[Cryptech Tech] Alpha board main CPU change
Benedikt Stockebrand
bs at stepladder-it.com
Mon Jun 1 06:12:11 UTC 2015
Hi Jacob and list,
Jacob <jacob at edamaker.com> writes:
> Also, realizing that I don't have any of the credentials required to
> enter into the sanctum sanctorum of the crypto world, and I know that
> this is bordering on heresy, but since M4 has a TRNG, will you trust
> it? If yes, what benefit is provided by the noise board, if any?
> If not, why would you trust the M4 to manage your external TRNG and
> FPGA core I/F?
there's a fundamental difference between an intentionally
non-deterministic TRNG and the otherwise deterministic behaviour of an
MCU/CPU. You can (at least to some degree) test the behaviour of a CPU
doing black-box testing against the spec, but a TRNG is inherently
indistinguishable from a manipulated one if viewed as a black box.
To allow a TRNG to be audited in any way you need to provide the
annotated design and the associated layout. We won't get that for a
CPU/MCU, and even if we did it would be infeasible for the majority of
people to x-ray the chip and audit it. So the only option really is to
use discrete circuitry for the non-deterministic part of the job.
That said, using any ICs adversely affects the auditability of our
design. The FTDI stunt showed what weird things can go wrong there;
most importantly, the fake chips used some generic MCU to implement the
FDTI functionality, so there's a real-world risk that we're provided
manipulated chips which may run some altered firmware but still pass for
the real thing.
There's still more to all this, but if there's anything I've learned
from all this then it's that the entire electronics and IT industries
have happily ignored the issue of auditable designs for several decades.
Cheers,
Benedikt
--
Benedikt Stockebrand, Stepladder IT Training+Consulting
Dipl.-Inform. http://www.stepladder-it.com/
Business Grade IPv6 --- Consulting, Training, Projects
BIVBlog---Benedikt's IT Video Blog: http://www.stepladder-it.com/bivblog/
More information about the Tech
mailing list