OK I see now.  The language used in SC isn't at all what I understand as a
security person.  I'd prefer that more direct language is used, but that
ship has sailed long ago.

Encryption w/ a KEK and/or public/private key pair is fine, and indeed what
I would expect a client to use.  (definitely not cryptographic hashes)

Deb

On Sun, Sep 21, 2025 at 6:53 PM Mahesh Jethanandani <[email protected]>
wrote:

> Hi Deb,
>
> On Sep 21, 2025, at 10:00 AM, Deb Cooley <[email protected]> wrote:
>
> inline w/ [DC]
>
> On Mon, Sep 8, 2025 at 8:26 PM Mahesh Jethanandani <
> [email protected]> wrote:
>
>> Hi Deb,
>>
>> On Aug 7, 2025, at 3:14 PM, Deb Cooley via Datatracker <[email protected]>
>> wrote:
>>
>> Deb Cooley has entered the following ballot position for
>> draft-ietf-opsawg-secure-tacacs-yang-13: No Objection
>>
>> 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-opsawg-secure-tacacs-yang/
>>
>>
>>
>> ----------------------------------------------------------------------
>> COMMENT:
>> ----------------------------------------------------------------------
>>
>> I've edited my ballot to be 'no objection'.
>>
>> I'm gathering that it not the Yang data model's responsibility to protect
>> (particularly) sensitive values in storage, but only in transport.  Given
>> that
>> raw private keys, epsk, and shared secrets for TLS connections are
>> apparently
>> stored in Yang data models in plaintext form, that there should be
>> something in
>> the Security Considerations as a warning.
>>
>>
>> That is not true. It is the responsibility of the YANG module to specify
>> how a particularly sensitive attribute is protected in storage. Storing
>> private keys, or shared secrets are expected to be protected using
>> nacm:default-deny-all, where nacm refers to NACM [RFC8341]. In addition,
>> sensitive data is expected to be stored as a cryptographic hash. This
>> document refers to and uses the groupings defined in RFC 9642 that defines
>> in Section 4 how the encrypting of keys should happen in configuration.
>>
>
> [DC]  Isn't NACM just specifying access controls?  Is that actually how
> keys - shared secrets and private keys (which aren't mentioned at all) are
> stored?  One cannot store shared secrets and private keys as cryptographic
> hashes, because they have to be available in plaintext to be used (hashes
> are one way functions, once the value is hashed, you can't go back to
> the pre-hashed value).
>
> [DC]  I guess my questions are these:
> 1.  Why are private keys never mentioned when raw public keys and
> certificates are?  There obviously need to be private keys, or one can't
> perform the certificate validate part of TLS 1.3.  Seems odd.
>
>
> [mj] What might help is to look at the full tree diagram in Appendix C,
> which mentions the raw-private-key used by the client. Here is a snippet of
> the part of the tree that defines raw-private-key.
>
>        |     +--:(raw-public-key) {tlsc:client-ident-raw-public-key}?
>        |     |  +--rw raw-private-key
>        |     |     +--rw (inline-or-keystore)
>
> The YANG module is essentially importing the grouping from RFC 9645
> (Groupings for TLS Clients and Servers) which in turn imports the grouping
> from RFC 9642 (A YANG Data Model for a Keystore).
>
>
> 2.  Are shared secrets (and presumably private keys) merely stored in
> access controlled locations?  How does this prevent extraction and exposure
> of those values?
>
>
> [mj] It is the Keystore module that defines how private-keys are encrypted
> and stored. See Section 4 in RFC 9642.
>
> Let me know if you have any questions.
>
> Thanks
>
>
>
>>
>> Are you seeing something in this document that gives you another
>> impression?
>>
>> Thanks.
>>
>>
>> --------------------------------------
>> Thanks to Robert Sparks for their secdir review.
>>
>> Section 4, grouping tls13-epsk:  'Selfie-style reflection' attacks?
>> Reference?
>>
>>
>>
>>
>>
>> Mahesh Jethanandani
>> [email protected]
>>
>
>
> Mahesh Jethanandani
> [email protected]
>
>
>
>
>
>
>
_______________________________________________
OPSAWG mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to