do we need to define key size?

[Sakurann]

just for the reference FAPI defines

• shall require and use a key of size 2048 bits or larger for RSA algorithms;
• shall require and use a key of size 160 bits or larger for elliptic curve algorithms;

[tlodderstedt]

@danielfett @paulbastian What do you think?

[paulbastian]

I would define a very limited number of crypto suites that shall be used, obviously ES256 but not so much else

[Sakurann]

related to #112

[nemqe]

Does it make sense to reference NIST recommendations (or any other body) when it comes to key sizes, which hash algorithms with which key size and etc for a subset of what is chosen (RSA / ES256 / Ed25519?)?

[jogu]

WG discussion: We should be careful not to contradict anything in ISO 18013-5 / 23220.

Nobody on call felt we needed normative text about key sizes.

Suggested that we add a security consideration that references SP 800-131A and/or SP 800-57 but doesn't make normative requirements.

[cre8]

The German Federal Office for Information Security has a Technical Guideline of recommended algorithm and key lengths. This is getting updated over the time, it may be linked as a useful resource, but not mentioned as a recommendation?