3 ms·
Can We Prove Time Protection?
- crote 8y agoSo if i understand it correctly, if all side effects are perfectly known and controllable, we can prove that they have not been used to leak any information? That's a bit obvious, isn't it? I'm a bit amused by the hardware requirements for this to work: everything must either be resettable, or fully partitioned. I guess this means a single DDR channel per core, no hyperthreading, no dynamic frequency scaling, no hardware features which are dynamically disabled to save power et cetera. Come to think of it, it's probably not even allowed to share a single package due to side effects due to heat creation and power usage? I guess we all have to go back to the equivalent of a very fast, single-core 8008. This is obviously a very serious problem, but it looks more and more like it's going to be impossible to solve this in something even remotely representing current cpu designs and it makes me wonder if it's even worth solving for most users.