Re: [secdir] re-review of draft-ietf-smime-cms-rsa-kem

[Sean Turner <[e-mail]>]

Stephen Kent wrote:
> I reviewed version 05 of this I-D in July of 2008.  The current version 
> is 10.
> 
> My original reviewed cited only a two major concerns:
> 
>     - the previous version was ambiguous about support for Camella. This 
> version clarifies this issue, making support for Camellia a SHOULD.
> 
>     - the pervious version called for using an algorithm ID (with very 
> complex parameters) in a cert to signal when a message recipient 
> requires use of RSA-KEM. The authors addressed this concern in Section 
> 2.3 (and Appendix B), by stating that these parameters MUST be absent 
> when this OID is used in a cert in this context.
> 
> I have corresponded with Sean and he suggested that he could provide 
> more explicit words re the fact that the parameters MUST be omitted when 
> the algorithm OID appears in the SubjectPublicKey field of a cert. I 
> encourage Sean to include this additional text.

In the working -11 version I have not yet submitted, I've got the following:

OLD:

The parameters are absent.

NEW:

When the id-rsa-kem algorithm identifier appears in the 
SubjectPublicKeyInfo algorithm field, the encoding SHALL omit the 
parameters field from AlgorithmIdentifier. That is, the 
AlgorithmIdentifier SHALL be a SEQUENCE of one component, the object 
identifier id-rsa-kem.

spt