search

mattrglobal / bls12381-jwk-draft

https://datatracker.ietf.org/doc/draft-denhartog-pairing-curves-jose-cose/
0 stars 0 forks source link

Add rationale for changing EC to OKP #4

Open kdenhartog opened 3 years ago

kdenhartog commented 3 years ago

Something to the effect of this:

"EC is the JWK form, but EC2 is the COSE form which requires an X and Y coordinate. The original usage of BLS12-381 used EC, but OKP form which is expected to have only a single X value seemed more compatible. Therefore this draft proposes changing from EC to OKP for the key type parameter consistency in CWKs and JWKs and to prevent further disalignment implied by the description of EC2 for COSE."

andrewwhitehead commented 3 years ago

Maybe a reference to https://datatracker.ietf.org/doc/html/rfc8037#section-2 where OKP seems to be first defined. I don't know what source you're referring to with EC2 and the COSE form.

Edit: I see EC2 is used in that referenced RFC 8152. EC in the common JWK usages also requires x and y coordinates, so OKP does seem to make more sense there.

Maybe for clarity it would be good to refer to 'the SEC1 format with point compression' rather than 'the SEC1 point compression format' since it supports both?

kdenhartog commented 3 years ago

Good suggestions. I think all of those are worth adding to the draft. I'm currently going to leave this for a bit while I focus on other stuff. In the mean time this issue is a good place to drop notes on the draft.

andrewwhitehead commented 3 years ago

Okay. I'm using JWK as the native format for keys in Askar at the moment so it would be nice to formalize the representation for BLS keys, especially since they're searchable by thumbprint.