Daniel Kahn Gillmor <[email protected]> writes:

> I have *not* done any analysis of the larger, less-corner-y cases to
> see whether there's a strong argument for or against treating data
> that came in under confidential cover differently once it's in the
> cache.

Technically, it's near impossible to completely protect privacy unless
you don't use a cache.  Imagine the case where someone goes to a coffee
shop first thing every morning that supports a TLS based resolver.
A second "customer" 5 minutes later can then perform queries to the
resolver (regardless of TLS or not) for a slew of names to find which
were "in cache" and responding quickly.  You now know that person #1
went to sites A, X and Y since they returned in < 5ms, but not the rest
of the alphabet which returned in > 5ms.

However, I don't think this changes the nature of whether or not the
caches should be separate.  If anything, it may argue for a shared cache
so that normal traffic from non-privacy protected lookups will mean
someone snooping caches for private-protected lookups won't know it came
from a TLS-based user.

[And, no, we shouldn't go down the road of "privacy requires you disable
the cache"]

-- 
Wes Hardaker                                     
My Pictures:       http://capturedonearth.com/
My Thoughts:       http://blog.capturedonearth.com/

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

Reply via email to