It is much clearer and simpler.  I like it better.  Thanks.

Deb

On Wed, Aug 27, 2025 at 4:58 PM Owen Friel (ofriel) <[email protected]>
wrote:

> Hi Deb – what do you think of this PR? It completely refactors and greatly
> simplifies section 8.2:
>
>
>
> https://github.com/anima-wg/brski-cloud/pull/257/files
>
>
>
>
>
> *From:* Deb Cooley <[email protected]>
> *Sent:* Tuesday 26 August 2025 21:00
> *To:* Michael Richardson <[email protected]>
> *Cc:* The IESG <[email protected]>; [email protected];
> [email protected]; [email protected]; [email protected]
> *Subject:* Re: Deb Cooley's Discuss on draft-ietf-anima-brski-cloud-16:
> (with DISCUSS and COMMENT)
>
>
>
> w/ [DC] below.
>
>
>
> On Tue, Aug 26, 2025 at 3:27 PM Michael Richardson <[email protected]>
> wrote:
>
> Deb Cooley <[email protected]> wrote:
>     > 1.  Section 8.2: I'm still struggling with finding the point here.
> I'd
>     > like to be a bit more cohesive.  State the issues/recommendations up
>     > front (i.e tell a story).  It looks like you are claiming that the
> end
>     > customer 'has to' run a private PKI, the manufacturer 'could' run a
>     > private PKI for the whole thing, or the manufacturer 'could' use some
>     > subset of Web PKI Trust Anchors (which you call something else). I've
>     > tried to do the 'use the public WebPKI' part below.
>
> RFC8995 (BRSKI) is for end customers that want to enroll a device into
> their
> private PKI (in a zero-touch way).
> So I really don't understand what you are complaining about here.
> There are options.
>
>
>
> [DC]  Outline the options in a clear and logical fashion.  Preferably
> without jumping around too much.
>
>
> 8.2 is about the tussle between being inclusive and reducing attack
> surface.
>
> Manufacturers that want to install IDevID into the devices (which BRSKI
> requires), already MUST have a private PKI for that.  The conflicting
> lifetime
> requirements prevent using a "Web PKI Trust Anchor".
> (There is also a SAN contents struggle, but there are workarounds for that)
>
> Is "Web PKI Trust Anchor" (set) even a well-defined term?
> It's not a single set, as much as people would like to think it is.
>
>
>
> [DC] that is absolutely correct.  Google, Apple, Mozilla, Microsoft, and
> many more (Adobe, Oracle, etc) all run Trust Anchor lists.  So saying that
> one can just use it (like there is only one) is odd, at best.
>
>
> AFAIK, CABFORUM does not define it, web browser and OS vendors decide on a
> case-by-case basis (based upon criteria established by CABFORUM), if they
> will
> accept any specific anchor into their trusted set.
>
>
>
> [DC] Right, the owners of the trust anchor lists may listen to CA/B Forum,
> but they aren't beholden to them.  They add the roots that they 'trust'
> when the rules that they set are satisfied.  This is why you don't want to
> go down this rabbit hole.
>
>
>     > 2.  Section 8.2, last three paras: These are actually just three
>     > sentences.  Let's see if we can make a paragraph out of them.  Put
> the
>     > last sentence first, and second sentence second, and first sentence
>     > last.  Something like: "It is recommended for Manufacturers to work
>     > with their VARs to determine if there is a subset of publicly trusted
>     > (Web) PKI Trust Anchors that would satisfy all their VARs, and to
> ship
>     > only that subset.  Manufacturers need to include enough trust anchors
>     > in their devices (the Pledges) so that all expected Cloud Registrar's
>     > can be validated.  While there is no requirement that Cloud
> Registrar's
>     > certificates are part of the public (WebPKI) Trust Anchors, it is
>     > likely simpler and cheaper for most such systems to use these easily
>     > obtained certificates."
>
> okay, I'll see wehat we can do.
>
>     > Nits: 8.1, para 2: 'device.is sheltered'/'device is sheltered .
> 8.2,
>     > para 3: 'operate enough of a private PKI'/'operate a private PKI'
>
> okay.
>
> --
> ]               Never tell me the odds!                 | ipv6 mesh
> networks [
> ]   Michael Richardson, Sandelman Software Works        |    IoT
> architect   [
> ]     [email protected]  http://www.sandelman.ca/        |   ruby on
> rails    [
>
>
_______________________________________________
Anima mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to