13 ms·
This view on sharing hardware seems too pessimistic to me. Strong isolation should be achievable when it is the design objective and and priority. On the end-u
by jssmith 8y ago
This view on sharing hardware seems too pessimistic to me. Strong isolation should be achievable when it is the design objective and and priority.
On the end-user/device side, JavaScript has proven that sharing hardware with untrusted code is just too valuable to give up, and a similar dynamic continues to play out in the cloud.
I'm also happy whenever a headline brings attention to these types of problems because I believe we can fix them.
- phkahler 8y ago>> This view on sharing hardware seems too pessimistic to me. Strong isolation should be achievable when it is the design objective and and priority. I think it's spot-on. Isolation has never been the top priority. Regardless of the priority, if one simply views all "untrusted code" as "code created by my enemy" the solution becomes clear - don't run it. >> JavaScript has proven that sharing hardware with untrusted code is just too valuable to give up... This is something I strongly disagree with. Javascript has become a common way to handle things but it's not the only way. In my opinion it's the lazy way - most people want to just use what's there and common rather than solve a few problems. >> I'm also happy whenever a headline brings attention to these types of problems because I believe we can fix them. One thing I'm seeing is that CPUs with speculative execution have lower performance/watt and performance/area. So in the cloud space it seem like just going to simpler CPU cores may be a solution (risc-v Esperanto Minions for example). Obviously that's not a solution where single thread performance is critical.