Re: [therightkey] draft-laurie-pki-sunlight-02

On 15 November 2012 02:09, Chris Richardson <chris@randomnonce.org> wrote:
> I have a few questions and comments on this document:
>
> A general comment: What should a log do if it receives multiple
> submissions of the same certificate?  It MUST detect and reject
> duplicates?  SHOULD detect?  What if it receives a certificate
> containing an embedded SCT from itself?  MUST/SHOULD/MAY reject?

MAY return the same SCT as last time - this is already fixed in the
next version (which I couldn't submit coz of cutoff dates).

> Section 1.1 fixes the hash algorithm as SHA-256.  It makes no mention
> of acceptable digital signature algorithms.
> http://www.certificate-transparency.org/sizes indicates the thinking
> is ECC.  Is RSA an acceptable signature algorithm?

Yes, but obviously expensive, since a certificate should usually
contain more than one SCT.

>
> Section 2.1: Shouldn't Version be covered by the signature in a
> SignedCertificateTimestamp?  I'd think it would be beneficial to be
> able to verify that the signature was intended for the same version as
> is claimed in the unsigned portion.

Yes, this is already fixed, also.

> Section 2.2 (minor edit): upon first read, the units of old_tree_size
> wasn't clear (leaf count?  bytes?)  The description of tree_size is
> explicit on the units ("number of entries").  I would appreciate it if
> old_tree_size had similar text.

OK.

> Section 2.3 (minor edit): the last bullet uses the term
> tree_signature, when the rest of the text uses tree_head_signature.

Oops.

I'm travelling this week, but will try to get an update out next week.

>
> Regards,
> Chris
> _______________________________________________
> therightkey mailing list
> therightkey@ietf.org
> https://www.ietf.org/mailman/listinfo/therightkey