Re: [Rats] draft-birkholz-rats-uccs

[Thomas Fossati <[e-mail]>]

Hi Carsten,

On Mon, Mar 15, 2021 at 10:49 AM Carsten Bormann <[e-mail]> wrote:
>
> On 2021-03-15, at 11:44, Thomas Fossati <[e-mail]> wrote:
> >
> > On Mon, Mar 15, 2021 at 10:26 AM Carsten Bormann <[e-mail]> wrote:
> >>
> >> On 2021-03-15, at 11:16, Thomas Fossati <[e-mail]> wrote:
> >>>
> >>> The oxymoron between "Unprotected" and the COSE in "CWT" has the
> >>> potential of creating cognitive dissonance.
> >>
> >> It is not the CWT that is unprotected; the CWT claims set is.
> >
> > I don't want to sound polemic, but if you take a CWT and you peel off
> > its security layer (i.e., you "unprotect" the CWT) what you are left
> > with is the CWT claims set :-)
>
> Exactly.
> But it is not a CWT, just as the people sitting in a tank are not a tank.

The analogy is correct but incomplete.  At least for the RATS context
(which I'm assuming is what we are interested here), you don't just
kick the people in the tank out and you're done with it: you need to
put them on a train and make absolutely sure that the train and the
tank came out of the same factory.  I.e., the requirement for UCCS is
that the signing key that identifies the secure transport endpoint is
- or can be linked to - the same key that would have signed the
attestation Evidence, were it encoded as "proper" CWT.

> >>> That IMHO deserves some discussion *in* the document.
> >>
> >> Definitely.  But piling up security considerations for when you are carrying around unprotected information seems a bit on the motherhood and apple pie side.
> >
> > I don't think I have suggested piling up stuff - IIRC I even used
> > "minimalist" upthread - but something needs to be said to clarify what
> > are the assumptions that need to hold for this to be usable.
> >
> > (I will review the draft and provide some more useful/actionable comments.)
>
> I definitely appreciate that!

Will do shortly, cheers!
--
Thomas