Timeline for Can we encrypt part of the XChaCha/XSalsa nonce?
Current License: CC BY-SA 4.0
10 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 3, 2020 at 8:45 | vote | accept | Modal Nest | ||
| Nov 6, 2020 at 23:54 | comment | added | Modal Nest | Thank you for your several responses. I can't add the detail here so I will edit my question (assuming I can). | |
| Nov 6, 2020 at 22:23 | comment | added | kelalaka | @ModalNest Updated the Answer with more details. I Hope this satisfies you. | |
| Nov 6, 2020 at 22:22 | history | edited | kelalaka | CC BY-SA 4.0 |
Added the XChaCha20 part to the beginning and addressed the OP's question more deeply.
|
| Nov 5, 2020 at 22:52 | comment | added | Modal Nest | I know, and I know the nonce is random. I've not suggested anything to do with randomness of the nonce. I am suggesting to use nonce [0,15] to randomize the key as normal, then run a 2nd round with the nonce [0,15] to encrypt the actual nonce [16,23]. Then run through the encryption with the actual nonce as normal. Apologies if I am not being clear, but I am simply suggesting to encrypt the nonce [16,23] with an additional 'key randomization' round using the nonce [0,15] again but with one of the nonce bytes incremented by '1', as they occupy the 'counter' blocks. | |
| Nov 5, 2020 at 22:24 | comment | added | kelalaka | nonce[0:15] is used to randomize the key, nonce[16:23] is used as nonce. The nonce should be already generated uniform randomly, so each bit of the nonce is already expected to be parted... | |
| Nov 5, 2020 at 22:14 | comment | added | Modal Nest | @kelalka Separate what? I am suggesting to encrypt the part of the nonce that isn't used to randomise the key, with the part that is used to randomise the key. I have said nothing about combining any separate processes into one. I am not suggesting to combine the nonce. In fact, I am suggesting running an extra hChaCha round which would further separate the two parts of the nonce. | |
| Nov 5, 2020 at 18:04 | comment | added | kelalaka | There is a good reason to separate them, the first part is used to randomize the key. | |
| Nov 5, 2020 at 18:01 | comment | added | Modal Nest | I know the nonce is random. To be clear, I'm suggesting to encrypt the final 8 bytes of the nonce but use the original (unencrypted 8 bytes) to run the cipher. On the other side decryption would only require the 16bits of the nonce (and key) to decrypt the other 8 bits of the nonce. | |
| Nov 5, 2020 at 17:53 | history | answered | kelalaka | CC BY-SA 4.0 |