RATS Agenda Wednesday July 26th - Session 1 (of 1)

Room: Continental 8-9

9:30 - 11:30 local time (PDT)

Time zone: UTC-7

9:30 - 9:35 Agenda Bash & Logistics

(5 min) WG Chairs - Nancy Cam-Winget, Kathleen Moriarty, Ned Smith

• Ned presents.
• Request for note take takers and Zulip watchers.
  • A.J. Stein and Mike Ounsworth volunteer for notes.
• Agenda bashing: no requests to amend agenda.

9:35 - 9:40 EAT Media Types (WGLC)

(5 min) Henk Birkholz

• All issues closed and prudent to bundle with EAT publication (note that EAT is alreday further along in the publication process), last call requested
• Chairs agree to last call after session

9:40 - 9:45 Unprotected CWT Claims Sets - UCCS (WGLC)

(5 min) Henk Birkholz

• Normative recommendation for a use case for this would be RATS (similar bundling recommendation to EAT dradt above given how far along it is)
• Ready for working group last call
• Ned agrees it is ready given feedback for last call

9:45 - 10:00 Concise RIM - CoRIM

(15 min) ~~Henk Birkholz~~ -- Yogesh Deshpande

• Henk has questions regarding slide 4 and earmarking for CBOR tag definitions with a range of potential future semantics, in this case CoRIM. This experimental EMR approach may need attention from the CBOR WG as it has value but has not been done before more rigorously?
• Dave Waltermire: in appendix there is a complete CDDL description that is copied from RFC 9393 (CoSWID spec), so this may be out of sync. It would be advised to not copy it verbatim as it will potentially go out of sync often.
  • Henk: yes we acknowledge this is a problem and we will move to CDDL 2.0 with modularization ID and ???? ID. We hope that authors will adopt this pattern of CDDL module includes.
• Dave Thaler: an editorial comment regarding how he likes slide 4, but has concerns re slide 5 title and how it reflects on the CoRIM appraisal process. Match slide to match doc.
  • Yogesh: concurs and agrees it should be aligned.
• Henk: wants to reiterate we do not remove inapplicable CoRIMs, but highlight them as such for auditing purposes.

10:00 - 10:10 Epoch Markers

(10 min) Henk Birkholz

• Henk: Request call for adoption by RATS WG
  • Ned: question regarding COSE and epoch markers and its relation and influence to signature formats, does that draft specify anything?
  • Henk: recommend but not require COSE but will potentially address this as a COSE profiling problem.
  • Ned: is there a clear demarcation line about what the COSE WG and the RATS WG would work on?
  • Henk: Not yet (my answer, not COSE WG's official answer)
  • Kathleen: ok to leave this slightly ambiguous so long as we have the appropriate cross-review when the time comes.
  • Dave Thaler: reviewed the document and 90% of this document is not specific to RATS, why is it applicable here?
  • Henk: the use case of freshness is important in the context of RATS and the expertise on freshness in regards to attestion are in this working group.
  • DaveT: Suggestion: this go to COSE flagged for RATS expert review.
  • Kathleen: Suggestion: get the expert opinions you need from this room on the 10% that pertains to RATS, then send it over to COSE.

10:10 - 10:15 Conceptual Message Wrapper - CMW

(5 min) Henk Birkholz

• Henk: request call for adoption by RATS WG
  • Ned: read from the room, who has read this?
  • Yogesh: read it and is using it heavily as part of attested TLS work.
  • Henk: TCG would like evidence from attested parties.
  • Roman: can we get commitment to review it from others in the room.
  • A.J. and Laurence volunteered

10:15 - 10:30 RATS Endorsements (Adoption Call)

(15 min) Dave Thaler

• Kathleen Moriarity: multiple endorsements slide and diagrams is identical to attested sets I worked on only tied to known standards (full disclosure: such as my employer Center for Internet Security on standards decided by consensus experts on such standards). So I have a very similar draft based on architecture from SP 800-191 and I would like to work with you on it.
• Henk: These boxes could be thought of as Class - Instance - Group. 31411 ??
• Andrew Draper: supportive of the slide and approach, good to have from my perspective in the CoRIM work
• Laurence: I want to talk about attestation keys and identity in relation to this draft. This is an architecture draft, so it will impact design of all sorts of future things. An important point is the key material to verify the evidence. It deserves a lengthy descussion in the draft. Don't call it "identity" because to most people "identity" does not mean "key material". I do not support adoption of the draft unless you frame up that attestation key material is different from identity.
  • Henk and Laurence (?) volunteer to contributing text.
• Henk: The one place that "identity" makes sense is "device identity" -- everything else can be hard to identify, so we should not call it "identity".
• Yogesh: Is it possible to collaborate with us on CoRIM rather than making a separate draft?
  • DaveT: this is architectural / Informational and so we should keep the separation between architecture guidance and a specific implementation (ex.: CoRIM). Think of this doc like an extension to the architecture doc.
  • Ned: the chairs agree.
• Henk: Confidential Computing Consoritum may have a SIG report with an approach to the SGX mapping question summarized
• Nancy: enough commonality with claims-set to review convergence with Kathleen Moriarity's other claim-set draft and Laurence's request regarding identity/key-material, the chairs would like an updated draft of the document.

10:30 - 10:45 Attested TLS

(15 min) Yogesh Deshpande

• Laurence: do you mean attesting both client or server, authenticating both parties, or both these actions?
  • Yogesh: both sides can be authenticated with attestations either instead of or in addition to certificates.
• Nancy: are you proposing that an certificate carry an attestation.
  • Yogesh: Yes.
• Ned: Does bilaterial attestation require bilateral authentication?
  • Yogesh: (Note taker missed the answer on this.)
• Russ Housley: confused by message flow slide. Is new cert format (in the TLS sense) in lieu of X.509 the finished message will contain a signature. How will you get authentication, I just see the attestation piece?
  • Yogesh: I will get back to you, I am not sure I understand the question.
• Mike Ounsworth: regarding message flow side are you putting evidence inside the certificate, or the certificate extension? So does the CA need to sign over the evidence?
  • Yogesh: it is inside the CMW extension, and the evidence is signed by the attestion key. By "certificate extension" we mean a TLS extension on the TLS Certificate message (and therefore is external to the cert itself), not an X.509v3 extension inside the certificate.
• Dave Thaler: people keep saying putting the evidence in there, this will only work for one of the models; doesn't that only work for BackgroundCheck evidence. Why not make it work also for Passport model?
  • Yogesh: I just showed this example in the slide, the intent is to work with both.
  • Ned: we should all requirements from TCG (DICE in particular) and other standards bodies
• Dave Thaler: are we also considering the use case where attested TLS will use a password?
  • Yogesh: yes
• Kathleen Moriarity: chair recommendation is to take feedback, consult interested parties, get feedback regarding those parties, and get review from the TLS WG to make sure it is sufficiently prepared for that.
  • Yogesh: we have shared this with the TLS WG last time (since last session, IETF 116?) and they are aware of this.
  • Kathleen: shore up on feedback from people in this session, and go back to TLS WG, as that will improve the TLS WG if you do the necessary pre-work we recommend here.
• Dave Thaler: what is the right-most vertical line of the mapping to RATS diagram
  • Yogesh: we didn't want to go in-depth on that subject, but there is one (of multiple) use cases for that dotted line you asked about.

10:45 - 11:30 Open Mic

(45 min www

• Dave Thaler: are we going to run the WG last calls in parallel or in series?
  • Nancy: in parallel, why?
  • Dave Thaler: we have only so many cycles and we will have difficulty getting feedback.
  • Nancy: I will give 4-6 weeks and they are short, and I think they will be fine.
  • Dave Thaler: OK, that's fine taking a page out of SUIT WG approach for getting more reviewers.
• Dave Thaler: TEEP WG wants more interoperability work and testing. Dave Waltermire and I discussed SUIT-TEEP shared meeting because of a lot of shared interest and overlap. There is no direct relationship between SUIT and RATS, but TEEP uses both SUIT and RATS. Is there any interest in extending collaboration in November?
  • Nancy: much appreciated.
  • David Waltermire: we can share info about this interrim event. My only concern is we have a lot of business to cover and we don't want to jam the agenda too full.
  • Nancy: Dave Thaler should post to RATS and TEEP mailing list to do so. We agree to put dates on the Doodle poll for SUIT and RATS working chairs, and integrate interest polled by Dave Thaler on ML. We can make coincide or following one another accordingly.

Chair Actions

• WGLC for draft-ietf-rats-media-type
• WGLC for draft-ietf-rats-uccs
• Epoch Markers also needs to be raised in COSE as its scope is there and RATs is a use case in need for them
• RATs Endorsements needs to be updated to address key management and identity, it should consider merging with the Claim Sets drafts that was previously presented by Kathleen Moriarty