But how did we end up in this horrible state of authentication? Why don't we have something as easy to use as the DNS, but for authentication?
Imagine what authentication would look like, if we all started running is the same direction, instead of implementing our own authentication again and again. If we had something open source, that would allow you to sign in to all the sites you use, while completely protecting your privacy, so none of them know who you are.
This dream can come true. Technically at least. I've taken the the first baby steps with https://promiseauthentication.org which proves that this is possible.
But, for this to become a reality, we really need to start running in the same direction. A collective movement towards a sane, privacy-first Single Sign-On provider that's easy to use for everybody.
Aligns really well with using your own domain for email instead of gmail.
Only problem is there aren't any password managers that implement it, so it's not actually practical to use as a primary authentication factor yet.
How does it prove that?
I think this would need serious widespread adoption until we saw benefits too. And you’d need some big names...like Google. Which probably will never happen.
Be aware that doing this now means your DNS provider and domain registrar become vectors for hackers to take over your email account, so make sure these are companies your trust and your access to these accounts is as secure as possible (ie strong unique passwords and app-based, not SMS-based two-factor authentication)
It seems like things should be more granular, such that being banned on YouTube doesn't make your thermostat quit working, ruin your phone contacts/photos/etc, or cut you off from your unspent AdWords funds.
If successful, this would impose a cost to Google for shutting down accounts capriciously and incentivize them to do better.
This would be a challenging lawsuit to win. You’d probably need support from an organization like EFF to manage it.
PS: I did not do anything wrong but still suffered lot of psychological pain due to this mistake by Amazon's internal security.
I think the conclusion of the article is flawed. I think the risk of getting locked out is far lower than the odds of any single, or even all of, other (non-major tech co) website you might join getting breached. It's fair to argue the impact might be less also - and I'm happy to have this debate.
In my experience, typical users aren't the ones that get their google accounts banned - they are always banned for doing something significantly more sophisticated.
Yes, I need to move away from gmail...
If Google can, without due process and fair warning, remove your existence then this is a power that should be delegated to the relevant authority, namely the "justice" system to make such considerations.
If your house could be removed at a whim because a bot decided you were a bad person it would likely cause an uproar, it wouldn't be tolerated.
Yet here it is. Google can offer their services and the legal system seemingly doesn't want to be involved.
Why?
You can get also locked out of your phone
You can get also locked out of the email that you actually use for signing in because you can never remember the password and they stupidly ask you to change it every 6 months with bizarre constraints
You can get locked out of your password manager
You can get hijacked
The business you're signing into can go under
The odds of these things happening are to be weighted against each other
Yes you shouldn't use third-party sign-in for the bank account that holds all your money (though most consumer bank 2-factor authentication mechanisms, sadly, rely on third parties such as phone and email provider)
Yes it's also ok to use third party sign-in for the odd website that you don't care about which somehow insists on asking you to create an account
There are no absolutes in security risk management </>
I always use my email to sign up. If I can't register by email, there's a good chance I won't use that service.
Although using e-mail sign-up actually provides a number of privacy-related benefits over using the Google account way, it still doesn't solve the main problem - because the e-mail usually is GMail anyway (and when it's not - you can get blocked by Microsoft, Yahoo or anything else too, and you can also loose your own domain).
"Every respectable service" should let and recommend (but not require) you set a secondary e-mail and/or another way to contact you but they usually don't.
The alternative would have been to use email, which, presumably would have been a gmail.com address.
If Google locked you out of your account, you wouldn't be able to access your email account either.
Finally, is it not possible to require that all such block critical to someone’s data require some form of govt approved appeals process?
I’m asking these questions so maybe someone can enlighten me on why they were not yet attempted, or if they where, why they failed? Is it legal complexity? Cost? Lack of large scale support, as in, is it only a niche concern that only the HN crowd is complaining about?
One thing I don't understand is: the author suggests a remedy is using your email address instead of third party sign in. But what if your email address is Gmail? For example, I just went to my Stack Overflow account and added my email address as a sign in method. But then of course I realized: my email address is Gmail. So what's the difference? How are we supposed to put this into practice without running our own email? Email is just another form of third party auth.
I have around 8 gmails. Theyre all connected to various things via OAuth2 and I have never once had any of them locked.
Maybe im ignorant to some detail here, but, this sounds like a spammer retaliating because they got caught.
However, for random sites, entering an email/password worries me because I have no idea how this password is handled server side, is it stored in plaintext or with a weak algorithm? The vast majority here don't care that much because they use a password manager but I'm worried about the ones that don't, they can be impacted if there is a database leak or if the site owner is shady and starts looking through its database for passwords that look reusable and try them on other important website. How easy would it be to set a nice honeypot website that requires a username/password?
A properly set up google sign in makes it impossible to do that at least. Thoughts?
I restored it but automatically had to start thinking about a backup plan where I’d have to point my MX records away from Gmail to something else immediately in order to prevent email downtime.
This way I should maintain my own email server, because I can be locked out of my email by any of cloud providers as easy.
Paying for your own domain also comes with its own troubles. If you're not using Google (or some other service) as your mail forwarder, good luck being able to email anyone. Stealing you custom domain is also a real possibility, and negates your investment in Gmail 2FA.
But if the account I'd use to sign in is my Gmail account, and they had locked that account, wouldn't I be locked out anyway?
Please explain it to me if I'm wrong (cause that happens often).
We need a name and shame site for websites that can't be bothered to write a back end database for the 3 columns needed to store emails, salts, and hashes.
Internally, on my website, i may have an account, that i then link to this Google or FB account.
1) If Google shuts down an account, authorization might still work for the purpose of logging in somewhere else. Your email might not work anymore, like any other services within Google. But authorization does. That it does not, is an implementation detail.
2) Since internally i have created an account, that is only linked to your Google account, i can always allow you to also login via any other method. Maybe with your facebook account.
3) I think, things will become better. You use your devices to authorize yourself. And i then trust your device. Then there is no 3rd party involved anymore.
I guess at least if you’re using your own domain, you’d be able to repoint mx records to do the recovery.
I tend not to use it for my personal accounts, but honestly, Google Sign In for our work systems has generally been a good experience. Works well for our small team, anyway.
Though the post does have a good point on that non-email auth providers add more risk to the equation.
I would have canceled my facebook account long ago if I had not chosen their login for a (unknown) number of service.
What would be a better alternative? Use same credentials everywhere? No, because it is just a matter of time it would leak out of one service. Use unique credentials for each service in local password manager? Nay, because most of us at least want to sync between desktop an mobile. Use something like Chrome's password manager? That bears similar dangers like those the article points out.
2. Developers should always allow restoring passwords for SSO only users, it is ridiculous for it to even be an issue.
3. As a user, refrain fro using free email accounts to identify on a platform, as others already said, buy a domain not an expensive one, and stick to it, remember to renew, and setup your email address with a reliable service, there are good providers for $1 a month.
Update: line separation...