mailing list archives
Re: DNS Cache Dan Kamikaze (Actual Exploit Discussion)
From: Paul Schmehl <pschmehl_lists () tx rr com>
Date: Tue, 15 Jul 2008 20:46:57 -0500
--On July 16, 2008 11:17:07 AM +1000 Mark Andrews <Mark_Andrews () isc org>
The real problem isn't signing or resigning zones, or even
successfully=20 completing the original configuration (although those
are not trivial for=20 the average person trying to setup their own
dns). It's the trust=20 anchors. Until the root is signed, trust
anchors are a PITA. And until=20 the root is signed, why should anyone
believe that DNSSEC will achieve=20 wide adoption?
Well there are a number of ccTLD's that are already signed.
RIPE sign their part of the reverse space. ORG is in the
process of getting signed. It's happening.
There are existing solutions to dealing with lack of support
in the infrastructure zones (includes the root). You let
someone you trust collect the trust anchors for you then
incorporate them on a regular basis.
We effectively do this everyday with https but for some
reason people are scared to do the same thing with dns
despite private parts of the keys never being available to
the entity doing the certification. With https the certifying
authority can spoof any site they certify.
Perhaps that's because a cert problem on a web server breaks a single
webserver. A cert problem with dns breaks an entire domain.
If it isn't already obvious,
my opinions are my own and not
those of my employer.
Full-Disclosure - We believe in it.
Hosted and sponsored by Secunia - http://secunia.com/
Re: DNS Cache Dan Kamikaze (Actual Exploit Discussion) Ureleet (Jul 15)