Yahoo Sued For Password Breach

twoheadedboy writes "Yahoo is being sued by one of its users, who has claimed the US Internet company was guilty of negligence when 450,000 passwords of the members of the Yahoo Voices blogging community were posted online. Jeff Allan from New Hampshire has turned to a federal court in San Jose, California, after his eBay account, which used the same password as his Voices account, was compromised. The breach at Yahoo followed similar hits on LinkedIn and Nvidia, which together saw millions of passwords leaked."

Comments

Guilty of Negligence

[O'Krap]

One could say that reusing a password is negligent....

Re:Guilty of Negligence

[Anonymous Coward]

But then one would be forced to be a complete idiot who implicitly stated that passwords were a good measure and that people have good enough memories and enough time on their hands to manage one unique strong password for every website they visit.

Luckily one wouldn't say that. (maybe you would though.)

Re:Guilty of Negligence

[icebike]

Its his accounts that are at risk. His choice to take the risk. Not Yahoo's choice. See the difference?

Re:Guilty of Negligence

[Geeky]

Regardless of whether passwords are a good measure, I do use a unique strong password for every important site I visit - i.e. ones that store personal or financial information. Not so bothered with forum logins and the like where it really doesn't matter all that much if they're compromised.

I only remember one password, though, and that's the one to my password database that's stored locally on my PC. I use KeePass, but there are plenty of other password safe applications.

Liability

[Malf.me]

Regrettably a liability lawsuit like this seems to be one of the only tools available to encourage large organizations to take computer security seriously.

TRWTF

[Anonymous Coward]

On the other hand, neither service X nor service Y should be storing your passwords in such a way that it is possible to recover the actual password.

Re:TRWTF

[The Mighty Buzzard]

On the other hand, a hell of a lot of services limit password length for some insane reason. Given the choice, I never use a password under 30 characters in length but there are sites I use that limit me to as little as eight characters. Nearly any (over 95%) possible password of eight characters or less can be looked up in a rainbow table in less than an hour by a single computer. With distributed rainbow table generation today, counting on hashing functions to be one way is rapidly becoming a thing of the past.

Re:TRWTF

[icebike]

Salted passwords don't matter - you can recover the password. Heck, you can reverse engineer hashing algorithms by just making a bunch of passwords then recovering them.

That would require you not only steal the password hash file but also the software used to create that file, including the salt, etc.

The point in the current case is that the passwords WERE NOT stored encrypted in any form. They were stored in clear text despite every recommendation never to do this on any system. Its inexcusable.

Every Linux distribution since the Pleistocene has defaulted to at least a minimally encrypted password file. Yahoo runs nothing but Linux. They would have had to intentionally bypass Linux security basics and roll their own to end up in such a mess.

They deserve to be sued. Still it will be a hard case to win because there is no law that says they have to be careful or competent.

Re:TRWTF

[bzipitidoo]

What's really fun are those services that let you enter a 30 character password, then silently truncate it to 8 characters.

Also thrilling when a service is able to tell you what your forgotten password is.

Then there was the login web page that would let you start typing in your credentials before it was finished loading, then move the cursor back to the username input box when it finished loading. I recall Yahoo's webmail did that for a while. Actually, it was a combination of bad design on both the web page and the browser's part. On those occasions when it was a bit slow to load, you'd be typing in your password just as the cursor warped back to the username box. If you hit enter before noticing what had happened, the overly helpful, advanced browser would remember your username...

Another fun one was the service that insisted on selectively applying capitalization rules to usernames. When I created my account, I used all lower case for my username, and it silently changed the first letter to upper case. But it didn't change the case when I attempted to log in. Took me a while to figure out why I couldn't log in...

Re:TRWTF

[loxosceles]

"It is always possible to recover a password."

This is not true. If a password has more entropy than the hash being used, there will be collisions that make it impossible to tell what the original password is.

This is a basic consequence of the fact that hash functions are irreversible and have fixed size. If you consider the space of all passwords of any length, there are infinitely many passwords (even if you limit passwords to those made of long strings of english words) that hash to a particular value.

For the vast majority of passwords in use, the entropy is lower than the entropy of the hash, so it's feasible to construct mappings of possible passwords to hashes and determine the most likely password that way. It is not *always* possible to recover a password, however.

Re:TRWTF

[Kidbro]

Salted passwords don't matter - you can recover the password. Heck, you can reverse engineer hashing algorithms by just making a bunch of passwords then recovering them.

That would require you not only steal the password hash file but also the software used to create that file, including the salt, etc.

No. The context of this subthread was that using the same password on two separate systems would give the owners of one of the systems access to the other. Presumably, they won't need to "steal" anything, as they already control both the relevant data and software.

Yahoo runs nothing but Linux

If the only boxes they have are forward pointing webservers, then this link is relevant. That is obviously not the case.

They would have had to intentionally bypass Linux security basics and roll their own to end up in such a mess.

Intentionally bypass? Please. Of course you don't create operating system level users for users in your web app. They are way too much overhead for that. Of course they (and everybody else that provide web services of some sort) "roll their own" user system.

Storing passwords in plain text is still inexcusable of course, but the rest of your post is nonsense.

Re:TRWTF

[Tom]

It is *always possible to recover* a password.

No, it is not. You need go back to Cryptography 101.

A properly seeded hash using a proper cryptographic one-way hash function is impossible to revert using todays and any technology within the forseable future. It's not a matter of raising CPU powers by a few orders of magnitude, but by a couple billion orders of magnitude.

Heck, you can reverse engineer hashing algorithms by just making a bunch of passwords then recovering them.

The proper term is rainbow tables, and they don't work for good salts because you need one table per possible salt value, meaning with 2 bytes you need 65k rainbow tables. If that frightens you, use 4 bytes for the salt and you need 4 billion rainbow tables. Good luck.

Public-key crypto has been around how long?

Someone has eaten a buzzword. :-)

Public-key crypto is not any more (nor less) secure than one-way hashing. Properly applying the appropriate crypto is what matters.

Bottom line, reusing a password is negligent on your own part.

Welcome to the real world, where people have dozens of accounts. Since we request password policies that are insane from the human memory perspective POV, very few people can reasonably be expected to remember more than 2 or 3 of those Sc8OOlB1hu0Bj%7 beasts you have the choice between re-use or storing it somewhere (paper or password manager). Both of these approaches have their own security issues.

disclaimer: I am an IT security professional.

Re:TRWTF

[Tom]

You really need to go back to some basics.

The strength of good cryptography lies exactly in that not one but many men, and not just any but the top experts in the field have been trying to break it - and have failed. A crypto algorithm is considered strong exactly if there are no known attacks against it that are significantly faster than brute-force, despite said experts looking for one. All the major ciphers in use today have withstood at least one, usually several decades of attempts to break them.

Is it theoretically imaginable that tomorrow, someone will publish something that drops the strength of a major cipher by at least a thousand orders of magnitude? Well, it is not impossible in the sense that FTL travel is, but it is about as unlikely as the existence of the famous teapot in orbit around Jupiter.

Nobody in the academic field of cryptography assumes humand infallability. Like any academic field, new discoveries will be made. However, just like in physics we will certainly improve upon, say, the theory of gravity, and modify the formulas to be more precise, but nobody sane gives a significantly different from zero probability to a complete falsification of gravity these days.

Likewise, there will certainly be more weaknesses found in say, the SHA family (we already know a few), which will provide attacks that are ten or even a hundred times faster than brute-force. SHA-1, for example, is considered broken by the cryptographic community. That is mathematically broken - you can crack it in a billion years instead of the trillion years that a brute-force would require. That is the dimensions we're talking about. Find another weakness in SHA-1 and make it a thousand times weaker, and you are still far beyond anything that will matter in our lifetimes.

You are assuming that there is a real "crack", a crucial weakness that will enable a practical attack, instead of a theoretical. Those things happen - Deep Crack - but they don't happen suddenly, they don't happen unexpected, and by the time you and I hear about them, the people who actually work in cryptography think of it as yesterday's news.

We already have SHA-2 which you can use if you don't think SHA-1 is secure enough. And we're working on SHA-3. It's not like encryption would stand still while people attack it, you know?

man can break it.

Check this history of SHA-3, and how many candidates were rejected for which reasons. That should give you some perspective - and humility.

Image of Trust

[Penurious Penguin]

Because Yahoo and other similar services pimp the image of being both sophisticated and virtually omnipotent, while offering to manage your affairs, organize your life, provide targeted news headlines and personal suggestions regarding your personal life, and then covertly subpimp your personal data while indifferently and deeply mining your grazing habits -- I think this lawsuit is, compared to others, reasonable, if a lawsuit without grievous injuries or loss can even be so.

Not everyone has a degree in IT. Perhaps instead of guerrilla advertisement, Yahoo (and other similar services) could cough up at least a token effort for their cattle, I mean customers. Maybe they could reserve some extra ad-space to discourage unknowing subjects from having shared passwords. Maybe they could do a lot more in general, and a lot less too, in a good way.

I sympathize with neither side in this case, but can empathize with only one. Altruism, despite modern Goliaths, doesn't always need an ulterior motive. Yahoo preys on the sea of humanity, and a few minnows nip back. Pardon me whilst I desiccate myself with tears.

I'm just a simple caveman...

[gubon13]

I'm sure there will be many valid points made about how utterly irresponsible it is to use the same username and password on both your email account and "financial" sites, and also about how terrible Yahoo! is for making very little effort to make amends to their customers whose privacy they clearly don't value... And sure, this lawsuit will likely tie up the court system and waste thousands, possibly millions of taxpayer dollars... And maybe Yahoo! will ultimately be required to send each of those 450,000 users a $10 voucher for food in the Yahoo! cafeteria... ...but I think we're all missing the elephant in the room here - Yahoo! is still around?