Much analysis and testing has been done surrounding SSD long term performance issues. In other words, it seems like the industry has a pretty good handle on how to manage things like garbage collection, TRIM, reserved space, write amplification, wear leveling, etc. when used on a normal file system.
My question is, has anyone ever looked at the implications (if any) of using FDE on an SSD? I don't know how most FDE software works, but my understanding is at the extremes you either have a completely new on-disk format with new filesystem drivers, or else just a huge encrypted file with its own internal filesystem, or an implementation that's somewhere in between.
So it sounds like something like this could potentially mess up the assumptions SSD firmware makes when implementing garbage collection and TRIM (i.e. would TRIM even be useful if all you're ever doing is overwriting sections of a huge file). Have there ever been any tests along these lines?
My question is, has anyone ever looked at the implications (if any) of using FDE on an SSD? I don't know how most FDE software works, but my understanding is at the extremes you either have a completely new on-disk format with new filesystem drivers, or else just a huge encrypted file with its own internal filesystem, or an implementation that's somewhere in between.
So it sounds like something like this could potentially mess up the assumptions SSD firmware makes when implementing garbage collection and TRIM (i.e. would TRIM even be useful if all you're ever doing is overwriting sections of a huge file). Have there ever been any tests along these lines?
Last edited:
Facebook
Twitter