On Tue, Jun 06, 2023 at 01:10:47PM +0000, Sandowicz, Krzysztof wrote:
> 
> I am architect at Intel working on Network Interface Controller
> implementation. We defined extensible manifests on devices, which
> are generated by ROM using the RFC9053 CBOR based encodings
> (Signing & Hashing algorithms) based on your 
> https://datatracker.ietf.org/doc/rfc9053/.
> Future implementation on NIC devices will require algorithms that
> are not defined in RFC 9053 (see below)
> 
>               +=======+=======+============+==================+
>               | Name  | Value | Hash       | Description      |
>               +=======+=======+============+==================+
>               | ES256 |   -7  | SHA-256    | ECDSA w/ SHA-256 |     P-256, 
> SHA2-256
>               +-------+-------+------------+------------------+
>               | ES384 |  -35  | SHA-384    | ECDSA w/ SHA-384 |     P-384, 
> SHA2-384
>               +-------+-------+------------+------------------+
>               | ES512 |  -36  | SHA-512    | ECDSA w/ SHA-512 |     P-521, 
> SHA2-512
>               +-------+-------+------------+----------------------+
>               | ????? |  ???  |SHA2-512/256| ECDSA w/SHA2-512/256 | P-256, 
> SHA2-512/256
>               +-------+-------+------------+----------------------+
>               | ????? |  ???  | SHA3-256   | ECDSA w/ SHA3-256    | P-256, 
> SHA3-256
>               +-------+-------+------------+----------------------+
>               | ????? |  ???  | SHA3-384   | ECDSA w/ SHA3-384    | P-384, 
> SHA3-384
>               +-------+-------+------------+----------------------+
>               | ????? |  ???  | SHA3-512   | ECDSA w/ SHA3-512    | P-521, 
> SHA3-512
>               +-------+-------+------------+----------------------+
> 
>                      Table 1: ECDSA Algorithm Values
> 
> Can you please let me know how can we request for changes in your RFC
> to define newer algorithms, so we could follow the standard.

The process to do this is to write a specification and publish it
(Internet-Drafts have been accepted) and then requiest expert review.

However, there are some techical issues with what you seem to be
proposing:

- SHA3-512 is very rarely used because it is slow. Usually SHAKE256
  is used instead, it runs at near twice the speed still at high
  security level.

- ECDSA with SHA3/SHAKE is somewhat nasty to implement in software, and
  is a interoperability nightmare.




-Ilari

_______________________________________________
COSE mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/cose

Reply via email to