Re: [Rats] Quantum-safe attestation

["Smith, Ned" <[e-mail]>]

If the architecture uses 'PKI' or words that imply PKI, please point to those sections in the architecture document.
-Ned

On 8/24/20, 7:22 PM, "RATS on behalf of Panwei (William)" <[e-mail] on behalf of [e-mail]> wrote:

    > Laurence Lundblade <[e-mail]> wrote:
    > 
    > I don’t want the RATS architecture to pick one algorithm or go into detail
    > about algorithms.
    > 
    > I expect the RATS architecture to be fully flexible to use what ever
    > algorithms can do the job. In particular it must not assume PKI. Most other
    > crypto-using IETF standards do not assume PKI. They often work better with
    > PKI, but they allow for other algorithms. RATS needs to be the same.
    > 
    > The particular manifestation I can think of for RATS is that Endorsements
    > must support confidentiality. If endorsements do not support
    > confidentiality an assumption that only PKI-based attestation can be used
    > is made.

    The characteristics of Endorsement depend on what the Endorsement is. I don't feel current architecture assumes the Endorsements are PKI-based or limits the algorithms to only use PKI.

    Regards & Thanks!
    Wei Pan

    > 
    > I don’t know of any other manifestations that moving to quantum-safe
    > crypto would have on the architecture, but it seems worth thinking
    > through.
    > 
    > LL
    > 
    > 
    > 
    > > On Aug 24, 2020, at 2:16 PM, Michael Richardson
    > <[e-mail]> wrote:
    > >
    > >
    > > Laurence Lundblade <[e-mail]> wrote:
    > >> We probably want RATS architecture to be able to use quantum-safe
    > >> algorithms. On low cost and low speed devices that might mean HMAC
    > is
    > >> used since SHA-2 and such seem to be quantum-safe. On higher cost
    > and
    > >> higher speed devices there may be alternatives that look more like
    > >> PKI.
    > >
    > > Yes/no.
    > >
    > > Use of a keyed HMAC requires a symmetric key, which will be a hassle
    > > to provision the verifier, and effectively locks the device to a single (likely
    > > manufacturer provided) verifier.   It will also raise questions of
    > > non-repudiation.
    > > Better *might* be RFC8778 HSS/LMS Hash-Based Signature Algorithms.
    > > They are bigger, and have a limited number of uses.
    > >
    > > I would say that the architecture does not need to say anything about
    > > the algorithms.  These are functional requirements, not design
    > requirements.
    > >
    > > There will have to be firmware updates to the Attesting Environment
    > > once either we have quantum-safe asymmetric algorithms, or when we
    > > have a QM breach.  That can be done with SUIT signed by RFC8778.
    > > I don't think that our architecture need worry about that.
    > >
    > > --
    > > Michael Richardson <[e-mail]>, Sandelman Software
    > Works
    > > -= IPv6 IoT consulting =-
    > > _______________________________________________
    > > RATS mailing list
    > > [e-mail]
    > > https://www.ietf.org/mailman/listinfo/rats
    > 
    > _______________________________________________
    > RATS mailing list
    > [e-mail]
    > https://www.ietf.org/mailman/listinfo/rats
    _______________________________________________
    RATS mailing list
    [e-mail]
    https://www.ietf.org/mailman/listinfo/rats