Timeline for Splitting a password for dual roles [duplicate]
Current License: CC BY-SA 3.0
14 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 7, 2021 at 7:18 | history | edited | CommunityBot |
replaced http://tools.ietf.org/html/draft with https://datatracker.ietf.org/doc/html/draft
|
|
| Dec 3, 2013 at 0:17 | history | closed |
hunter user6961 archie Cryptographeur rath |
Duplicate of How to derive two keys from one password | |
| Dec 2, 2013 at 17:13 | answer | added | Ilmari Karonen | timeline score: 0 | |
| Dec 2, 2013 at 16:51 | history | edited | Ilmari Karonen | CC BY-SA 3.0 |
copyedit
|
| Nov 6, 2013 at 19:22 | comment | added | disk eater | I'm sorry I know this is a weird question my example with HMAC was designed to avoid just this confusion but it remains. My question is not as much as protecting each password as it is ISOLATING the fate of derived passwords to trace back to a testable common password.. see "box fish" vs "puffer fish" comments above. If you can guess 'pass1' of "box" 'pass2' = "fish" portion cannot be tested offline without attacker having knowledge of "pass2"... there is not an equality in risk exposure to each derived key. | |
| Nov 6, 2013 at 14:44 | comment | added | John Deters | disk eater, try separating the concept of "password" from "key". The password is what the human remembers, the key is what the algorithm needs. Algorithms like PBKDF2 translate a password into a pile of key material. So instead of splitting the password, you divide the key material. Knowing half the pile reveals nothing about the other half. Of course, guessing the password now has two independent systems that can test your guesses, and if you guess right, you can generate both keys, but that's a risk inherent to your requirements - not to the technology. | |
| Nov 6, 2013 at 14:01 | answer | added | K.G. | timeline score: 3 | |
| Nov 6, 2013 at 3:24 | comment | added | disk eater | if an all knowing algorithm cut input password "box fish" into separate passwords assigned pass1 = "box", pass2 = "fish" knowledge of 'pass1' is more or less useless to derive 'pass2'. Whereas splitting "puffer fish" into "puffer" and "fish" yields disaster. | |
| Nov 6, 2013 at 3:17 | comment | added | disk eater | They are different questions. If I derive an input yielding 'pass1' by brute forcing 'pass1' this input also reveals 'pass2'. I am asking about ways to construct outputs such that difficulty of deriving 'pass2' exceeds the simple entropy of the input. For example | |
| Nov 6, 2013 at 1:27 | review | Close votes | |||
| Dec 3, 2013 at 0:17 | |||||
| Nov 5, 2013 at 23:59 | comment | added | disk eater | Thank you, I'm looking for basic understanding of best that can be done with a given effective input entropy (amplified or not, good or bad) | |
| Nov 5, 2013 at 21:11 | comment | added | hunter | Have you considered using a slow KDF like PBKDF2 or Scrypt to compensate for the low entropy of the password, and then simply split the output into two keys? | |
| Nov 5, 2013 at 20:52 | history | edited | user991 | CC BY-SA 3.0 |
improved grammar
|
| Nov 5, 2013 at 20:44 | history | asked | disk eater | CC BY-SA 3.0 |