21–30 of 109 posts
We can't actually tell much about their password storage mechanism from that post. For one, they mention "encrypted hashes", which might mean encryption, might mean hashes, or might mean both; for two, they don't describe how they determined which users had weak passwords. If they were using some kind of reversible encryption scheme, it's possible that they ran that function against their database and determined comm…
I think availability bias is kicking in, with each additional announcement, more links about announcements are posted (for karma) or more sites feel like they can announce, under the cover of larger leaks, that they themselves have been compromised.
Probably just me.
The platform guys at Riot were the sharpest group of engineers I've worked with. Smart and able to get stuff done without bike shedding. They did stuff right.
Is there a good authentication system that doesn't involve storing passwords, hashed passwords, or encrypted passwords in the master database?
There's Secure Remote Password, which stores a non-reversible verifier instead: http://en.wikipedia.org/wiki/Secure_Remote_Password_protocol Apparently it's quite tricky to implement properly though.
(I don't see how there's any possible protocol that behaves otherwise: if Alice is capable of proving she possesses a secret to Bob through a protocol, then Bob can always discover Alice's secret by playing that protocol with himself, enumerating every possible secret until the protocol succeeds. And so can anyone who knows everything that Bob knows about Alice's secret.)
The SRP paper doesn't say much about this except that brute-force dictionary attacks are "expensive". I don't think they make specific claims about computational hardness like scrypt. The paper was written in 1997.
"An attacker who captures the host's password file cannot directly compromise user-to-host authentication and gain access to the host without an expensive dictionary search."
We can't actually tell much about their password storage mechanism from that post. For one, they mention "encrypted hashes", which might mean encryption, might mean hashes, or might mean both; for two, they don't describe how they determined which users had weak passwords. If they were using some kind of reversible encryption scheme, it's possible that they ran that function against their database and determined comm…
In case I do: 1) The site doesn't need my password in plaintext. 2) If the server is compromised, the attackers don't just have salted or unsalted hashes THEY HAVE FULL PLAINTEXT PASSWORDS (unless it's some weird case where they have DB but not FS access).
No, either way, they're wrong.
(Sigh, I'm sorry for the snippiness, I'm just tired of coming into everyone of these threads with people who earnestly try to regurgitate what they heard from the last thread about this and manage to give bad advice or implications.)
We can't actually tell much about their password storage mechanism from that post. For one, they mention "encrypted hashes", which might mean encryption, might mean hashes, or might mean both; for two, they don't describe how they determined which users had weak passwords. If they were using some kind of reversible encryption scheme, it's possible that they ran that function against their database and determined comm…
You can't do that with salt. That's half the point of salting.
It says encrypted not hashed, these are not the same thing. Encryption uses keys and without the keys you can't get the value without a huge brute force effort. Hash without salt can be broken trivially. It'd bad either way but encrypted is way better than leaked hash.
I was going to ask: is there actually any benefit salting a password that will be encrypted?
Encrypt("salt"+"password", key) = ENCRYPTED_DATA
Decrypt(ENCRYPTED_DATA, key) = "salt"+"password"
I'm nearly positive the US code salted the hashes (it's been a long time since I was in that code :). I say US code because (again, memory is dim) I think they had a partnership with a company in the EU and that (might have) included authentication. The platform guys at Riot were the sharpest group of engineers I've worked with. Smart and able to get stuff done without bike shedding. They did stuff right.
Is it just me or have the number of announcements on HN about password leaks gone up in the last few weeks? I think availability bias is kicking in, with each additional announcement, more links about announcements are posted (for karma) or more sites feel like they can announce, under the cover of larger leaks, that they themselves have been compromised. Probably just me. [1] - http://en.wikipedia.org/wiki/Availabil…
I'm nearly positive the US code salted the hashes (it's been a long time since I was in that code :). I say US code because (again, memory is dim) I think they had a partnership with a company in the EU and that (might have) included authentication. The platform guys at Riot were the sharpest group of engineers I've worked with. Smart and able to get stuff done without bike shedding. They did stuff right.