3 ms·
Ok, now we can run a verified OS on multiple cores. Now, what value has verified software which runs on buggy or (theoretically) compromised hardware? In my op
by progman 10y ago
Ok, now we can run a verified OS on multiple cores. Now, what value has verified software which runs on buggy or (theoretically) compromised hardware?
In my opinion safe computing requires verified software AND verified hardware. Since proprietary SW/HW cannot be trusted in general truly safe computing vitally depend on open software and open hardware.
- satai 10y agoIt is one (huge) step in the right direction. We are not there yet but solving part of the problem is improvement anyway. HW exploits are harder then SW ones so the practical implication is more safety regardeless the closed HW/ quality SW is not perfect.
- naasking 10y ago> Now, what value has verified software which runs on buggy or (theoretically) compromised hardware? The vast majority of bugs and vulnerabilities are due to software, not hardware. There's tremendous value in pushing that boundary back to the hardware layer.
- curried_haskell 10y agoBut does everyone know that there's tremendous value in pushing that boundary? Do they know that pushing the boundary has the best value? You should tell them how tremendous your value is. That you have the best value.
- posterboy 10y agoWhen hardware comes with multiple levels of firmware, there is hardly any difference.
- naasking 10y agoA search for "linux" in NVD returned over 4,500 entries. "firmware" has barely 900. So I'd say there's definitely a meaningful difference, and operating system kernels have been used in firmware (Linux and L4 have at least), so formally verifying kernels can eliminate firmware vulnerabilities too.