[Cryptech Core] alpha board paper sketch

Peter Stuge peter at stuge.se
Tue Mar 17 14:00:35 UTC 2015


Would have liked to be there.

I like the idea of a button to exercise tamper detection.

So the tamper detector is the root of trust, since it owns the MKM.

Is AVR8 a Harvard architecture? I think that would be appropriate.

Why is CPU allowed to send *data* to tamper detector? I would expect
one or more discrete signals instead.

Also, should there be a discrete component for
monitoring/auditing/accountability, e.g. watching FPGA's MKM use?

And what expectations are there on the two USB-SPI interfaces from CPU?


//Peter



More information about the Core mailing list