Hi Ben!
Thanks for taking the time to review the draft!

> For example, some insecure records may be present in cache that
> were fetched previously over UDP.  Should they be refetched over
> ADoX to prevent forgery, even though this reduces privacy?

Good question! I personally think a resolver should refetch the data
it has in the cache that does not live up to a client's request for STRICT
privacy. I'm not sure whether the draft should specify this, though.
What if some operators prefer to let the currently cached records
expire first before refetching them over an authenticated connection?
That should be allowed too (unless everyone thinks this is a stupid idea,
then we can update the draft to specify what the resolver should do).


> What is the privacy response status of a query whose recursive
> stages used a mix of cached and uncached, secure and insecure,
> ADoX and non-ADoX queries?

It depends on what the client requested. If the request was STRICT,
a mix of non-cached and cached queries that were fetched over
authenticated encryption is allowed and the response on success
will be AUTHENTICATED.
If the request was OPPORTUNISTIC-NO-FALLBACK, cached records
can be reused without considering how they were fetched and
encrypted connections do not have to be authenticated.
Upon success, the resolver will respond with CACHED if the records
were already in the cache or ENCRYPTED otherwise.


> What about QNAME-minimized queries that only reveal a TLD?

We tried to explicitly state in the draft that it does not care about
QNAME-minimization at all. I think your reasoning is sound: maybe
it would be okay to "leak" a TLD query and still be able to honor
a OPPORTUNISTIC-NO-FALLBACK request. But that opens up too
much complexity. Easier to just keep them separate and hopefully
we can have both at some point in the future.


> Should queries with EDNS Client Subnet be treated as more
> sensitive in some way?

Good question! Will have to do some homework on this one!


Hope this clears up a few things, at least!

//
Leon
_______________________________________________
DNSOP mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to