153 reputation
5
bio website blog.jeremymartin.name
location
age
visits member for 3 years, 1 month
seen Feb 3 at 15:26

May
15
comment Cryptographic Security of Dynamically Generated, Non-Random Salts
Just realized I never accepted this answer - sorry for the long overdue accept! Not sure I ever got my question resolved, but I definitely appreciate the time and thought you put into this.
Sep
21
comment Cryptographic Security of Dynamically Generated, Non-Random Salts
@ThomasPornin Thanks again. I'm starting to wonder how to actually get an answer to my question though. c) may indeed be true, but it doesn't really address 1), 2), or 3) in my question. This question is academic in nature and focuses on whether or not the proposed approach has negative implications for the cryptographic properties of the secured password. This has all been good input and I appreciate it - I just want to see 1), 2), and 3) explicitly addressed...
Sep
21
comment Cryptographic Security of Dynamically Generated, Non-Random Salts
Thanks for the comment. As I've stated already, though, I'm entirely comfortable with accepting that "trusting the experts and proven solutions" is sound advice. That's an answer to a different question though (e.g., "should I roll my own solution?"). I have an academic interest in this, however, and I've very explicitly outlined what my exact questions are, with regards to the proposed solution. (Also, apologies, as this isn't really targeted just as you, and I really do appreciate you taking the time to answer.)
Sep
16
comment Cryptographic Security of Dynamically Generated, Non-Random Salts
@Thomas - I swear I plan on giving some upvotes and accepting like it's Christmas soon, so thank you for sticking with me. I still feel like my initial question hasn't been answered though. Trusting the experts and vetted utility classes for this stuff is sound advice, but that doesn't help me understand it any better. Providing a plain-text salt along with the hash is either; a) not as useful to a hacker as I think it is, or b) a dismissible concern for other reasons that haven't yet been mentioned.
Sep
16
comment Cryptographic Security of Dynamically Generated, Non-Random Salts
@Thomas - agreed on the effectiveness of a unique salt in rendering precomputed tables useless. Should I or should I not be bothered, though, by the greatly reduced order of complexity that having the unique salt associated with a hash provides a brute force algorithm (i.e., the 80^24 --> 80^8 reduction mentioned above)?
Sep
16
comment Cryptographic Security of Dynamically Generated, Non-Random Salts
Thank you for all the replies. I'd have to suggest though, that, depending on the site, a single password valid only on the targeted site can still be of great value to a hacker. With regards to your statement "if you are using a random salt per password then you are already increasing the time it takes to get anything of value to the point of silliness", that's actually part of the question itself. It certainly makes precomputation-based attacks pointless, but I'm hoping for some math proving that giving them the salt doesn't nullify the "silliness" of a brute force attack.
Sep
16
comment Cryptographic Security of Dynamically Generated, Non-Random Salts
Specifically, let's say that the range of valid password characters is 80. If I know that all passwords use the same salt, my labor is reduced from 80^24 iterations (for an 8 byte password and 16 byte salt) to only 80^8. It just seems that having salt values readily available, even if they're unique per password, has the same effect. Is this not the case? Or, perhaps, is that indeed the case, but the iterations aren't as important as I'm making them?
Sep
16
comment Cryptographic Security of Dynamically Generated, Non-Random Salts
@Thomas - Thank you for the detail in this answer. I understand your point about economies of scale. What I'm struggling with is that it seems to me that having the salt readily available also lends itself to economies of scale (not in the form of precomputation, but in the same way that a fixed/static salt would (and as described by @Nick)). cont...