Colm,

At 2016-04-11 10:59:02 -0700
Colm MacCárthaigh <[email protected]> wrote:

> On Mon, Apr 11, 2016 at 10:54 AM, Robert Edmonds <[email protected]> wrote:
> 
> > Unbound actually does support both fixed and randomized, and the entropy
> > is taken from an interesting place: the ID field from the query.
> 
> I think if there are more than 8 RRs, that won't be enough entropy to
> generate all possible permutations.

At first I thought this was shocking. SHOCKING.

But if the goal is merely to mix up the answers a bit so that software
that uses the first RR does not always get the same answer, then it is
Good Enough(tm). Even with a fully-predictable QID this is fine.

It's only when you consider the order of the RR private that this
becomes a concern.

I admit I don't fully understand how information leakage is supposed to
happen. But if it is a valid concern then probably a cryptographically
secure random number generation is needed, and needs to be documented.

Cheers,

--
Shane

Attachment: pgpP93rHJOgkH.pgp
Description: OpenPGP digital signature

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

Reply via email to