Re: [openpgp] SHA3 algorithm ids.

[Bill Frantz <[e-mail]>]

On 8/16/15 at 8:47 AM, [e-mail] (Derek Atkins) wrote:

>Bill Frantz <[e-mail]> writes:
>
>>I think in the IoT space, we will need to have signed software
>>updates. I don't think there is much of an issue taking several
>>seconds to verify an update signature, but these 8 bit processors seem
>>like the right level of hardware for these IoT devices.
>
>Yes, signed software is definitely one use case.  However, often on
>these systems it's more than just authenticating a software update;
>sometimes it might actually want to check the signature on every bootup
>(to prevent an attack on the flash/firmware)!

I hope we don't have to worry about attacks via physical access, 
so the only attacks available will be through the upgrade mechanism.

We also need to worry about authentication and replay prevention 
for the instructions delivered to these devices through the 
internet. One can imagine an architecture with a controller with 
the power of a Raspberry Pi giving orders to dumber devices 
using authenticated symmetric crypto as a solution. That system 
would prevent my favorite "neat hack" attack, turning your 
neighbor's living room into your own light organ.


>I'll note that there are alternate algorithms that run much faster than
>ECC (e.g. Algebraic Eraser can run in the tens of miliseconds instead of
>the ones of seconds of ECC)!  However my real point is that we should
>not ignore these platforms, and more specifically we should remember
>that they might not have the power to run the same algorithms that work
>fine on our x86-64 servers.

I think we are in violent agreement here.

Cheers - Bill

---------------------------------------------------------------------------
Bill Frantz        | If you want total security, go to prison. 
There you're
408-356-8506       | fed, clothed, given medical care and so on. 
The only
www.pwpconsult.com | thing lacking is freedom. - Dwight D. Eisenhower