The dangerous-titled 'unhackable' processor has survived its biggest hacking test remarkably unscathed. Created by the University of Michigan (UoM), the Morpheus chip design has now been attacked by more than 500 cybersecurity researchers resolutely going at the chip for three months straight. Just imagine what that room smelled like by the end...
Best wireless gaming mouse: ideal cable-free rodents
Best wireless gaming keyboard: no wires, no worries
Best wireless gaming headset: top untethered audio
But in all that time not one has managed to exploit it. Now, three months doesn't necessarily mean it really is completely unhackable, but it certainly makes it pretty damned secure.
Naming a chip 'unhackable', much like a certain GPU limiter we could mention, is tantamount to begging someone to come along and crack it, and that is exactly what the UoM has done in putting Morpheus into the Finding Exploits to Thwart Tampering (FETT) Bug Bounty hunt from DARPA.
Bounty hunting. FETT. Cute, eh? The research was also supported by the DARPA SSITH program too.
Anyway, the bug hunting hackathon ran from June through August last year, and the RISC-V based chip resisted all attempts to compromise it.
Dubbed Morpheus by its creators, the chip design uses a mixture of 'encryption and churn' to first, randomly obfuscate key data points—such as the location, format, and content of a program's core—and then re-randomise them all while the system is in operation.
"Imagine trying to solve a Rubik’s Cube that rearranges itself every time you blink," says team lead, Todd Austin (via Hexus). "That’s what hackers are up against with Morpheus. It makes the computer an unsolvable puzzle.
"Developers are constantly writing code, and as long as there is new code, there will be new bugs and security vulnerabilities. With Morpheus, even if a hacker finds a bug, the information needed to exploit it vanishes within milliseconds. It’s perhaps the closest thing to a future-proof secure system."
This level of obfuscation and churn might sound like an absolute nightmare to code for, or even use, but Austin claims that the design is completely transparent to both software developers and users, because all the randomness occurs within data known as 'undefined semantics.'
What's more, the action of churning those data points is proved to have negligible impact on the actual performance of the system, with the architecture paper (PDF warning) itself claiming just a one percent performance hit.
So far, it's looking good for the 'unhackable' chip then.
