Re: [lamps] [Pqc] Multiple drafts with PQ algorithm key encodings that are not compatible.

[Dieter Bratko <[e-mail]>]

Hello,

> I was considering changing the DilithiumPrivateKey ASN.1 to match 
> other private key formats 
> (https://datatracker.ietf.org/doc/draft-ietf-lamps-kyber-certificates/ 
> and https://www.rfc-editor.org/rfc/rfc5915) This means proposing the 
> new sequence structure:
>
> DilithiumPrivateKey ::= SEQUENCE {
>
> version                               Version,
>
> privateKeyAlgorithm      PrivateKeyAlgorithmIdentifier,
>
> privateKey                        OCTET STRING,
>
> publicKey          [0] IMPLICIT DilithiumPublicKey OPTIONAL
>
> }
>
> i.e., the parameters are defined/contained within privateKeyAlgorithm. 
> This means we can support the byte array encoding for the private key 
> proposed by the algorithm designers in privateKey an OCTET STRING.
>
To be clear (if the DilithiumPrivateKey shall be wrapped by an 
OneAsymmetricKey/PrivateKeyInfo): the PrivateKeyAlgorithmIdentifier of 
the privateKeyAlgorithm field identifies (by id) the Dilithium parameter 
set, and the PrivateKeyAlgorithmIdentifier of the same-name 
privateKeyAlgorithm field of the wrapping OneAsymmetricKey identifies 
the key as being a Dilithium key? Perhaps the privateKeyAlgorithm field 
of DilithiumPrivateKey should be given another name. Another possibility 
(as done in the draft of Christine) might be to omit the 
privateKeyAlgorithm field from DilithiumPrivateKey and use the 
parameter-set id for the privateKeyAlgorithm field of the 
OneAsymmetric/PrivateKeyInfo (thereby implicitly identifying the key as 
Dilithium key).

Regards,
Dieter