Re: [Rats] Adoption call for draft-lundblade-rats-eat-media-type

[Dave Thaler <[e-mail]>]

Anders Rundgren <[e-mail]> writes:
> > And FYI to respond to Anders, the TEEP use definitely cannot use an in-object
> tag, and I expect some other uses outlined in the RATS architecture cannot
> either for the same reason.  Specifically, the TEEP use is to distinguish between
> (among other things) an EAT using the TEEP Profile, vs existing proprietary
> formats (such as SGX reports) that do not use CBOR.
> 
> Hi David,
> As we all know there are usually multiple ways to address a problem and this
> case is no exception :)
> 
> The established way dealing with different objects in an HTTP world is using
> different paths.  This has native support in most server frameworks while the
> support for media types is practically non-existent.
> 
> If for some reason SGX reports and EAT attestation must invoke the same
> application, wrapping SGX reports with with a tag ("wrapped EAT") seems like
> a reasonable alternative.
> 
> In short: EAT/TEEP/RATS would (IMO...) gain by sticking to de-facto standards.
> Media types are great but their primary use are in REEs.

The TEEP use of this draft is not in HTTP per se, but in TEEP messages.
HTTP carries TEEP messages.  TEEP messages carry attestation payloads.
Attestation payload formats are denoted with a media type carried in the TEEP message.

Separately, for use in HTTP, see RFC 9205's discussion of media types.  This was recently published by the HTTPbis working group.   It's possible (I don't know)
that it could be a useful Informative reference from the eat-media-type draft.

Dave