That's neither true, nor is the guaranteed security you profess absolute with your "are not really that accessible due to MD5." Besides the countless security breaches where the passwords were unencrypted, there are many other ways in which passwords are distributed to and from the user in an insecure manner (http://plaintextoffenders.com).
But even if you've verified that Yahoo! has followed every standards guidelines to utmost, you still can't be sure that your password (all passwords) weren't intercepted as encrypted packets and then unencrypted. For example, had you been piggybacking on a network to capture huge dumps of data using Wireshark for years, only to find that the juice stuff was encrypted with SSL, you could go back and unencrypted everything from Apple devices for many years do to Goto Fail. For that reason alone, you should've systematically changed all your internet-facing passwords.
Soli said ...there are many other ways in which passwords are distributed to and from the user in an insecure manner...
Sure but that has nothing to do with the hosting site. If you go to a coffee shop that has an unencrypted guest WiFi account, then you are responsible for your password being sent in clear text. Have you ever tried to unencrypted an MD5 password? Almost impossible even with a super computer. And that would be case by case situation, not a massive breech.
volcan said: If you go to a coffee shop that has an unencrypted guest WiFi account, then you are responsible for your password being sent in clear text.
You're not following if you think I mentioned anything about passwords "being sent in clear text."
volcan said: How exactly do you believe hackers decrypt MD5 passwords? I want to know because that is what I use.
How exactly is that a serious question? Just like the aforementioned Goto Fail bug (as well as countless others not mentioned) you can't be certain that any security measures you've taken are 100%, guaranteed, I never have to change a password ever again, no matter what, all my data is safe for eternity. It sounds like you're not even aware of Goto Fail, which I brought up specifically because it's the most common of the major security gaffes affecting Apple that has hit the news in force in the last couple years.
If you think you never have to change a password and can have the same password for every internet-facing account because MD5 has your back. Fine, but please don't try to convince others to follow your false idol.
I'm not saying that it is a good idea to use the same password for every site, only that hackers cannot crack an MD5 password. An MD5 password contains 32 characters, 128 bits. Which means that any password more than 32 characters is only evaluated up to that character limit. If a hacker gains access to the database they can copy the hashed password. Then if they hack into another site and gain access to that database they can paste that hash into the password field, but they still don't know what the actual password is so it really does them no good.
As I mentioned, the passwords are not the main target. It is the underlying free text information in the database such as names and addresses that they are after.
You're not following if you think I mentioned anything about passwords "being sent in clear text."
An assumption on my part. How exactly do you believe hackers decrypt MD5 passwords? I want to know because that is what I use.
MD5 has been unsafe for quite a few years now.
What you're doing when you MD5 a string is perform a set of mathematical transforms on it such that it will always generate the same output given the same input; it should be impossible to actually reverse. In the case of a MD5, the output will always be 32 characters long. You don't actually ever 'decrypt' a MD5, it's what we call that a one-way hash. There are many implementations of one-way hashes, md5 is far from the only game in town and is kind of old at this point.
What happens when you enter your password into a website login, for example, is they take what you typed, and then they re-run the hashing algorithm (md5) on what you typed in. Normally they also modify in some way and add additional data to what you typed in order to protect it in the case of a leak/dump/breach. That's called 'salting'. In 2016 still not everyone does this, and that's a mess.
What hackers do when they get their hands on a login database is create what are known as 'collisions'. They throw gigantic datasets against the hash they have of your password in an attempt to come up with a string - any string - that ends up creating matching output. That obviously wasn't supposed to be possible when the hashing algorithm was first conceived, but security researchers are constantly poking and prodding at these things to enure they are actually secure. When they find a collision they then use that to log into whatever site you used that same password on. Doesn't matter if it isn't your exact password, when its run back through the hashing algorithm it will be recognized as such.
You're not following if you think I mentioned anything about passwords "being sent in clear text."
An assumption on my part. How exactly do you believe hackers decrypt MD5 passwords? I want to know because that is what I use.
MD5 has been unsafe for quite a few years now.
Yeah well three failed log ins and you are locked out.
If you have a dump of the data, you don't need to try and brute force a login. You create the collision first with the offline copy and you login first try ; )
Good to know. So how long do you think it takes to crack so many accounts using your techniques?
This gets technical. Depends heavily on the resources of the person doing the cracking. GPU accelerated cracking has moved a lot of this stuff from "that would take forever" to "holy crap, everything is broken now."
I suggest taking a look athttps://en.wikipedia.org/wiki/MD5#Security if you want to know more info about the security pitfalls MD5 specifically.There are ways to iterate over the passwords faster than a brute force attack too, some of which could potentially reveal the actual password itself -- not by reversing the MD5 mind you, but by working with extremely large sets of data known as "rainbow tables".
Basically if you know the password is MD5 hashed, and it's not properly salted, you can remove the need to brute force anything by simply comparing the stolen hashed output against a precomputed set of outputs. If i spend the time to figure out what most common passwords would be when run through the algorithm, I can skip doing any potentially expensive computations. This is what is known as a "rainbow table". You can do a whole lot of those comparisons every second on modern hardware.
Essentially, If your password gets hashed to 1234567890123456798012 and I already know the possible inputs that could return that output, or (often) just the most likely inputs, then I can use that knowledge to know what your password actually is.
Isnt that billions of computer hours for slim possibility that they would find something useful?
'Useful' means different things to different people. They're interested in things like your contact list so they can add them to their spam lists (either for personal use or resale), credit card numbers, security question information, etc. Once (in the case of Yahoo Mail) they get a hold of an email account, they can use it to try to reset your password on other, potentially more valuable targets. If, for example, that same email was used to log into your Wells Fargo bank account and your security question is your mother's maiden name... there's probably a good chance they can figure that out. That can be a real problem.
Also, none of those companies (like Wells Fargo) really know what you sound like on the phone, so it becomes very easy to use social engineering to trick them to do stuff that is outside of normal policy. You can appear to be an authentic person and even 'prove' you are them by sending an email from their account.
If i spend the time to figure out what most common passwords would be when run through the algorithm, I can skip doing any potentially expensive computations. This is what is known as a "rainbow table". You can do a whole lot of those comparisons every second on modern hardware.
Ha! interesting. I've been using a lot of different passwords that are easy to remember yet very complex like "i95@exit26" as an example. Just a location. Haven't used a password manager app yet.
Comments
But even if you've verified that Yahoo! has followed every standards guidelines to utmost, you still can't be sure that your password (all passwords) weren't intercepted as encrypted packets and then unencrypted. For example, had you been piggybacking on a network to capture huge dumps of data using Wireshark for years, only to find that the juice stuff was encrypted with SSL, you could go back and unencrypted everything from Apple devices for many years do to Goto Fail. For that reason alone, you should've systematically changed all your internet-facing passwords.
If you think you never have to change a password and can have the same password for every internet-facing account because MD5 has your back. Fine, but please don't try to convince others to follow your false idol.
As I mentioned, the passwords are not the main target. It is the underlying free text information in the database such as names and addresses that they are after.
What you're doing when you MD5 a string is perform a set of mathematical transforms on it such that it will always generate the same output given the same input; it should be impossible to actually reverse. In the case of a MD5, the output will always be 32 characters long. You don't actually ever 'decrypt' a MD5, it's what we call that a one-way hash. There are many implementations of one-way hashes, md5 is far from the only game in town and is kind of old at this point.
What happens when you enter your password into a website login, for example, is they take what you typed, and then they re-run the hashing algorithm (md5) on what you typed in. Normally they also modify in some way and add additional data to what you typed in order to protect it in the case of a leak/dump/breach. That's called 'salting'. In 2016 still not everyone does this, and that's a mess.
What hackers do when they get their hands on a login database is create what are known as 'collisions'. They throw gigantic datasets against the hash they have of your password in an attempt to come up with a string - any string - that ends up creating matching output. That obviously wasn't supposed to be possible when the hashing algorithm was first conceived, but security researchers are constantly poking and prodding at these things to enure they are actually secure. When they find a collision they then use that to log into whatever site you used that same password on. Doesn't matter if it isn't your exact password, when its run back through the hashing algorithm it will be recognized as such.
Hope that helps.
Isnt that billions of computer hours for slim possibility that they would find something useful?
I suggest taking a look at
Basically if you know the password is MD5 hashed, and it's not properly salted, you can remove the need to brute force anything by simply comparing the stolen hashed output against a precomputed set of outputs. If i spend the time to figure out what most common passwords would be when run through the algorithm, I can skip doing any potentially expensive computations. This is what is known as a "rainbow table". You can do a whole lot of those comparisons every second on modern hardware.
Essentially, If your password gets hashed to 1234567890123456798012 and I already know the possible inputs that could return that output, or (often) just the most likely inputs, then I can use that knowledge to know what your password actually is.
'Useful' means different things to different people. They're interested in things like your contact list so they can add them to their spam lists (either for personal use or resale), credit card numbers, security question information, etc. Once (in the case of Yahoo Mail) they get a hold of an email account, they can use it to try to reset your password on other, potentially more valuable targets. If, for example, that same email was used to log into your Wells Fargo bank account and your security question is your mother's maiden name... there's probably a good chance they can figure that out. That can be a real problem.
Also, none of those companies (like Wells Fargo) really know what you sound like on the phone, so it becomes very easy to use social engineering to trick them to do stuff that is outside of normal policy. You can appear to be an authentic person and even 'prove' you are them by sending an email from their account.
TLDR; Damn Internet, You scary.