Hi Michael

> On 28 Jul 2020, at 17:22, Michael Richardson <[email protected]> wrote:
> 
> Signed PGP part
> 
> Hi Eliot,
> 
> 0) PLEASE ADOPT THIS DOCUMENT.
>   (Also I have some other MUD related documents which await the chairs' 
> attention)

+1, but let’s not expect a rush on approval.  Lots of work needed.

> 
> As I said in the meeting, I don't really understand sbom-local.
> I asked this online, but let me repeat.
> 
> 1) It seems that just putting coaps:///.well-known/sbom in the URI might work 
> as well.
>   If necessary, then maybe: coaps://127.0.0.1/.well-known/sbom or 
> https://[::1]/.well-known/sbom
>   CoAP does have a Content-Type for returned content, but it is true that
>   there isn't as much negotiation by default in the form of Accept:

I don’t think /// is a real construct.  There are two issues.  We don’t know 
the hostname and we don’t know the scheme.  But the MUD manager does know the 
hostname, and we need to be provided the scheme.  There is no convention for ME 
without a pre-existing context.  That might be something to work on, but it 
would hav to happen in coordination with the COAP and HTTP working groups, and 
I don’t know how big a job it would be.  Thus what’s there.

> 
> 2) I think that I'm interested in what the format of the sbom is than I am
>   about how to get it.

Well, you have your choice of at least three different classes and multiple 
encoding variants.  Nice thing about standards, eh?

> 
> 3) I think that, in general, that the SBOMs will seldom be hosted on a
>   constrained device that is capable of being described by MUD.
>   {that is, devices that can self-host an SBOM, will tend to be servers,
>   desktops,    etc. that are not easily described by MUD.

Perhaps.  I can’t say.  We’re going to need some operational experience.  There 
are a number of use cases in the IoT field where you plug in different 
components that require different software loads.  One aspect of SBOM, by the 
way, that hasn’t been explored is the difference between software being 
installed and the code being placed in service.  Particularly drivers and 
supporting user-level software.

> 
>   So I am not convinced that this is use case is reasonable.
> 
> Alternative paths:
> 
> a) There is also a draft,  draft-moran-suit-mud-00
>       Strong Assertions of IoT Network Access Requirements
> 
>   which would provide a MUD URL as part of a Software Update process.
>   It will be at secdispatch on Thursday.

How did this go?

> 
> 
> b) There is a draft:       draft-birkholz-rats-mud-00
>   which uses a MUD file to point to a list of appraisal evidence (ras-uris),
>   a list of RIM CoSWID (SBOM!), and edt-uri (endorsements of Roots of Trust)

Indeed.

> 
> These are *not* mutually exclusive, and we should have multiple paths for
> different ecosystems.
> 

Righto.

Eliot


Attachment: signature.asc
Description: Message signed with OpenPGP

_______________________________________________
OPSAWG mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/opsawg

Reply via email to