Re: "BIND 9 DNS Cache Poisoning" by Amit Klein (Trusteer)

[Theo de Raadt <deraadt () cvs openbsd org>]

> On 24 Jul 2007 17:40:35 -0000, securityfocus () networkontap com
> <securityfocus () networkontap com> wrote:
> > I don't exactly see how this is new "News" since Zalewski's paper on TCP =
> sequence number analysis (which included analysis of versions of BIND):
> > http://lcamtuf.coredump.cx/newtcp/
>
> That article does not deal with attacks on BIND's PRNG.
>
> As far as I can tell, Joe Stewart extended Zalewski's TCP sequence
> number analysis to BIND's transaction IDs - however I don't think
> Stewart's paper "DNS Cache Poisoning =96 The Next Generation" (
> www.lurhq.com/dnscache.pdf ) goes as far as the recent BIND advisory
> here - http://www.isc.org/sw/bind/bind-security.php:
>
> "The DNS query id generation is vulnerable to cryptographic analysis
> which provides a 1 in 8 chance of guessing the next query id for 50%
> of the query ids. This can be used to perform cache poisoning by an
> attacker."
>
> I don't think that Amit's attack has been described before.

The problem comes from ISC writing an incomplete solution to a problem
initially described in 1997 (and solved, I might add).

http://www.openbsd.org/advisories/res_random.txt

Before 1997, the attack was even easier -- take Amit's attack and
delete all the complicated math and replace it with id++.

Amit just shows that ISC ignored a better solution; that of using
a LCG-based generator.