Re: 2 forwarded messages...DNSEXT discussion of Day and Kaminsky

On Wed, Feb 11, 2009 at 04:09:54PM -0500, Dean Anderson wrote:

> These two explain the combined birthday attack pretty decently, I think.
> 
> I'm going to be looking at Jeff King's patches to see if they implement
> the combined birthday attack. The 'merge similar outgoing queries'
> sounds on the surface a lot like the plan to 'reuse queryids' that does
> suffer from the combined birthday attack.

OK, I see now why you think this would make things worse, but it
doesn't. The merge-outgoing patch does not reuse ports or reuse queryids
for a new packet (which would, as you say, make it easier for an
attacker to succeed).  They actually avoid sending another packet
entirely if we have a query that would have a coincident name (and
instead wait for the other packet to show up, and use the answer for
both).

So with stock dnscache, you can have MAXUDP queries oustanding to
authoritative servers that differ by port and queryid, but _not_ in the
query contents or destination IP. And that means instead of a 1 in port
* queryid chance, you have a MAXUDP in port * queryid chance of matching
* (i.e., 200 times as easy by default), since you only have to match one
* of them.

With the merge-outgoing patch, the idea is that dnscache will never have
two packets outstanding which have the same query contents and
destination IP.

But don't believe me; read the patches yourself and see if they make
that guarantee (and if they don't, please let the list know, because
that is a bug).

-Peff