Re: [radext] WGLC for draft-ietf-radext-radiusv11-02

[Alan DeKok <[e-mail]>]

On Oct 26, 2023, at 1:07 PM, Peter Deacon <[e-mail]> wrote:
> I don't think this document should move forward unless security standards related claims are either withdrawn or objectively substantiated.
> 
> As I provided reference for earlier and has been conveyed by others use of unapproved algorithms for non-security reasons is NOT disallowed by FIPS-140.

  From the text you quoted:

> "In addition, the dependency on
>   MD5 makes it impossible to use RADIUS in a FIPS-140 compliant system,
>   as FIPS-140 forbids systems from relying on insecure cryptographic
>   methods for security."

  That text addresses your concerns, does it not?

  Perhaps it could be touched up slightly to say:

	... the reliance on MD5 for security ...

> Further unless there has been a credible review of these systems WRT relevant standards text implying compliance is inappropriate.

  The text doesn't suggest that removing MD5 means that the systems are FIPS compliant.  It suggests that removing MD5 makes it easier for them to be *capable* of FIPS compliance.

  I don't see any text in the document which says "systems implementing this standard are by definition FIPS compliant".

  As such, I don't see how the above comment is relevant to the text in the document.

> " Systems which implement this transport profile are therefore capable of being FIPS-140 compliant. "

  That statement can be word smithed a bit.

  The issue people run into is that it's hard to audit software for FIPS compliance.  If the software uses MD5, is that being used for security, or for something else?  Unless they read the source code to all of the programs, they can't really tell.

  As a result. the simplistic approach is to just ban MD5 completely.  If MD5 is not used at all, it can't be used for security.  And therefore it's easier to demonstrate compliance.

  So the intent of the above statements is that it's easier to show FIPS compliant if they don't have MD5.

  Perhaps instead:

	Systems which implement this transport profile are more easily verified to be FIPS-140 compliant.

> " A major benefit of this extensions is that a home server which
>   implements it can also choose to also be fully FIPS-140 compliant."

  Perhaps instead:

	A major benefit of this extensions is that a home server which implements it can also be more easily verified for FIPS-140 compliance.  That is, a home server can remove all uses of MD4 and MD5, which means that those algorithms are provably not used for security purposes.

  I think these minor changes should address all of your concerns.

  Alan DeKok.