[Cryptech Tech] AVR for tamper subsystem

[Fredrik Thulin]

On Wednesday, March 18, 2015 05:07:08 PM Peter Gutmann wrote:
> Fredrik Thulin <fredrik at thulin.net> writes:
> >Yesterday on the face to face meeting, we talked about the two suggestions
> >for the tamper subsystem - MSP430 and a PIC. We decided on an AVR ;).
> 
> The most important criterion for this is power consumption, since this is
> the part of the device that will potentially need to run for years off a
> single Li-ion cell.  So it depends on whether the ATtiny picoPower(tm)
> beats Microchip's XLP(tm) or TI's ULP(tm) (TI and Microchip have been
> slugging it out for awhile now over this, with claims and counterclaims
> from both camps).

Well the most important criterion would be security, but I take that as 
implied.

The arguments brought forward against my first suggestion of an MSP430 were 
ones of security actually. The MSP430 has a boot loader and the PIC mentioned 
on the list does not. While we didn't really investigate whether the boot 
loader in the MSP430 would be a real problem or not, we quickly found middle 
ground in AVR. Although I'm not familiar with AVRs, it is my understanding 
that they don't come with a ROM boot loader, so they are about as easy to work 
with as the MSP430 and arguably easier to trust.

I think the order of priority for the alpha is

  - security
  - ease of use
  - power consumption

while, for later designs, it will be

  - security
  - power consumption

/Fredrik