Wiki

Clone wiki

fapi / FAPI_Meeting_Notes_2024-01-17_Atlantic

FAPI WG Agenda & Meeting Notes (2024-01-17)

The meeting was called to order at 14:04 UTC.

1.   Roll Call (Nat)

  • Attendees: Nat Sakimura, Joseph Heenan, Brian Campbell, Daniel Fett, Filip Skokan, George Fletcher, Kosuke Koiwai, Mark Andrus, Matt Belanger, Peter Stanley, Peter Wallach, Robert Gallager, Victor Lu, Kelley Burgin, Bjorn Hjelm, Dave Tonge, Dima Posnikov, Matt Belanger, Rifaat Shekh-Yusef
  • Regrets:

3.   Events (Mike L.)

3.1.   OIDF Hybrid Workshop in Tokyo

Tokyo, Japan on Thursday, January 18, 2024

https://openid.net/registration-oidf-workshop-tokyo-2024/

In-person registration is full, virtual attendance is still available.

4.   External Orgs & Liaisons (Joseph)

4.1.   OBIE

Ceasing FAPI 1.0 RW ID-2 Certification at the end of the next year.

We probably can close https://bitbucket.org/openid/fapi/issues/570/deprecation-removal-of-fapi-1-implementers

4.2.   Columbia

Mandating FAPI 2 Security Profile. It came out last year.

5.   FAPI 2.0 Issues (Dave/Nat)

5.2.   PR #455 - Renumber attackers, fix editorial stuff

https://bitbucket.org/openid/fapi/pull-requests/455

May be external dependency on attacker numbering and attacker numbers are mentioned in security analysis but refers to one specific version of the spec

Changes were made to attacker model in response to analysis

Suggested to add note in attacker model stating that analysis was based on previous version of the attacker model

Added table to point out differences in the attacker model

Text starting on line 335 does not seem to be explicit enough.

Joseph is going to add comments

5.4.   Issue #670 - Use of FAPI with mandatory MTLS

https://bitbucket.org/openid/fapi/issues/670/use-of-fapi-with-mandatory-mtls

Various ecosystems are mandating use of MTLS everywhere

Conformance suite has ecosystem specific tests

Goal is to move away from developing ecosystem specific tests and standardize methods

If a new variant is created, it will work like the current way; vendors will test to make sure it works for each ecosystem

Currently, if non-plain FAPI profile is selected, MTLS is used everywhere and will pull in ecosystem specific tests

May not be addressed with just creating a new variant

Goal is to make it easier for new ecosystems and not make more divergence

If would be beneficial to have some note/guidance in implementation advice what to watch out for if you want to choose certain approaches/variants

Expand section on MTLS everywhere and make recommendations when MTLS everywhere is selected

Filip will create PR

5.5.   Issue #659- No normative statement on id_token encryption

7.   AOB (Nat)

The meeting adjourned at 14:__.

Updated