Re: Media Excitement!

[Cody Hatch <bytejump () gmail com>]

I completely agree with pageexec here. RBAC is good - no doubt - but
the difficulty of implementing a proper and secure policy seems almost
insurmountable. Foolproof doesn't exis because fools have proven far
too clever in the past.

It seems to me to be a better approach to use PaX, grsecurity, and
systrace to make the kernel and applications behave appropriately
rather than build a monolithic set of policies that govern the
permissions (roles) of users and processes.

Don't get me wrong - I see the benefits or RBAC - but I view
complexity as the law of averages working against you - the more
complex something gets, the more likely it is that mistakes will be
made.

Regards,
Cody

On 4/22/05, pageexec () freemail hu <pageexec () freemail hu> wrote:
> On 22 Apr 2005 at 10:09, robert () dyadsecurity com wrote:
> > The goal here wasn't to say "This one is more secure than that one".
> > It's to say "We have this level of sensitivity and require these
> > particular security mechanisms, and need this assurance level as to the
> > effectiveness of the security mechanisms".  Basically, choose the right
> > technology for your environment.
>
> i understood this much ;-), the real question is, which of the solutions
> in the mentioned URL gives *appropriate assuarance* against exploitation
> (remember the original question about alternatives to patching)? based
> on my experience and instinct, none of them does (EAL 4 is little more
> than a joke), but i'd like to be *proven* wrong.
>
> > > side question, which one of those didn't have security patches since
> > > their evaluation?
> >
> > I believe every product listed has had patches since their evaluation.
> > As I pointed out though in an earlier post, the containment of the
> > compromise, or rather the inherent ability to limit intrusion should
> > be designed into the TCB, not bolted on afterwards.
>
> does any of the mentioned products at that URL contain a compromise
> (thinking of kernel bugs)? or to be more precise, does any real-life
> policy (since any deployed MAC system implements one) exclude the
> compromise of the TCB? if not (which would match my experience), then
> what's the real point (of getting certified at EAL<7)?
>
> _______________________________________________
> Dailydave mailing list
> Dailydave () lists immunitysec com
> https://lists.immunitysec.com/mailman/listinfo/dailydave
_______________________________________________
Dailydave mailing list
Dailydave () lists immunitysec com
https://lists.immunitysec.com/mailman/listinfo/dailydave