Re: [pkix] fyi: Sovereign Keys: an EFF proposal for more secure TLS authentication

[Ben Laurie <[e-mail]>]

On Fri, Dec 9, 2011 at 5:32 PM, Miller, Timothy J. <[e-mail]> wrote:
> On 12/9/11 9:30 AM, "Ben Laurie" <[e-mail]> wrote:
>
>>Then you need to read the full paper, because it is _not_ equivalent
>>to a second CA.
>
> I have, and I said *functionally* equivalent, not *operationally*
> equivalent.
>
> Look at it this way:  The CA tells a user "I say this name is bound to
> this key."  Your notary basically says "Yeah, what he said."  The words
> are different but the message is the same.
>
> Note that the only way you can support detection of a colluding notary is
> to use what you've defined as a side channel.  The client needs to check
> with the notary's public list to show that a server-provided proof set
> isn't valid--a fact you recognize in the paper but not on the summary.

The summary says "clients check these proofs".

> This is part of the "reconsider your arguments" comment.
>
> Also, you missed a fraud case in your discussion.  What happens if a CA
> fraudulently issues a cert *and* registers it with a non-colluding notary?

The legitimate owner notices and takes action.

> The notary has no interaction with the entity the cert identifies, and
> can only take the CA's word for it. So the cert goes on the viewable list,
> the proof set goes to the fraudulent user, and the user gets spoofed.

Correct. However, this fact is detected.

>
> Finally, IMHO you place undue faith in the ability to "untrust" a notary;
> if the notary list comes from the browser makers--who by your own argument
> have a *proven* inability to manage roots of trust--then why should the
> notary list supplied and updated by the same source be somehow more
> trustworthy?

Because their current inability is thrust upon them, not caused by
incompetence or unreliability.

> On a side note, these aren't personal comments.  You placed the work in
> the public sphere in what appears to be a solicitation for comments and
> criticism.  I'm not particularly invested other than just trying to help
> improve the work.

Indeed, we are happy to receive comments, thankyou for your interest.