Timeline for Reducing size of hash function
Current License: CC BY-SA 3.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 23, 2015 at 22:28 | comment | added | Nova | @Harry: Yes, with this you will need first-preimage resistance. 80 bit should be enough. | |
| Feb 23, 2015 at 21:51 | comment | added | Harry | I'm not really asking about a password. I am distributing a hash of something and I want it to not be bruteforced until the original data is shown and then verified against the hash. So I'm looking for the smallest size hash of that data that can't be bruteforced between the time the hash is shown and the data is produced to verify it matches the hash. I know 128bits or higher for a hash is "Recommended" but I guess I'm asking if there's a way to go smaller than 128 and still be in the realm of acceptable. | |
| Feb 23, 2015 at 13:25 | comment | added | Nova | @Harry: For the number of iterations of PKBDF2, yes, 1 million is good. That won't help you with passwords like "password". Even the highest amount wouldn't help you there. The iterations are only to help mediocre passwords to get a good security value. Bad passwords will stay bad. See here for more informations: stackoverflow.com/questions/6054082/… | |
| Feb 23, 2015 at 13:24 | comment | added | Nova | @Harry: It's all about the entropy and against what attack you want to protect yourself. 80 bit (10 output bytes) is (in the moment) enough for let's say password hashing for a web site (first- and second-preimage resistance), but not for hashing a document to sign it with asymmetric encryption (collision resistance). Be warned, 80 bit could be in realm of possibility to be breakable in 2020 or so. Use it only if you really need to. 128 bit is recommended. | |
| Feb 23, 2015 at 12:09 | vote | accept | Harry | ||
| Feb 23, 2015 at 12:07 | comment | added | Harry | Would making n=80bits and doing a million or so iterations of SHA256 add enough security that a one month bruteforce is out of the question? Do you have any recommendations on minimum bit size, and key stretching needed to throw out the possibility of something being bruteforced to find a collision in a reasonable amount of time? | |
| Feb 23, 2015 at 11:03 | history | answered | Nova | CC BY-SA 3.0 |