Printing chips allow a way around hardware trojans?

Assuming all the other layers are in place and SAFE is going to make the world transparent and finally give us inclusive vice extraction political and economic systems and you’re on the edge about to upload the most important piece of info ever divulged but you can’t just trust any old hardware. A Black Phone for instance wouldn’t cut it because its full of proprietary secret sauce that screams NSA back door. Maybe you find a PC from 1981 that still works and meets your risk level for hardware trojans but then how do you get it to interface with the SAFE mesh without the risk of a hardware trojan in the intervening chips?

Or print the solution from community verified files:

Verified open source code to print a blown up version of an ancient Intel 4004 and wireless antenna other super basic pieces on a fabber. Verified file that can be loaded in to a very high resolution cell phone camera to optically confirm what was printed out was as intended. Verified file to soft proof super simple design from the inside that so that anything that might have been added in the fabbing wasn’t. The simplicity of the design makes it possible to virgin build and then double check with open source verified code.

What do you mean by “print”? I thought 3d printers could only be used to make PCB:s.

If you meant etch silicon, then there already exists methods to make a hardware hack visually indistinguishable in the silicon.


http://www.chesworkshop.org/ches2013/presentations/CHES2013_Session4_3.pdf

But as state here https://eprint.iacr.org/2014/508.pdf it’s distinguishable with a scanning electron microscope (SEM).

Buy a SEM or print your own circuit… the price of a safe SAFE hardware is going up! :wink:

1 Like

Use a ancient simple design with 60 k transistors almost visible to the eye and either scale up to make a more visible (slower) target or use a microscope with a wide view plane so a photograph can be verified by software.

If you’ve got a hardware problem, then you have a hardware problem. It’s an important problem but it is a different one.

If you wanted to move across an airgap and upload a copy to SAFE then you could encrypt it first. Indeed you would want to do that, so that anyone else who knew the file wouldn’t know it was uploaded.

To solve the hardware liability what we need, I wonder, are chipsets that are fully readable and ideally writeable; so, for example we could know that a HDD chipset hasn’t been compromised, where now it’s only writeable.

The second route to knowing the hardware is sound, is winning the argument against spooks compromising our ability for their own gain. We need to remind them who they work for. If they change their understanding and see that supporting the public’s use of encryption is “a Good Thing”, then perhaps they will work to evidence and enable us rather than break those tools that keep us SAFE. I guess we’ll see what rounds of consultations and spook bill writing bring over the next year or two.

1 Like

Agreed with step 2. So what you are suggesting is something like a system made completely of programmable gate array so that the software defines the hardware, which may not be exactly known ahead of time even if open source or aspects of the design are randomized and then proofed by the software once checked. Its interresting because one of the buzzwords out now is software defined radio.