Re:A password hash storage scheme that prevents efficient pa...
by azinman2 on Wednesday September 16, 2015 @09:17PM
(#10231014)
Similarly what do you do for the first couple of users that need to legitimately login but you don't know the secret yet?
(I also haven't read the paper, but the concept is intriguing).
Re:A password hash storage scheme that prevents efficient pa...
by mvksaa on Thursday September 17, 2015 @10:48AM
(#10233643)
There is the option to "leak" a portion of the users' salted hashes, so that the users can be verified upon reboot. This leaked information harms overall security but, even with this option, the scheme is still better than regular salted-hashes.
Re:A password hash storage scheme that prevents efficient pa...
by est on Wednesday September 16, 2015 @10:56PM
(#10231330)
Hmm, hard coded bootstrap dummy users? (Evil)
Re:A password hash storage scheme that prevents efficient pa...
by darkmighty on Wednesday September 16, 2015 @09:37PM
(#10231086)
Maybe the probability that your dummy accounts share the same "line" (their terminology) will be low if the number of users is large enough? (and the number of "lines" is also large enough) Although this conflicts somewhat (but not assymptotically) with the need to have enough users at any "line" so it can be easily computed by the server from scratch when it needs to restart.
Re:A password hash storage scheme that prevents efficient pa...
by fryguy on Thursday September 17, 2015 @03:16PM
(#10235464)
They have the concept of "shielded" users, which are users that are shielded by the shared secret, but don't contribute to it. Though, they don't really specify when to put people into the shared-secret group and when to leave them as shielded. I suppose it's an exercise for the user.
But what stops a malicious attacker from registering a bunch of dummy accounts, and then using the known passwords to reconstruct the Shamir Secret?