Minutes IETF120: rats: Tue 20:00
minutes-120-rats-202407232000-00

Meeting Minutes Remote ATtestation ProcedureS (rats) WG
Date and time 2024-07-23 20:00
Title Minutes IETF120: rats: Tue 20:00
State Active
Other versions markdown
Last updated 2024-08-20

minutes-120-rats-202407232000-00

RATS Agenda, Jul 23rd - Tuesday Session II (1 of 1 sessions)

Room: Plaza C -
https://datatracker.ietf.org/meeting/120/floor-plan?room=plaza-c
Scribes: Hannes Tschofenig; A.J. Stein; Thomas Fossati

13:00 - 13:05 Agenda Bash & Logistics
     (5 min) Nancy Cam-Winget, Kathleen Moriarty, Ned Smith

No changes requested during agenda bash.

13:05 - 13:20 draft-ietf-rats-endorsements - WGLC
(15 min) Dave Thaler -
https://datatracker.ietf.org/doc/draft-ietf-rats-endorsements/

Dave walked us through his slides starting with the history of the
document.
Goal: getting ready for WGLC
Ned's review comments addressed (some in -02, some only in GitHub):

Next steps: submit -03 and start WGLC
If there is no feedback, I'll submit -03 today.
Chairs: how many people have read -02?
... underwhelming response.
Chairs: we should probably do multiple WGLCs for this one.

The endorsement draft should be flagged as Updates:9334; the RATS
architecture draft because it adds definitions that should be part of
core RATS Architecture.

13:20 - 13:35 draft-ietf-rats-corim
(15 min) Thomas Fossati -
https://datatracker.ietf.org/doc/draft-ietf-rats-corim/

Thomas presented recent updates to CoRIM draft.
Updated -05 submitted before 120 cutoff; 18 changes in PRs merged; 4 PRs
pending

Muhammad Sardar (MS) wanted to talk about side meetings. Why is the work
on the draft not done in the working group. What is the work not done
transparently.

Thomas: In the meetings we go through the open issues and work on the
PRs. We have added an extra hour to the meeting to discuss the
restructuring.

MS: I was specifically interested in the work in these meetings
regarding the application Horn clauses to the draft and this work was
not brought back to the group.

Nancy: It would help to post the Github link and to share information
about the status.

MS: What was discussed should be brought back to the mailing list. I
never seen anything being brought back to the group.

Henk: there is an IETF process that dictates the venue to discuss the
document, and it is specified in the I-D. Any meeting that results in
changes in the content of the draft, it inevitably leads to public
changes, comments, and approvals (on GitHub).

MS: I understand, the one example issue I highlighted is not discussed
in the GitHub.

Yogesh: We welcome Usama and others to join. I checked the meeting
information and we have already invited Usama to said meetings.

Deb: My suggestion, re-send emails for invitations to the meeting.

(Mike Ounsworth (MO) was very close to activating his mic, but seems to
have failed. WG participants were amused, so we owe him thanks for
that.)

MO made his point in text chat: This is not the first time that people
have asked for more transparence from the RATS design teams.
It has previously been compared to the "PKIX Attestation" design team
that runs under LAMPS where I try (but don't always succeed) at
publishing meeting minutes back to the WG list. For example:
https://mailarchive.ietf.org/arch/msg/spasm/bIrtTYYjBfFHqEdD0EWU8jTnm9Y/

13:35 - 13:45 draft-ietf-rats-reference-interaction-models
(10 min) Henk Birkholz -
https://datatracker.ietf.org/doc/draft-ietf-rats-reference-interaction-models/

Henk briefed the group on changes to the draft. The high-level theme of
the old draft was prescriptive about dedicated TPMs, industry and the
draft have changed to address that not being the case anymore.

Addressed Hannes's review (he approved the relevant PR so it looks
like he's OK with the changes).

Henk asked if we should ask for WGLC, AD review, other approaches to
check that the document is aligned with the rest of the RATS WG output?
This I-D is relatively old at this point.

Nancy: we can do early SEC review and then an early directorate review.

Carsten in chat agreed early directorate approach is good.

Giri, Monty, Yogesh, Kathleen, MCR, and John volunteered to review

Giri: RFC 9334 and attested TLS need to be bridged in a more consistent
way. Can we review this document with that in mind? I.e., in
relationship to attested TLS?

Henk: Taking on the PoV of an ATLS, the ATLS work should be able to map
to one model in the I-D (e.g., streaming remote attestation) and
highlight which optional items are used, which additional ones are
added, and so on. If that is not possible with the current document:
that would also be a good test for the applicability of the reference
model I-D. If it is not possible to map ATLS interaction model to the
reference interaction model, then the reference interaction model needs
fixing

*13:45 - 14:00 draft-ietf-rats-uccs - WGLC**
(15 min) Carsten Bormann -
https://datatracker.ietf.org/doc/draft-ietf-rats-uccs/

Carsten recaps what UCCS is and why it is important to wrap the trivial
idea (a CBOR tag) in a set of security considerations that describe when
and why using it - esp. in context of RATS interactions.

The "grand unification" between UCS and JCS is not in scope.

-10 received a lot of editorial feedback.

Carsten and others believe it is ready for WGLC.

Deb: this document has already gone through IESG review and Roman's
comments need to be addressed instead.

Nancy concurs.

Thomas: A tangential concern, we think as UCCS is released an EAT media
type draft can be released as well. EAT media types can be progressed at
the same pace.

Deb: I am holding media types until UCCS and something else she can
recall is done.

Thomas: UCCS is the one holding up.

Deb: another dependency draft is in the editor's queue I want that
corrected as well.

Henk: the authors believe that we have responded to Roman's comments
with PR #29 (July 4). Perhaps it is a datatracker bug that the action
holder is not assigned correctly?

Thomas: another clarifying question, it is the EAT draft itself that
went back to IESG then to the RFC Editor's queue.

Deb: the draft-ietf-rats-yang-tpm-charra had a comment on it that
affected the YANG module and made some other changes were not correct.

Henk: we reversed the one change that may resolve the issue. He is
relatively sure.

Deb: clarification, I am only holding UCCS because of media types (note:
scribe doesn't understand clarification re the above).

14:00 - 14:10 draft-birkholz-rats-epoch-markers – Adoption?
(10 min) Henk Birkholz -
https://datatracker.ietf.org/doc/draft-birkholz-rats-epoch-markers/

Henk describes purpose and terminology of epoch markers. Also, the
underlying data structures that can instantiate an epoch marker (e.g.,
RFC3161 TSA, CBOR time-tag, nonce, nonce list, etc.)

Request adoption from the RATS WG (after some ping-pong with the COSE
WG).
Poll for "who has read": 4 Yes, like 16 No / No Opinion.
Nancy: will do the adoption call on the list.

14:10 – 14:20 draft-moriarty-rats-posture-assessment – Adoption?
(10 min) Kathleen Moriarty -
https://datatracker.ietf.org/doc/draft-moriarty-rats-posture-assessment/

Kathleen presented on the history of the draft and recent addition of
authors and latest updates. Draft is to summarize an otherwise verbose
attestation process that is conducted locally and forward the report
remotely to, e.g., a GRC system, a cloud controller(?), etc.

MS: is the scope limited by device type (TPMs vs TEEs)?
Kathleen: The scope is not limited.
Usama: I suggest adding it to the document because I don't see TEEs
being mentioned.
Kathleen: Monty is planning to add this, in addition to mentioning some
non-standards-conformant TEEs.

Yaron Sheffer: If I am in a cloud env and I don't care about the
platform components, how does that affect the architecture?
Kathleen: it's completely flexible: once you have the report you can
decide whether to forward it or not depending on context. The only
important thing is that the report exchange format is standardised.
Justin Richer: To follow-up on Yaron's question: can you choose only the
platform things, or only the cloud things, etc?
Kathleen: that's dictated by the specific benchmark structure.

Daniel Smullen: Can you remark on "where from"?
Kathleen: from the container management system

Kathleen continues with her slides: The document needs to define an EAT
profile.
Ask for adoption.

Nancy: let's poll who has read the latest draft?
5 people confirmed (and will also ask for adoption on mailing list).
Nancy: "Do you support adoption into RATS?"
Yes: 6
No: 1
No Opinion: 10

Deb: Will we offer a chance for the 1 person who said no to come to the
mic and say why not adopt?

Justin: I didn't vote no, but I have some opinions. The concept of the
document is sound, the doc needs a lot of work. A lot of syntactical
work to be aligned with JWT and other stuff. (Example: claim names are
all upper case.) There are some nuggets of good ideas, but it has a long
ways to go, evidenced by questions Yaron and I brought forth.

Kathleen: Agreed, we wanted

Usama: before I vote, what plan of trusted execution environments are
you proposing.

Nancy: I will ask those kinds of questions because the implementation
call (on mailing list) will be used to guage adoption.

Nancy: Who will review? Who will implement? (Nancy will defer to asking
on mailing list.)

14:20 – 14:30 draft-onsworth-rats-pkix-evidence – Adoption?
(10 min) Hannes tschofenig, Jean-Pierre Fiset -
https://datatracker.ietf.org/doc/draft-ounsworth-rats-pkix-evidence/

Mike Ounsworth has already presented to the RATS WG the LAMPS DT working
on CSR attestation.
That document is now in WGLC and attention has shifted to a format to
describe attestation evidence for HSM attesters. Includes a bunch of EAT
claims translated into ASN.1 (wanted by the CA and HSM communities) plus
some other specific "key attestation"-related claims.
Slide 5 has several links to various resources.

Yaron: Why are we not starting with ACME?

Hannes: ACME first and foremost a conveyance protocol, but this I-D
defines an attestation evidence format. Besides, the community
sponsoring this work is not interested in ACME. There is other ACME
efforts for remote attestation, with OAUTH as well, they are not
immediately related and the communities are interested in different
things.

Henk: a document about key attestation? yes please.

Nancy: have you presented this to LAMPS?

Hannes: no. The LAMPS meeting is later this week.

Tim Hollebeek: RATS and LAMPS are working closely on this stuff. As a
co-chair, I can't speak for the group, but I am happy with the approach.

Nancy: adoption will be on the mailing list

14:30 - 14:45 draft-fft-rats-eat-measured-component
(15 min) Thomas Fossati -
https://datatracker.ietf.org/doc/draft-fft-rats-eat-measured-component

AJ question regarding the import syntax with import digest from corim as corim namespaced as corim.digest after import for the new CDDL
syntax. He will read and follow up with additional questions.

Monty: Does this cover the event log format and IMA?
If not, I volunteer to help.

Poll: who has read version -03 of the EAT measurements draft?

4 yes, 7 no, 1 no opinion

Nancy: will do a call for adoption on the list

14:45 - 14:55 Arm CCA Token
https://datatracker.ietf.org/doc/html/draft-ffm-rats-cca-token-00
(10 min) Giri Mandyam

Giri presented on CCA Token in behalf of other authors.

Usama: the word "token" is used ambiguously in this presentation and not
relating it to accepted terms for this WG which is "Evidence".
Thomas Fossati points out that the "T" in EAT is for "token".

Henk tries to summarize. Where does the verifier behavior go?

Giri: I think this is better covered in papers that Arm publishes.

Thomas: We are going to define the Corim parts in a separate document.

Yogesh: There is already a CCA endorsement draft. We also have a full
open source implementation in Veraison where endorsment and evidence
processing is handled.

Usama: Regarding the verifier behvaior, Sec. III. D. 2) of our
paper covers both split and integrated verifiers.
I like this work and I suggest other vendors do also publish their
documents.

14:55 - 15:00 Open Mic
(5 min)

Peter: We are proposing a new initiative this Thursday morning, 9:30 -
11:30 in the Recency E/F room, that builds on top of RATS to allow
routing based on properties of attested routers (security enhanced
connectivity).

Chairs Actions summary