Re: [Secdispatch] [saag] The Mathematical Mesh

[Nico Williams <[e-mail]>]

On Mon, Apr 22, 2019 at 03:33:22PM -0400, Phillip Hallam-Baker wrote:
> On Mon, Apr 22, 2019 at 3:03 PM Nico Williams <[e-mail]> wrote:
> > Is it fair to characterize this as something of a PGP trust mesh on
> > steroids?  If not, how does it differ?
> >
> > (Yes, I do see that your scheme has decentralized device management for
> > individuals, which is a bit of a new thing and very welcome.)
> 
> The objective is to give users the power to control their own security
> environment.
> 
> This includes but is not limited to the ability to make use of Web-of-Trust
> approaches or delegated trust, delegated distrust or a combination of all
> three which is actually the approach I think works best.

OK, so, all-of-the-above trust?

But there's nothing new yet, besides hierarchical[0], and
web-of-trust[1], TOFU, and things in between[2] right?

Just a new way of doing all-of-the-above?

This is not a critique.  I think a simple characterization will help.

"All of the above on steroids" seems fair.

But I do feel that the device management part is separable, even though
it's certainly critical for the UX.

[0] E.g., PKI, DNSSEC, Kerberos.
[1] E.g., PGP.
[2] E.g., WebPKI, which isn't exactly hierarchical, and definitely isn't
    webby as to trust.

> The primary focus is enabling real users to manage public key pairs on
> their devices without being aware that they are doing it. Securely
> establishing a set of public key pairs on each device and providing a
> validation path to the user's personal axiom of trust is the main idea
> here. Because if we achieve that, we are 80% of the way to securing almost
> any communication pattern.

Agreed.

Vendors will (already do) gladly help you do this within their walled
garden, natch.  The result is... a bit lame.

> Trust topology is something I think we should be opportunistic on and use
> every available option. For example:

Ah, "opportunistic".  TOFU.  Sure.

> * We meet at an IETF and exchange our master profile fingerprint directly
> using the QR code exchange.
> * You meet Alice at another meeting and directly exchange credentials using
> the QR code exchange.
> * I add Alice to the GitHub repo for the project based on your
> recommendation.
> 
> * I add my personal banker to my account based on an Extended Validation
> certificate authenticating his bank and an assertion issued by the bank
> credentialing him as a current employee.

OK, so WebPKI, PKI (if we had it, which, well, we do, since DNSSEC is a
PKI).

> Right now, I am doing everything in JSON and the Mesh Assertion
> infrastructure. It is pretty obvious that some of this is SAML territory.
> But where should the boundary be and should we make use of the XML SAML
> binding or switch it to a JSON encoding? I don't know what the right
> decision is there or have time to think about it.

It's always tempting to start from scratch...  But it's already been
done via OAuth/JWT.  Don't feel bad for not using SAML/XML.  Just make
the right choice.

Nico
--