3 ms·
> The original password is never stored thus keeping it a secret even from the website you provided it to Never stored, hopefully, but the website has a chance
by pmontra 6y ago
> The original password is never stored thus keeping it a secret even from the website you provided it to
Never stored, hopefully, but the website has a chance of seeing it everytime one signs in. A malevolent site or developer or admin could store the password somewhere or try to reuse it on a number of other well known sites. Hence, one different password per site to protect also against the sites themselves.
- mooreds 6y ago> Hence, one different password per site to protect also against the sites themselves. Preferably one generated by a password manager, large and random and totally unrelated to any other password of the user.
- hinkley 6y agoThere was a class of password algorithm that used your password and some other information as a seed for generating an asymmetric cypher, but the original Stanford proposal was apparently a bit light and nobody ever implemented it. I bumped into some new algorithms in the same category, but I'm blanking on the clever name they used to describe them all. The server would only ever have your 'public key'. I'm not sure where you get the saltiness to prevent lookup tables but still let me log in from three devices with the same credentials.
- dinosaurdynasty 6y agoDo you mean <https://en.wikipedia.org/wiki/Zero-knowledge_password_proof> https://en.wikipedia.org/wiki/Zero-knowledge_password_proof>?
- tialaramex 6y agoThe Stanford proposal you're talking about was named pwdhash. This approach is sometimes called a "Deterministic password manager" and that might be the phrase you wanted. There are several problems with this approach, especially with regard to the need to sometimes change passwords. But the point your parent was making is that in the modern era (certainly this century) none of this was necessary, there are asymmetric PAKEs which render it unnecessary for a site to know your password at all in order to confirm that you still know what it is.
- pauldino 6y agoI think you're referring to PAKE (password-authenticated key exchange). SRP (secure remote protocol) might be the Stanford proposal you're referring to, although it's actually fairly commonly implemented as far as PAKEs go (Apple for instance uses it for their iCloud Keychain). https://blog.cryptographyengineering.com/2018/10/19/lets-talk-about-pake/ https://blog.cryptographyengineering.com/2018/10/19/lets-tal...
- hinkley 6y agoThis is it, thanks. PAKE. No wonder I couldn’t remember it!
- kaoD 6y ago> the website has a chance of seeing it everytime one signs in Indeed. I don't understand why challenge-response auth isn't more widely deployed, or even part of the web APIs. Sure, it doesn't mitigate ALL attacks, but it does mitigate SOME attacks even if you're already on TLS.
- ziml77 6y agoWe're finally seeing challenge-response in the form of WebAuthn, though I'm not sure why it's taken so long to get to where we are now. It's not like challenge-response is a new concept. And you don't need a hardware security key to handle challenge-response. If you don't have that as an option, you could use the password to directly generate the response. Of course the browser has to support this, otherwise you'd be relying on the page's JS which doesn't help with the attacker controlled server scenario.
- tialaramex 6y agoWebAuthn is a lot more than just challenge response though. I think other schemes that did a lot less had plenty of time to displace passwords if they were going to. And they didn't. So, implement WebAuthn and have done with it.
- ziml77 6y agoAgreed. My point here is that there was no reason that we couldn't have had something better than passwords in the time between first using passwords on the web and now.
- hnick 6y agoDebug logging of requests is sometimes left on by mistake too.