By default, most FDE have horrible performance hits and require significant tweaking, configuring and benchmarking to get it right depending on hardware, use cases, conditions… I’m sure there are quite a bunch of people out there who don’t want to do any tweaking while still having the performance they paid for.
Unless what you are doing is heavily I/O dependant (mostly heavy database workloads), that’s not really true anymore, especially with a modern CPU and say, LUKS encryption. Phoronix has a recent review of FDE using LUKS, and apart from synthetic I/O tests, the difference isn’t really observable.
Try cryptsetup benchmark on your pc and look at the results for aes-xts for example.
What I’m getting at is that for people using FDE, any performance hit is worth it compared to worrying that you’ve covered every angle.
By default, most FDE have horrible performance hits and require significant tweaking, configuring and benchmarking to get it right depending on hardware, use cases, conditions… I’m sure there are quite a bunch of people out there who don’t want to do any tweaking while still having the performance they paid for.
Unless what you are doing is heavily I/O dependant (mostly heavy database workloads), that’s not really true anymore, especially with a modern CPU and say, LUKS encryption. Phoronix has a recent review of FDE using LUKS, and apart from synthetic I/O tests, the difference isn’t really observable.
Try
cryptsetup benchmarkon your pc and look at the results for aes-xts for example.