I agree with Filip. Ideally there would be test vectors for every algorithm registered, although that would be quite large given the size of the signatures involved.
Regarding the size, the smallest parameter set produces JOSE signatures of over 10KB in size, after base64url-encoding. So they are probably only going to have niche usage in the JOSE ecosystem, but it’s harmless to register them. Also, SLH-DSA public keys are quite small so there may be value in using them in JWKs independently, to bootstrap another protocol.
— Neil Furthermore, if you intend to keep JOSE in then please update the JOSE examples appendix section with actual working vectors. The existing "example" leaves a lot to be desired.
Since this draft registers JOSE algorithms and defines JWK representations it would be prudent to send its WGLC notice there as well. cc @JOSE WG
I appreciate the algorithm set is kept at a minimum. But I still don't see these as general purpose algorithms that we necessarily "need" to have in JOSE (unlike ML-DSA/FN-DSA). I'll bite tho and say that it doesn't hurt to have them registered as backup given the novelty and some small uncertainty surrounding the other PQC algs in general.
That being said I would welcome it if the draft did mention something along those lines, these algorithms are either targeting a niche purpose or serve as backup, the former is more likely. General purpose JOSE libraries shouldn't bother implementing these. I for one certainly won't, being mindful of the library footprint. Also none of the Web Cryptography API implementers currently plan to support them despite being included in the API's Modern Algorithms extension.
Speaking of which, the Web Cryptography extension will register all remaining SLH-DSA parameter sets in JOSE IANA for JWK representation purposes only (Algorithm Usage Location(s): "JWK"). It currently lists the ones from this draft too but that's merely because at some point it was uncertain whether this is going to move forward or not. I will update the extension proposal accordingly depending on what gets published in this draft.
|
