Re: [secdir] review of draft-hollenbeck-rfc4933bis-02

[Sandra Murphy <[e-mail]>]

(Note to all.  I sent multiple secdir related messages yesterday 
afternoon, and they all were mangled or ultimately not delivered.  So this 
is a repeat that some og the recipients may already have seen. 
Apologies for the technology failure.)



On Tue, 7 Jul 2009, Scott Hollenbeck said:


>There was no follow-up, so I remain a little unsure of how to address
>the comment.  Similarly, I need a clarification to know if the text
>change suggested below is best made in 4933bis.

Yes, I am guilty of not following up.  I was unsure of my understanding
of channel binding issues, none of the other reviewers of the rfc493*
suite made any notice of the issue, and, as you note here, the
issue is really in the base spec, not 4933.  I just had no idea
what the right process would be, if any.

Subsequent discussion and review of material make me more confident
that there is, indeed, a channel binding issue here.

The question is what to do about it.  This is an established protocol.
Would a security considerations section in 4930bis that pointed out
that there's a MITM attack possible here, because of the lack of
channel binding, be sufficient?  Or would pointing to the sasl-gs2
work as a protection be mentioned?  suggested?  mandated?

I sure would love to see AD or subject matter experts weigh in here
on the technical and process aspects of this.

--Sandy