[Rats] Re: Updated version of Distributed Remote Attestation

[Michael Richardson <[e-mail]>]

Hi, I had read draft-wang-rats-distributed-remote-attestation-00.
I have reviewed -02 quickly.

I was not convinced that the distributed component was either necessary or
sufficient.    Even proof-of-stake DLs are only temper-evident with **effort**

DLs are appropriate when there are mutually distrusting entities, who can
invest equally in verification.  I point to midnight settlement for banks as
a successful use of the technology.

In your Figure 1, Reference value providers and Endorsers contribute (write)
to the DL, while Verifiers most read from it.  They are not equal in any way.
What is the motivation for Endorser 1 to verify the DL that contains Endorser
2's contribution?   Imagine that they are business competitors.

I am also skeptical that writing Attestation Results is a useful way to
communicate them to the RP.

I think that I said before that this is DL looking for a problem.
This isn't a problem it solves well.

jiangyuning \(A\) <jiangyuning2=[e-mail]> wrote:
    > We’d like to (re-)introduce our individual draft,
    >  [1], and share that
    > it now has an open GitHub page [2] for collaboration.

    > Many recent discussions (multi-verifier, CoSERV, conceptual-message
    > subscription) touched on how verifiers obtain and share endorsements,
    > reference values, and appraisal results at scale. Our draft addresses a
    > closely related but complementary problem:

    > ⦁ Cross-domain attestation transparency: In many deployments, each
    > trust domain operates its own Remote Attestation Service (RAS). In
    > cross-domain scenarios, a relying party or verifier in domain-A often
    > cannot directly interface with devices/verifiers in domain-B, making it
    > difficult to obtain and validate attestation measurements and results
    > across domains.  ⦁ Many-to-many distribution of security artifacts:
    > Verifiers appraising heterogeneous attesters need to obtain endorser
    > public keys, endorsements, and reference values from multiple
    > providers. Conversely, these providers need to distribute artifacts to
    > multiple verifiers across domains. This many-to-many distribution
    > problem doesn't scale with point-to-point integrations, and similarly
    > applies when verifiers share attestation results for reuse.

---

    > In this draft, we propose a mechanism to make selected attestation
    > artifacts more reusable across multiple verifiers and domains, and a
    > shared publication channel for endorser PKs / endorsements, reference
    > values, and (optionally) attestation results, with provenance and
    > access control in mind.

I don't think sharing the artifacts this way makes them more
interoperable/re-useable.   The problem you seem to be solving is NAT44.

    > The draft currently discusses a distributed-ledger-based realization;
    > we are also open to reshaping this as a more generic solution if that
    > is a better fit for RATS WG.

    > We appreciate feedback on whether this work is (a) correctly scoped for
    > RATS and (b) useful to the WG, and we’re also looking for potential
    > collaborators/reviewers.

    > BR, All authors

    > [1]
    > https://datatracker.ietf.org/doc/draft-wang-rats-distributed-remote-attestation/
    > [2]
    > https://github.com/DistributedRemoteAttestaion/Distributed-remote-attestation

    > ----------------------------------------------------
    > Alternatives:

    > ----------------------------------------------------
    > _______________________________________________ RATS mailing list --
    > [e-mail] To unsubscribe send an email to [e-mail]