Re: [dns-privacy] Eric Rescorla's Discuss on draft-ietf-dprive-dtls-and-tls-profiles-09: (with DISCUSS and COMMENT)

[Eric Rescorla <[e-mail]>]

On Mon, May 8, 2017 at 1:21 AM, Sara Dickinson <[e-mail]>
wrote:----------------------------------------------------------------------
I've reviewed the diff:

1. The bit about meta queries still seems kind of muddled. If you do an
unauthenticated
meta query to get the IP and then use that to do strict queries for the
actual data, you suggest
that that offers a DoS attack (if the attacker gives a bogus response) but
couldn't they
just inject an unverifiable response if you did it in strict mode? Also,
doesn't it
introduce a possibility of amplification: the attacker sends the client the
IP of
some victim and now the client sends multiple packets to that victim (as it
retries)

2.  I don't think the "limited or no mitigation" text in S 1 is right for
opportunistic.
Rather, it provides defense against passive attacks but not active ones.
Note
that an HSTS-like mechanism would provide more here....


     widespread adoption of Strict Privacy.  It should be employed
> when
>      the DNS client might otherwise settle for cleartext; it provides
>      the maximum protection available.
>
> I don't think this statement is accurate. It provides the best
> protection
> that the attacker will allow.
>
>
> Agreed - will update.
>

I don't think the parenthetical quite does the job here. The point is that
it provides protection against passive attackers but not active ones.



> Table 1 seems to have N and D paired, so maybe you can coalesce them?
>
>
> That specific question was asked on the WG mailing list and the answer was
> ‘no, please keep both’:
> https://www.ietf.org/mail-archive/web/dns-privacy/current/msg01541.html
>

OK.

-Ekr