Re:League of Legends database compromised, passwords hashed ...
by teamonkey on Friday June 08, 2012 @11:31PM
(#4087077)
I was going to ask: is there actually any benefit salting a password that will be encrypted?
[ Parent
| on HN ]
Re:League of Legends database compromised, passwords hashed ...
by DanielRibeiro on Friday June 08, 2012 @11:40PM
(#4087094)
It is really kind-of very good practice to salt passwords. We've talked a bit about this recently here on HN:
Re:League of Legends database compromised, passwords hashed ...
by drivebyacct2 on Saturday June 09, 2012 @12:25AM
(#4087179)
Despite the other reply you got (which was my original 'lol'), no, there isn't an advantage to salting a password that is going to be encrypted with symmetrical encryption. I mean, what would that even mean?
Encrypt("salt"+"password", key) = ENCRYPTED_DATA
Decrypt(ENCRYPTED_DATA, key) = "salt"+"password"
Re:League of Legends database compromised, passwords hashed ...
by dmak on Saturday June 09, 2012 @01:15AM
(#4087280)
In short, it is BEST to HASH with a salt. Hash with a salt and don't use symmetrical key encryption methods.
Re:League of Legends database compromised, passwords hashed ...
by drivebyacct2 on Saturday June 09, 2012 @01:17AM
(#4087283)
Yup, always hash with a salt (or use an algorithm that does it for you). I added another example below for this guy. And yes, definitely don't use symmetric encryption.
Re:League of Legends database compromised, passwords hashed ...
by teamonkey on Saturday June 09, 2012 @01:08AM
(#4087260)
Thanks. I'm no security expert but I couldn't see how it would make a difference.
Re:League of Legends database compromised, passwords hashed ...
by drivebyacct2 on Saturday June 09, 2012 @01:10AM
(#4087270)
Sure, to help illustrate it further: (don't use SHA1, I'm using it to avoid a discussion of why you wouldn't manually use a salt with bcrypt)
user1: SHA1("salt123"+"puppydog11") = SOME_HASH
user2: SHA1("salt456"+"puppydog11") = SOME_DIFF_HASH
This way, even for the same password, there are different salts, and thus different hashes. No one can run a rainbow attack unless they've generated rainbow tables for that salt (frankly, it would be possible, but useless to generate a rainbow table for a specific salt)