Re: [Lake] 1 week 2nd WGLC on requirements and scoping text

[Eric Rescorla <[e-mail]>]

On Tue, Jun 2, 2020 at 6:48 AM Göran Selander <[e-mail]>
wrote:

>
>
> *From: *Eric Rescorla <[e-mail]>
>
>
> The scoping text added was:
>
>    As illustrated above, the setting is much more diverse
>    in terms of credentials and trust anchors than that of
>    the unconstrained web.  In order to deliver a timely
>    result, there is a need to initially focus on what is
>    considered most important at the time of writing: RPK
>    (by reference and value) and certificate by reference.
>    Information about validity of a certificate may be
>    omitted from the AKE if available over unconstrained
>    links.  The case of transporting certificate validation
>    information over the AKE may be specified in the initial
>    phase if there is a lightweight solution that matches
>    existing standards and tools.
>
>    A subsequent extension beyond the initial focus may be
>    inevitable to maintain a homogenous deployment without
>    having to implement a mix of AKE protocols, for example,
>    to support the migration path described above.  The AKE
>    needs to make clear the scope of cases analysed in the
>    initial phase, and that a new analysis is required for
>    additional cases.
>
>
>
> Stephen
>
>
>
> It's not clear how to read this in the context of other parts of
>
> the document, for instance:
> https://tools.ietf.org/html/draft-ietf-lake-reqs-03#section-2.2
>
> which says:
>
>
>
>    In order to allow for these different schemes, the AKE must support
>
>    PSK- (shared between two nodes), RPK- and certificate-based
>
>    authentication.  These are also the schemes for which CoAP is
>
>    designed (see Section 9 of [RFC7252] <https://tools.ietf.org/html/rfc7252#section-9>).
>
> How is one supposed to interpret this text?
>
>
>
>
>
> [GS] Since OSCORE is an extension to CoAP it is expected to support the schemes for which CoAP is designed. But in the discussion following the virtual IETF 107 LAKE WG meeting we restricted the initial focus, leading to the addition of section 2.2.1 and the following text in the paragraph after the one you quoted:
>
>
>
> ”In order to provide a clear initial effort, Section 2.2.1
> <https://www.ietf.org/id/draft-ietf-lake-reqs-03.html#initial-focus>
> lists a set of credential types of immediate relevance; the mechanism for
> selecting credential scheme is presumed to enable future extensibility if
> needed.”
>
>
>
> The ability to extend beyond the initial focus is also repeated in the
> text from Section 2.2.1 which Stephen quoted in his mail:
>
>
>
> ”A subsequent extension beyond the initial focus may be inevitable to
> maintain a homogenous deployment without having to implement a mix of AKE
> protocols, for example, to support the migration path described above.”
>
>
>
> Yes, my point is that these three paragraphs together contradict each
> other. One says the AKE MUST support PSK and the other says it's not part
> of the initial focus. Does that mean that the AKE as published will or will
> not support PSK?
>
>
>
>
>
> [GS] The AKE which we will start work on will not include the PSK ECHDE
> variant.
>

I don't think this answers my question, which is not about where we start
work, but what is an acceptable output of this WG.

To restate it more crisply: would an AKE which was extensible but did not
include PSK be conformant with the requirements? It seems to me that the
document is unclear on this point.

-Ekr

The process for how to include that variant is as far as I know not
> defined, but it must be possible to extend to that variant at a later stage
> given that the appropriate analysis etc. is made.
>
>
>
>
>
> Göran
>
>
>
>
>
>
>
>
>
>
>