What's the /g/ approved cipher mode of operation

>>106131953 (OP)
>nonce counter

>>106131953 (OP)
Fuck AES, I prefer xChaCha20 (a stream cipher).

>>106131953 (OP)
>block cipher
ngmi
stream ciphers are where it's at

I can't find anyone implementing this.

Basically, a one time pad hard drive, and the other one is the ciphertext hard drive. Actually, the otp hdd can generate the random data on the fly. But anyway, instead of having any copies of the otp hdd, you just have one.

So really they are of equal meaning, since you only write to both drives at the same time during use.

I would call this raid -1. (raid negative one).

So, if you take either hdd with you, you basically have taken the key to the plaintext.

Except both are the key to the plaintext.

idk why this "RAID NEG ONE" isn't implemented, but it could be rather useful.

In case you're wondering, I this isn't meant to replace regular cryptography. You continue to use full disk encryption as usual, and you use well, not gpg, because web of trust is a total nsa rape, but symmetric ciphers.

Ideally we'd have a system of "chain of communications" where you have to alternate between persons, and this ping pong would work like this:
>an archive file is generated containing
>your stupid letter.txt
and
>the key you will use for your reply (very long and generated by the sender)

It could work over literally anything.

But anyway. bouncy privacy has yet to be implemented. And also RAID NEG ONE.

>>106132885

>>106132995

>>106132125
Someone should make the ChaCha-RS method.

>>106133091
What I want is RAID NEG ONE and BounceCrypt

>>106132885
>I can't find anyone implementing this.
because it's retarded
Not only is the encryption completely pointless, since if the feds are gonna raid you, anything you could do with one of the two drives you could have done with just one drive alone without encryption (I guess if they can only find one drive you're safe). Except now you have a RAID-0-like situation where if one drive fails, the data is completely irrecoverable (and not just practically but mathematically) but with half the data storage capacity.

>>106133307
we assume perfect drives (there are mitigations that are obvious, so I think you're retarded)

You're sperging out, but this is practically just like a keyfile.

Again, I think you are retarded, because obviously you use whatever other encryption you want, but this creates a very interesting scenario:

>you absolutely require physical control of two locations to get the data
>there is no alternative, except catching you at the moment of use/creation

Unlike a keyfile, this is perfect encryption.

I mean, given that it's perfect, it's wild nobody has done this.

>>106133479
came here to say this

>>106133479
GCM Is CTR with some Galois thing to generate an authentication tag. Also max file size of 64GB.

>>106132228
What is the point of stream ciphers if you still have to frame packets