[Cryptography] Data remanence on solid state storage

[John Levine]

It appears that Tom Mitchell <mitch at niftyegg.com> said:
>-=-=-=-=-=-
>-=-=-=-=-=-
>
>On Wed, Aug 7, 2024 at 2:46 PM Justin Goldberg <justgold79 at gmail.com> wrote:
>
>> Does anyone have updated guidance or can point to a good study on data
>> remanence with SSD storage?
>>
>> Also perhaps a study that includes successful recovery on the "faster"
>> Secure Erase command, which afaict, has only been implemented on solid
>> state disks and was the reason for it's creation.
>
>The secure erase command is one NECESSARY aspect.
>
>I have no secret secrets so I simply  generate a block of random data and
>with a script make a couple hundred thousand files to fill up the device
>after loading a "different operating system" or hang it on another machine
>that allows access to the raw device. ...

Except that SSDs do all sorts of clever wear leveling. As far as I
know you have no way to tell where on the physical device your files
went. For all we know, it noticed that you were writing the same data
over and over so it helpfully adjusted its internal pointers so there
are a million pointers to the same block and you should be grateful
that it avoided those 999,999 redundant writes.

I suppose if you write different junk to every block that will
eventually force it to overwrite most of the SSD but even then, who
knows if there were blocks marked bad that it didn't overwrite but
might still be recoverable.

R's,
John