Re: [pkix] PKCS #8? Re: Updated EdDSA/Ed25519 PKIX document

[Simon Josefsson <[e-mail]>]

Anders Rundgren <[e-mail]> writes:

> Pardon me for being somewhat unclear, I meant the distinction of EC versus ED.
...
> Essentially the question is weather you extend EC or create a new type.

Ah, now I understand.

Well, the situation is that GnuTLS implements a new key type (see [1]
but it should have more text) and there is ALSO the possibility of using
the RFC 5915 ECPrivateKey together with the NamedCurve defined in [2]
but it is a bit underspecified (see below).  I'm not negative about
doing a PKCS#15 format in addition to this.

The problem with re-using RFC 5915 is that has a number of references in
it that doesn't make a lot of sense for EdDSA (e.g., it uses I2OSP from
RSA for encoding!).  So it is not really clear how you would actually
re-use the existing EC type for EdDSA since it is so tailored around how
ECDSA works.  I assume this was the reason Nikos didn't re-use it but
created a new simpler private key format for EdDSA.

To have more alignment, we could describe an ECPrivateKey structure for
EdDSA keys that would be ASN.1-compatible but have different semantics
that would work for EdDSA.  I'm not convinced that is worth the effort,
as it will create implementation complexity and confusion.

/Simon

[1] https://tools.ietf.org/html/draft-josefsson-pkix-eddsa-04#section-8.3
[2] https://tools.ietf.org/html/draft-josefsson-pkix-newcurves-01