Re: DNS cache poisoning attacks -- are they real?

[Florian Weimer <fw () deneb enyo de>]

* Brad Knowles:

> At 12:09 AM +0200 2005-03-28, Florian Weimer wrote:
>
> >  I doubt this will work on a large scale.
>
>       It's already been done on a large scale.
>
> >                                            At least recent BIND
> >  resolvers would discard replies from the abused caching resolvers
> >  because they lack the AA bit, so only clients using the resolvers as
> >  actual resolvers are affected.
>
>       Incorrect.

Indeed.

> The resolver requiring that the AA bit be set would prohibit anyone
> from forwarding queries to another server, which might be answering
> from cache.

Would you point me to such a configuration?  I don't think it will
work reliably for this purpose because BIND 9 only waives the
requirement for the AA bit if the authority section of the response
remotely looks like a referral.  I doubt that this is the case if you
simply redirect to a cache.