On Feb 17, 2013, at 3:34 PM, Paul Hoffman <[email protected]> wrote:
> Because I am my own operator. Yes, you want to be my nanny; no, I didn't ask 
> you to be.

I don't want to be _your_ nanny.   But you are one of the other two dozen 
people in the world with some inkling of how DNSSEC works.   You aren't the 
person that I mean to protect when I say this UI MUST NOT happen.

> That would work for me too, but not "you're a host, so you cannot use 
> negative trust anchors", which is what Jason's proposed text said.

Hence the discussion.   But it is worth paying attention to the way that this 
solution breaks the assumptions of the DNSSEC security model.   In the case of 
a validating resolver at the ISP level, it doesn't, because if you're trusting 
a validating resolver without a secure path to it, you're basically just 
looking for protection against cache spoofing.

If you have a validating resolver on your host, this gives you much stronger 
security guarantees; if you then enable negative trust anchors from an ISP 
list, you are back to the external validator level of security, which is a 
potentially _substantial_ downgrade, depending on what kind of security your 
provider offers.

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

Reply via email to