Room 7, RATS Session 2
Time zone: UTC, 1 hr
Note takers: Eric Voit, Giri Mandyam
RATS chairs
Eric Voit (EV), Laurence Lundblade (LL), Kathleen Moriarty (KM)
(1) draft-voit-rats-attestation-results - draft-AR
(2) draft-moriarty-attestationsets - draft-AttestationSets
(3) draft-ietf-rats-eat - draft-EAT
EV: Introduces slide on attestation results, and interrelationship between. 3 drafts
EV: The draft-EAT defines claims. draft-AttestationSets defines evidence sets, but can leverage claims for conveyance. No overlap.
DT: draft-EAT can be a generic format - not just for data coming from an attester.
KM: Main purpose is to define what a 'set' looks like. Not covered in draft-EAT.
EV: draft-AttestationSets is defining types and sets of Evidence, which doesn't have an overlap with generated attestation results from a Verifier.
EV: draft-EAT contains claims on identity, while draft-AR has claims on types of identity (non-normative). There are new claims in draft-AR could be taken as claims into draft-EAT. Subsequent revisions of draft-EAT can adopt encoding of identity types.
HB: draft-AR can re-use identity claims of draft-EAT. Additional identity claims can be created later, but don't need to be in the core draft-EAT.
EV: Not all identities will come from the same place. No normative conflicts in the existing 3 drafts.
Giri: There's a pull request outstanding on draft-EAT. Is there no conflict between that pull request and your draft (draft-AR)?
EV: currently, that is correct. draft-AR evaluated the overall trustworthy of the device iteself, so the claims are independent?
Giri: Can we pull that claim into draft-EAT?
EV: I don't have a problem with it, but it's an independent question. Personally, I think relying parties should be relying on high level conclusions from a verifier. YMMV.
GF: Need a place where these document relationships are recorded.
EV: we don't have a description of how these object relate. That is an open question that we haven't addressed as a group. It's a large conversation, making sure our info models have reasonable ties between them.
AW: I would like to see justifications captured on why things are separated the way they are.
LL: draft-AR and draft-EAT relationship is subtle. draft-EAT can be used as a substrate for implementation of draft-AR. draft-AR can also be done in SNMP or YANG. The request was made that if draft-AR is adopted that it not be called "attestation results" as other drafts can describe attestation results.
Henk Birkholz (HB)
HB: Need clarification on adoption call. Where are we right now?
NW: Are there comments? There was a call for adoption ending today. We should ask for adoption right now?
NS: Work is still needed before adoption (e.g. cover Intel's EPID)
HB: We should do this work in the WG.
Poll results:
Raise Hand: 15
Don't raise hand: 5
RD: How do we make sure this doesn't conflict with TCG?
HB: The authors are the same on both sides. And have the authors on both sides working the intersection is a good thing.
RD: The WG mailing list result was only one vote for, versus 15 on this session
DT: This is important for the WG to cover, it is important beyond the TCG intersection.
NW: Do we need to re-confirm?
RD: It is not needing a re-confirmation as both were accomplished via the process.
NW: If you have an objection, please speak up now.
GF: This would be a lot of work, is it important enough?
NW: This is why we are asking the question. It looks like there is consensus.
Poll: would you be willing to work on the DAA draft:
Raise Hand: 8
Don't raise hand: 8
RD: We need to work on sequencing across the level of effort needed across the WG.
NW: DAA is adopted.
Henk Birkholz (HB), Thomas Fossati (TF)
HB: Extensibility is something which needs to be emphasized.
HB: Would span integrity measurements for TCG (DICE and TPM), ARM PSA Token, EAT
HB: We are seeing running code in places like Veraison.
HB: Evidence reporting is consistent with RATS charter (assertions about system components, protection with COSE structures, consumed by Verifiers)
HB: Requesting WG adoption. Github repo is active. Thrice-weekly design meetings.
GF: What is the difference between CoRIM and CoSWID?
HB: CoRIM (Concise Reference Integrity Manifest) = umbrella manifest, CoSWID (Concise SoftWare ID) = software filesystem hierarchy
Editors note: CoMID (Concise Module ID) - is a schema within CoRIM that describes device hardware and firmware especially where the filesystem abstraction doesn't make sense.
RATS Chairs
NW: Need a status from the TEEP discussion
DT: Should be in mailbox > believes concepts are not TEEP specific. TEEP will define specific values.
(On whether we can go to WGLC for RATS).
LL: There were comments from Thomas & Jeremy on the list which doesn't match current draft. Needs more discussion here.
LL: Suggested the draft name is misleading as other drafts could contain attestation results content.
Poll: Is draft -v11 ready for WGLC?
Raise Hand: 10
Do not raise hand: 4
NW: Ready for WGLC.
(on RATS Architecture IPR)
NW: If anyone has any feedback on if the IPR asserted should impact the progression of the architecture draft, the last call for comments on the IPR submitted Friday of next week.
NW: Mailing List Adoption Calls will be made on:
(1) draft-voit-rats-attestation-results
(2) draft-moriarty-attestationsets
Editors note: Subsequently, the draft-AR authors propose changing the draft name to draft-ietf-rats-attestation-results-secure-interactions