Hello,

I went through some discussions on null ciphers and didn't see this, my
apologies if this view has already been covered: There are cases where
you either can't (due to legal reasons) or don't want to (performance
reasons) encrypt, but would still like to have integrity protection and
be certain you are talking to the right party.

One example of a legal reason is amateur radio, which by definition in
most jurisdictions requires plaintext communications, but allows
integrity protection and authentication. Granted this is a tiny minority
of TLS users and use, but I don't really see much harm in keeping null
ciphers in play. Turning the knobs by default to positions that don't
allow null is fine, but the users should be able to turn those knobs
into other positions if they want to.

So the draft-ietf-uta-tls-bcp-08 section 4.1 first MUST NOT would in my
view be better as SHOULD NOT, with a rationale acknowledging those cases
where you don't want or can't have confidentiality. If you want to keep
the MUST NOT wording, at least add a few words to the rationale section
explaining the above exceptions. This way those software authors that
read the BCP might make an informed choice and let the users turn on
null ciphers themselves if they want to.

  Tapio

_______________________________________________
Uta mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/uta

Reply via email to