lainchan / sec

DIY Hardware Encryption

Miho Iwakura · 2018-07-25 13:14
Did any of you make a homebrew cryptoprocessor of sorts? Looking to mess around with some microcontroller to encrypt datastreams for fun. I doubt anyone knows how to also make it tamper-proof without given access to a fuarrkin chip factory, so that part is off the table. I know this is actually /sec/ + /tech/ + /diy/, but I didn't know where to post so it ended up here by flipping a coin twice.
Donna Hawthorne · 2018-07-25 14:27
>Looking to mess around with some microcontroller to encrypt datastreams for fun. Are you looking to solve any particular problems? >I doubt anyone knows how to also make it tamper-proof without given access to a fuarrkin chip factory, Tamper proof is always either a misnomer or marketing copy. Real systems can be tamper evident and tamper resistent. >I know this is actually /sec/ + /tech/ + /diy/, but I didn't know where to post so it ended up here by flipping a coin twice. You should have flipped at least three times. This is your daily reminder that you are bad at math and logic and should not roll your own crypto.
The Oracle · 2018-07-25 15:46
>>5263 > solve any particular problems My lack of ideas for my bachelor's thesis. >Tamper evident and tamper resistant Yep, my bad, I meant tamper resistant. Wipe-it-all sort of thing, though I guess if it's diy it'd be really evident and easy to simply desolder off. >flipped 3 times Um, no? >should not roll your own crypto I guess everyone on this board already knows this
Joe Bloggs · 2018-07-26 09:33
>>5263 >You should have flipped at least three times. This is your daily reminder that you are bad at math and logic and should not roll your own crypto. If a coin lands on the same side twice what is the advantage of flipping it a third time?
Miho Iwakura · 2018-07-26 16:34
>>5263 >should not roll your own crypto you gotta learn somehow. OP did specify "for fun".
Lady 3Jane.. · 2018-09-23 19:38
Here for the hardware temper proofing https://youtu.be/n9zPq4_Evvw
Unspecified · 2018-09-23 21:30
>tamperproofing Make your cryptographic device powered by a battery, store the keys in volatile memory that is lost if power is interrupted. 3D print a shell around it, embed a alternating wire mesh in the plastic as you print it. Use put current through that wire mesh and then put charge through it find something in that to measure that changes depending on the wire-meshe's 3D form, then use that to generate the keys. No idea if that would work, but it's based on half remembered principles of those self powered tamper wiping cryptographic data drives Sorry if some of this info is inaccurate, it's only a fuzzy memory of a news article.
Random · 2018-09-23 21:42
Homebrew cryptoprocessors are a pain to make since side channel sniffing attacks. The reason all real mass manufactured cryptoprocessors are reliant is because their INTERNAL filtering and smoothing of input voltage of supply. If there was none then you could put a really sensitive ADC on supply of microcontroller or homebrew cryptoprocessor you made and sniff out internal code or even private keys simply through the amount of current chip drains from source and voltage drops. This is why SIM cards have filtering on internal voltage supply inputs. This is why mass manufacturing is usually much more reliable than making your own and hoping a hacker won't fuarrrk you over. But if the PCB and its application can't be easily distinguished or reverse engineered (adding a bit epoxy with color dye to the pcb won't be bad idea) then I say power to you. Ain't that hard to program a microcontroller and add encryption abilities. Depending on what encryption you want to implement it might be a challenged depending on what microcontroller you want to use, but other than this general discussion we can't progress unless you give specifics.