Hi COSE WG,

As promised at IETF 126, please find my review of
draft-bormann-cose-turbo-kanga-kmac-00
<https://datatracker.ietf.org/doc/draft-bormann-cose-turbo-kanga-kmac/00/>
"COSE Algorithms for KangarooTwelve, TurboSHAKE and KMAC" below.


Section 2. MAC Algorithms Based on TurboSHAKE128, TurboSHAKE256, KT128, and
KT256 for COSE
<https://www.ietf.org/archive/id/draft-bormann-cose-turbo-kanga-kmac-00.html#name-mac-algorithms-based-on-tur>

The MACs are described only as "a MAC using [TurboSHAKE128/etc]", but the
precise construction is unclear. Is this HopMAC, or the "MAC on short
messages" MAC = KT128(M, Key, L) described in the last paragraph of section
4 of RFC 9861, or something else? It looks like the construction is MAC =
TurboSHAKE128(M' || K, D?, 16), but it's not at all clear to me. And if so,
why does this not use either of the two constructions recommended in RFC
9861?

Question for the group: Does the group want to have a D value?

No opinion from me.

Question for the group: L being 16 bytes (128 bits) is fine?

I reckon if 32 bytes is appropriate for HMAC-SHA-256, then 16 bytes should
be appropriate for TurboSHAKE128 (etc. for the other 3)? (same chance of
MAC collision as the chance of hash collision)


Section 2.4. KT256-MAC
<https://www.ietf.org/archive/id/draft-bormann-cose-turbo-kanga-kmac-00.html#name-kt256-mac>

The last paragraph names KT128 in both sections.


Section 3. KMAC128 and KMAC256 for COSE
<https://www.ietf.org/archive/id/draft-bormann-cose-turbo-kanga-kmac-00.html#name-kmac128-and-kmac256-for-cos>

Question for the group: Does the group want to have S to be an empty
> string? Or, is there a specific string the group would like to use?

No opinion from me.

Cheers,

Emil Lundberg

Staff Engineer | Yubico <http://www.yubico.com/>
_______________________________________________
COSE mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to