Sorry for the long delay in responding to this--I needed to do some reading to 
figure out if I agreed with what you said.

The piece I was missing is that I didn't realize there was a process in place 
for requiring certificate transparency for any cert.   The plan to require 
certificate transparency in Chrome for Extended Validation certs sounds like a 
good start on making this happen, and does to some extent address my concern, 
although ultimately CT has to be required by the browser for _all_ certs before 
it really mitigates the key pinning attack I described.   But that's a path 
forward that I think is plausible.

So from my perspective, if the security considerations talks about the hostile 
pinning attack and suggests mandatory CT as a future way of addressing the 
problem, that would satisfy my DISCUSS.   I don't entirely agree with your 
observations about DNSSEC--I think I failed to effectively communicate the 
solution I was proposing, and so your response isn't actually addressing what I 
proposed.   But I don't care as long as the problem is addressed, and I do 
agree that once certificate transparency is mandatory, that will in fact be a 
better mitigation strategy than the DNSSEC-based strategy I suggested.

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

Reply via email to