I was thinking about Michael Dexter's BSD presentation last night.
One of my concerns about software is the insertion of logic bombs.

While open source code is decently reviewed by many eyes, the
review process is unlikely to catch intentional malware whose
behavior is scattered through many interacting  modules.  A
line of code here, a regexp there.  It would be difficult to
make all that add up to an exploit, but not impossible.

So, I assume that well funded agencies with enough brainpower
(US National Security Agency, Chinese People's Liberation
Army, Iranian Sepāh) can, and have, inserted logic bombs in
my Linux systems through seemingly innocent contributions to
open source software.  The insertion process would have to
be continous, and uncertain, as patches introduced by others
might deactivate parts of the behavior the logic bombs depend
on.  An arms race, where one side is trying to hide their
manipulations, and the other side is unknowingly defeating the
manipulations through the general process of code improvement.

I assume different code trees, like Linux and BSD, do not share
enough commonalities for the same subtle exploits to work on
both.  So if the two operating systems are running side by
side, processing the same inputs for the same intended outputs,
a third system could monitor the outputs of both and look for
differences.  This is a very high level abstraction; of course
the outputs and their sequence will differ, even if they follow
the same overall specification.  But if the specifications are
specific enough, the differences will be small and predictable,
and serious discrepancies detectable.  Both systems might have
some of the same overall exploits, but the time-to-exploit would
likely be different.  That should be enough to get attention
and trigger intervention.
  
On a less paranoid level, a "two OS plus detector" system
might be useful for testing code, or looking for failures in
systems needing ultra-high reliability.  Yes, the maintainers
of such systems will need big staffs to deal with a lot of
false alarms, but their code will become very well tested as
the sources of such alarms (bad specs and noncompliant code)
were eliminated.  

While I personally do not have the resources necessary to
maintain multiple OS'es (production Redhat and dabbling
with Ubuntu is all I can manage), those who can support a
heterogeneous collection of systems might consider setting
up some test systems like this. 

So, I'm glad some of us geeks are running BSD!  Keeping that
knowledge alive and ready to spread will be vitally important
in an emergency.

If we geeks ever find ourselves defending the region's
infrastucture from large scale attack, we may need to rapidly
deploy such systems to keep the generators from melting and
the gas pipelines from exploding.  From what I've read, the
US government and military are focused on cyber offense, and
the defense of their own systems, not protecting the general
population.  We are on our own - someday, the people on this
list may save Portland.

Keith

-- 
Keith Lofstrom          [email protected]         Voice (503)-520-1993
KLIC --- Keith Lofstrom Integrated Circuits --- "Your Ideas in Silicon"
Design Contracting in Bipolar and CMOS - Analog, Digital, and Scan ICs
_______________________________________________
PLUG mailing list
[email protected]
http://lists.pdxlinux.org/mailman/listinfo/plug

Reply via email to