Re: [keyassure] publishing the public key

[Paul Wouters <[e-mail]>]

On Sat, 19 Feb 2011, Peter Gutmann wrote:

> It's not impossible, but close to it (and for ECC algos it may well be
> impossible).  This is a really, really bad way to identify a cert.

The whole point is we do not want to identify a cert. We want to identify a key.

> cert, for which you call memcmp(), thus the "all" in quotes above).  For ECC
> keys there are so many ways to store the same key (different parameter
> encodings, optional parameters, different ways of identifying the same
> parameters and in different formats) that in general there's no really
> reliable way to compare them.  Even with the simplest key type, RSA, it was
> hard enough, I've got code that mostly works most of the time and that's way
> more complicated than it has to be.

But such a method can be standardised in an RFC? And possibly be re-used for
other standards - like DNSKEY.

> If you want to uniquely identify a certificate, use the SHA-1 hash like
> everything else does.

Sure. That's one of the options in the TLSA draft right now.

Paul