Hi Mike,

Thanks for discussing and addressing my concerns. I will update my ballot
to No Objection once the new draft lands.

Paul


On Thu, May 8, 2025 at 8:18 AM Michael Jones <[email protected]>
wrote:

> Hi Paul.  Thanks for your diligent review.  My replies are inline below,
> prefixed by "Mike>".
>
> -----Original Message-----
> From: Paul Wouters via Datatracker <[email protected]>
> Sent: Thursday, May 8, 2025 3:56 AM
> To: The IESG <[email protected]>
> Cc: [email protected];
> [email protected]; [email protected]; [email protected]; [email protected]
> Subject: Paul Wouters' Discuss on
> draft-ietf-jose-fully-specified-algorithms-11: (with DISCUSS and COMMENT)
>
> Paul Wouters has entered the following ballot position for
> draft-ietf-jose-fully-specified-algorithms-11: Discuss
>
> When responding, please keep the subject line intact and reply to all
> email addresses included in the To and CC lines. (Feel free to cut this
> introductory paragraph, however.)
>
>
> Please refer to
> https://www.ietf.org/about/groups/iesg/statements/handling-ballot-positions/
> for more information about how to handle DISCUSS and COMMENT positions.
>
>
> The document, along with other ballot positions, can be found here:
>
> https://datatracker.ietf.org/doc/draft-ietf-jose-fully-specified-algorithms/
>
>
>
> ----------------------------------------------------------------------
> DISCUSS:
> ----------------------------------------------------------------------
>
> Thanks to Kathleen Moriarty for the SecDir review. Thanks to the authors
> for
> resolving her comments. I have a few hopefully minor DISCUSS items I would
> like to raise:
>
>
> I think the Abstract should say a little more on what it updates on the
> mentioned
> RFCs? eg the Recommended status of some algorithms specified?
>
> Mike> Agreed - will do.  Some of the updates were to deprecate registered
> algorithms.  Some of the updates were to the instructions to the designated
> experts.
>
> Section 2.1:
>
> What does "COSE Recommended" mean? Is this defined anywhere? Eg for TLS
> this
> is defined in the IANA registry. It needs to be clear whether it means
> mandatory
> to implement or not.
>
> Mike> This is defined in
> https://www.rfc-editor.org/rfc/rfc8152#section-16.4, with the definition:
>    Recommended:  Does the IETF have a consensus recommendation to use
>       the algorithm?  The legal values are 'Yes', 'No', and
>       'Deprecated'.
>
> Mike> Of course, this spec expands the set of legal values, as described
> in Section 4.4.  Notably, this definition was not carried forward into the
> specifications that obsoleted RFC 8152.  I think it makes sense to carry
> this definition forward into this specification, with the corresponding
> updates made.
>
> Section 2.2:
>
> See above, but in addition also "JOSE Implementation Requirements". Are
> "Recommended"
> and "Implementation Requirements" different things or the same thing?
>
> Mike> They are similar things.  The "JOSE Implementation Requirements"
> definition from https://www.rfc-editor.org/rfc/rfc7518.html#section-7.1
> is:
>
>    JOSE Implementation Requirements:
>       The algorithm implementation requirements for JWS and JWE, which
>       must be one the words Required, Recommended, Optional, Deprecated,
>       or Prohibited.  Optionally, the word can be followed by a "+" or
>       "-".  The use of "+" indicates that the requirement strength is
>       likely to be increased in a future version of the specification.
>       The use of "-" indicates that the requirement strength is likely
>       to be decreased in a future version of the specification.  Any
>       identifiers registered for non-authenticated encryption algorithms
>       or other algorithms that are otherwise unsuitable for direct use
>       as JWS or JWE algorithms must be registered as "Prohibited".
>
> Mike> This document doesn't update this definition, nor is RFC 7518
> obsoleted, so I don't feel the need to carry it forward here.
>
> Perhaps this document can ask IANA to add a Note: similar to TLS at
> the top of their registry explaining this? I am also puzzled that for
> example ed25519 is "Recommended Y" for COSE but "Optional" for JOSE? Is
> this difference by design?
>
> Mike> This carries forward the existing recommendation values in the JOSE
> and COSE registries, applying them as-is to the corresponding
> fully-specified algorithms.  The working group scoped this work only to
> making it possible to use fully-specified algorithms.  It did not take on
> re-evaluating any existing algorithm recommendations.
>
> Section 6.1:
>
> How does an implementation know the RSA key size of the peer ? I assume
> since
> this doc isn't fixing it, that implementations have either a common
> unspecified
> default (in which case why not mention that here) or a way to deduce this
> (in
> which case why not mention that here) ? Why not help implementations
> interoperate
> by clarifying what is done when processing these incomplete polymorphic
> identifiers?
>
> Mike> The implementation knows the RSA key size by inspecting the key.  It
> does not know, in general, if the other party is capable of using a key of
> that size.  In practice, for RSA software implementations, they tend to
> work with all key sizes.  Only for HSMs and highly controlled environments
> might some key sizes not work.  (Also, for security reasons,
> https://www.rfc-editor.org/rfc/rfc7518.html#section-3.3 already says "A
> key of size 2048 bits or larger MUST be used with these algorithms.".)
>
> Section 7:
>
>         A cryptographic key SHOULD be used with only a single algorithm
>         unless the use of the same key with different algorithms is
>         proven secure.
>
> I would use MUST instead of SHOULD here. The _only_ valid escape is the
> unless proven
> clause. No other unmentioned clauses exist, so MUST is more appropriate.
>
> Mike> Will do
>
> ----------------------------------------------------------------------
> COMMENT:
> ----------------------------------------------------------------------
>
> Abstract:
>
>         This specification creates fully-specified algorithm identifiers
>         for registered JSON Object Signing and Encryption (JOSE)
>         and CBOR Object Signing and Encryption (COSE) polymorphic
>         algorithm identifiers and deprecates those polymorphic algorithm
>         identifiers, enabling applications to use only fully-specified
>         algorithm identifiers.
>
> Please split up this runaway sentence. With the many "and"'s it is too hard
> to parse.
>
> Mike> OK.  (It was shorter before the acronyms JOSE and COSE were expanded
> in response to another IESG comment. :-) )
>
>          Allow only authenticated content encryption algorithms.
>
> Maybe add "(e.g. an AEAD)" to give people a hint what this means?
>
> Mike> Sure
>
> Thanks again for your useful review!
>
>                                 -- Mike
>
> P.S.  As I also write in my response to Mike Bishop, I plan to make these
> updates soon after the telechat - possibly later today.  I normally get
> these kinds of updates done before the telechat, but I was giving a keynote
> this morning at the European Identity Conference and am moderating sessions
> most of the day, so my bandwidth has been temporarily limited.
>
>
_______________________________________________
jose mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to