Re:The FBI stole an Instapaper server in an unrelated raid
by pkteison on Thursday June 23, 2011 @06:13PM
(#2690141)
Salts, done properly, vary per user not per server.
[ Parent
| on HN ]
Re:The FBI stole an Instapaper server in an unrelated raid
by engtech on Friday June 24, 2011 @12:06AM
(#2690975)
but I'm assuming that if they have the code base, the plaintext user names (emailaddys) and the salted password, then they would have whatever the per user salt is.
Re:The FBI stole an Instapaper server in an unrelated raid
by Kadin on Friday June 24, 2011 @02:00AM
(#2691195)
Right, but if the salt varies per user, then you end up doing a bruteforce on each user's password; it's no longer a precomputation attack. There are no "Rainbow tables" in this case.
However, if you find Hale's bcrypt page (http://codahale.com/how-to-safely-store-a-password/) convincing, and I do, salting really doesn't matter because with modern GPUs you can bruteforce a reasonably-sized alphanumeric password, if the hash algorithm is a general-purpose (read: fast) one.
The solution is not salt, the solution is to use a purposely slow hash function.