
Discussion:
- Richard: The analysis is confusing. We do need identifiers and a
base credential type. The draft may need to be rebooted.
Brendan McMillion: Presentation on draft-mcmillion-mls-subgroups
Call for Adoption
Discussion:
- Rohan: If Alice has multiple identities, this model does not
reflect that well. -
Raphael:
- There are two more drafts in MLSWG.
- No changes are required in wire format, but additional
modifications are needed for MIMI. - Regardless of which draft we adopt, privacy concerns remain—we
do not know which clients emit updates or which devices are
active. - A policy could mandate updates.
- Ideally, we should use one of these drafts in MIMI, but
implementation remains uncertain. - We need a mechanism to securely roll out clients.
-
Mojtaba: The subgroup name is incorrect; it should be in
MLSWG. The identity issue is significant, and the server needs
to play a role. - Rohan: Love the idea. Regarding functionality, servers do not
store messages indefinitely. If I go on vacation with my phone on
but my laptop off, what happens to those messages? - Daniel: Thanks for the discussion. MLS should drive how we move
forward and decide which draft to adopt. - Raphael:
- We should discuss this in MLSWG, it will helps the MLS Group to
be more efficient. -
Tim:
- MIMI is concerned about the problem, but there is another
related issue in the MLS WG. This draft should be discussed in
MLSWG.
- MIMI is concerned about the problem, but there is another
-
Brendan:
- Already presented this to MLSWG, which referred it to
MIMI.
- Already presented this to MLSWG, which referred it to
-
Rohan:
- MIMI also needs to do additional work, but the necessary
primitives should be handled in MLS.
- MIMI also needs to do additional work, but the necessary
-
Raphael:
- Brendan is right—not changing the wire format is a good
approach, but subgroups require modifications in MIMI.
- Brendan is right—not changing the wire format is a good
-
Brendan:
- Disagree. Based on the slides, those changes are not necessary.
-
Raphael:
- Still not convinced.
-
Brendan:
- would like to have show hands
-
Tim:
- it was clear to go for MLSWG
-
Orie:
- we can check on the MIMI List, still not sure what's the right
path, thanks for the discussion, I'll be available for more
discussion
- we can check on the MIMI List, still not sure what's the right
-
Rohan:
- the changes of wire format are not the case, the case is on MLS
Result:
- Will talke to our AD and will be communicated to Brendan later