Hi Marco, hi COSE WG, First of all, thanks a lot for your time reviewing this document. The remarks are very useful. They will be very much taken into account, but due to the cutoff date, we will integrate them already after the IETF 127.
We hope this will help pave the way for the adoption of this draft. Best, Dmytro ----- Mail original ----- De: "Marco Tiloca" <[email protected]> À: "cose" <[email protected]> Envoyé: Mardi 3 Novembre 2026 20:13:42 Objet: [COSE] Review of draft-ochkas-cose-ascon-05 Hi all, Please find below some comments on this document. I support this work: Enabling the use of the Ascon family in COSE is certainly important and useful. Best, /Marco [Abstract] * I think that the abstract can be limited to the first paragraph, possibly with further information at the same high level. The content in the second paragraph would be a better fit in Section 1, e.g., for introducing the text in Section 1 that starts with "As a winner of ..." [Section 1] * s/only serializations/but only COSE serializations * s/existing cryptographic systems/cryptographic constructs [Section 3.1] * "justified security" I think you mean "justified security compromises" or similar. [Section 3.2.1] * I suppose that the fourth column is meant to have "Recommended" as its header, which would make the fifth column unnecessary. * It says: > Using this algorithm requires risk analysis or must be limited to > noncryptographic use cases (see Section 2 of [RFC9054]). Could you phrase it more precisely using normative language? Building on the phrasing in Section 3.1 about 64-bit tags, this can be: > In the absence of a careful and favorable risk analysis, using this algorithm > MUST be limited to noncryptographic use cases (see Section 2 of [RFC9054]). [Section 3.2.3] * Table 8: For consistency with the "COSE Header Algorithm Parameters" registry, I suggest to have the currently fourth column as the second one, while understandably keeping its header "New Algorithm" here. [Section 3.3.2] * Table 12: For consistency with the "COSE Header Algorithm Parameters" registry, I suggest to have the currently fourth column as the second one, while understandably keeping its header "New Algorithm" here. [Section 5.1] * It says: > The "Recommended" fields for Ascon-AEAD128/32 and HMAC Ascon-Hash256/64 are > set to "Filter Only" to discourage unreflected usage. - Based on the following subsections, shouldn't the list in the quoted paragraph also include Ascon-Hash256/64 for direct usage and Ascon-MAC/64? - "Filter Only" is set in Section 5.1.5 (Ascon-Hash256/64) and in Section 5.1.7 (HMAC Ascon-Hash256/64). The former relates to Ascon-Hash256 for direct usage defined in Section 3.2.1, which also includes text about risks for Ascon-Hash256 truncated to 64 bits (right after Table 4). The latter relates to Ascon-Hash256 for MAC operations defined in Section 3.2.2. Shouldn't Section 3.2.2 also include text about risks for Ascon-Hash256 truncated to 64 bits? - In Section 5.1.12, Ascon-MAC/64 has also "Recommended: Filter Only". However, it is not mentioned in the list in the quoted paragraph, and Section 3.3 does not provide requirements/statements analogous to those in Sections 3.1 and 3.2 about risk analysis for Ascon-AEAD128/32 and HMAC Ascon-Hash256/64. * Within Sections 5.1.1-5.1.15: * There has to be one more field "Change Controller" with value "IETF", between the fields "Capabilies" and "Reference". * The field "Reference" should also include a placeholder reference to this document. [References] * The informative reference [asconv1.2-caesar] is not used. [Nits] * Abstract --- s/of CAESAR/of the CAESAR --- s/selected Ascon family/selected the Ascon family * Section 1 --- s/(KDF) and Message/(KDF), and Message * Section 3.1 --- s/Ascon AEAD/The Ascon AEAD --- s/r, and the/r, the --- s/a and the/a, and the --- s/present, and it/present and it * Section 3.2 --- s/(KDF) and Message/(KDF), and Message --- s/Ascon hash algorithm/The Ascon hash algorithm --- s/the length of output/the output length * Section 3.2.1 --- s/version the Ascon/version of the Ascon * Section 3.2.2 --- s/In COSE, integration/In COSE, the integration of --- s/implemented via HMAC/implemented via the HMAC --- s/assigment/assignment (2 instances) * Section 3.2.3 --- s/In COSE, integration/In COSE, the integration of --- s/via HMAC-based/via the HMAC-based --- s/arbirtary/arbitrary --- s/assigment/assignment (3 instances) * Section 4 --- s/at List of Published Analysis/at the "List of Published Analysis" * Appendices A.1.1, A.1.2, and A.1.3 --- Please insert line breaks for readability within the values of Encrypt0_hex and Encrypt0_diag _______________________________________________ COSE mailing list -- [email protected] To unsubscribe send an email to [email protected] _______________________________________________ COSE mailing list -- [email protected] To unsubscribe send an email to [email protected]
