1
$\begingroup$

I'm using Diskcryptor to encrypt a 2TB external USB HD and the process is taking over 2 days. I chose a 100-character strong encryption key passphrase (because, why not?) and am now wondering if the length of the key passphrase has any impact on how long it takes to encrypt the drive. I'm also wondering if the length of the key passphrase will have an impact on how long it takes to write and read files from the encrypted drive. If I had chosen a 25-character key passphrase would this process be faster? Would reading and writing from the drive be faster?

Update: It has been explained to be below that what I was calling the encryption key is in fact the passphrase.

$\endgroup$
  • 3
    $\begingroup$ Hard disks are already terribly slow, but USB-2 is also capped at 30 MB/s. Now that's just reading, writing is usually slower at about 17-25 MB/s. Now you've got 2.000.000 MB of disks, so the absolute minimum time should be about 2 days indeed - that's excluding the encryption. $\endgroup$ – Maarten Bodewes Sep 1 '16 at 19:04
  • 1
    $\begingroup$ In general, ciphers take keys of fixed lengths. AES-128, for example, always uses a key of exactly 128 bits. Any time you are able to enter a password or key of a variable length, the software is using a Key Derivation Function (KDF) of some type to turn it into a key of the proper length, and then using that. $\endgroup$ – Reid Rankin Sep 3 '16 at 12:18
3
$\begingroup$

If I had chosen a 25-character key would this process be faster?

That would depend entirely on how Diskcryptor was designed; but it appears it would not.

According to the Diskcryptor homepage, Diskcryptor selects a random symmetric (AES, Twofish or Serpent) key, and uses that to encrypt the disk; it uses the passphrase (what you call the strong encryption key) to encrypt that random symmetric key and put it in the boot sector. When it comes time to use the disk, you'd enter your passphrase, which would allow Diskcryptor to decrypt the random symmetric key; that symmetric key is then used to decrypt (and encrypt) everthing else.

Because the vast bulk of the work is done using that random symmetric key (which is the same size independent of the size of the passphrase), the length of the passphrase doesn't matter.

Now, the page does mention that it can use Cascaded ciphers; say, AES + Serpent. If you have it configured that way, it would be considerably slower than if you were just using AES (which would be easily the fastest because of the AES-NI instructions); that's the only thing I can think of which might be slowing you down.

$\endgroup$
  • $\begingroup$ As it turns out, I am cascading: AES-256 + Twofish + Serpent. (Again, I figured, why not?) Perhaps that's just overkill, however, I'm assuming that the length of the passphrase still doesn't matter here. Thanks! $\endgroup$ – Mark Cramer Sep 1 '16 at 19:10
  • 2
    $\begingroup$ @MarkCramer: well, if your external USB-2 drive limits you to 30MByte/second, that's (and not your cascade ciphers) are likely to be the limiting factor... $\endgroup$ – poncho Sep 1 '16 at 19:19
  • $\begingroup$ That is most certainly the case. So, might as well crank the encryption to 11! $\endgroup$ – Mark Cramer Sep 1 '16 at 19:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.