On Oct 3, 2012, at 7:42 AM, Paul Wouters <[email protected]> wrote:

> On Wed, 3 Oct 2012, Paul Hoffman wrote:
> 
>> I fully agree with all of this, but it leaves the question: what about 
>> tunneling DNS in TLS-over-HTTP? The earlier statement about why this would 
>> not work (corporations getting MITM certificates from bad actors in the root 
>> pile) doesn't actually apply because the client will have a single TLS trust 
>> anchor, possibly even one not even in the root pile.
> 
> Why would the client even need a single trust anchor for this?

For non-validating stubs.

> Current unbound dns-over-tls completely ignores the TLS. It is only used
> to get out, not for any type of authentication of transport or data.

Right: a validating stub who is using HTTP-over-TLS only as tunneled DNS 
transport has no need to known the identity of the other party.

--Paul Hoffman
_______________________________________________
dns-operations mailing list
[email protected]
https://lists.dns-oarc.net/mailman/listinfo/dns-operations
dns-jobs mailing list
https://lists.dns-oarc.net/mailman/listinfo/dns-jobs

Reply via email to