Take the 2-minute tour ×
Information Security Stack Exchange is a question and answer site for Information security professionals. It's 100% free, no registration required.

Most users tipically use the same password for multiple applications. Let's say all of these applications hash the password in some way. Would it be easier for an attacker to get the original password if it got access to all the different hashes? Of course the attacker probably looking at the hashes may know which is the weaker one and only try to bruteforce this, but has he gained anything having access to other hashes of the same password? Also assuming the attacker does not have any rainbow table containing any of these hashes.

So, if an attacker has access to all these hashes of the same password, can he be faster in breaking it than just try to bruteforce one of the md5 hashes (as in this case is the weaker link)?

  • 33c38d9db9759da0d0a813968bc3c1fb
  • 7b8758cf3cb6e540d60b3e9999c48fbef0502ac6
  • faa81286aa90907253a58acfd83d2d89e58120031630d60dc05d0cd365b6df8d
  • $1$NPeNiNCA$AwpOWKuur2LZvXiJxNM6U1
  • $5$VHVZwDnI4aAEUG2m$ZhSJ537JmYJ5ISDxPQ6doWxPOrz9AMjyvxYxecViK23
  • $5$3bRXBWCpG9GEHdm$a2xIBbgK.fyr5k26BhUGo2QCvl25yQU1I9mwFp7mmh5p
  • $6$Hcdqud1hDgz$5MOEbgtOdxfF289OSXrmIevt7NnZBLaQkNGKeijR/X09ZDEbQ/ZfObJjNo0J64t4haSSqihRhfPe4z8l.ptro1
  • $6$hqaQ3RzX$bk8dgMsAlmDPpOX0IWWIrJ9T3awblvI.PCipeqJDdJZSVDQJgCRsQRX8pKpFU2XNcvRr56e3MARcTQR5oJ94V/
  • $2a$13$Hdz8T8vlEqwCvhixyu4rlel2cjj.TfA1qXEZ2dXhabogN35Idd8Je
  • $2a$15$hLPoYhbVJNA48A/Wmv1I5.5XBv/G/1s8BGfDLU7mt37ojGhNjETd2

btw, the original password is hello12345world12345, I'm not trying to (make you) crack someone's password ;-)

share|improve this question
1  
Yes, it's as weak as the fastest / weakest hash function used. –  Polynomial Jan 23 '13 at 10:05

2 Answers 2

up vote 5 down vote accepted

In theory, having several hashes may help; in practice, not so much.

The Theory

Let's consider SHA-256. From this function, we can define another hash function, which we will call SHB-256, such that SHB-256(m) = SHA-256(m) XOR m. In simple words, I compute the XOR of the SHA-256 and the first 256 bits of m.

This "SHB-256" is as good a cryptographic hash function as SHA-256: it is similarly resistant to collisions, preimages and second preimages. However, if I give you both SHA-256(m) and SHB-256(m), then a simple XOR of both value will yield the first 256 bits of m.

The Practice

Taking advantage of several hashes computed with the same password then relies on how "similar" the hash functions are, so as to allow "canceling" each other. This is a very ill-defined property, and (as yet) not thoroughly investigated. To my knowledge, no such shortcut is known with existing hash functions. There are a lot of details to mind; iterated functions and usage of salts make the matter even more complex.

What a realistic attacker can do when faced with multiple hashes is to use the one which will be easiest for him to attack, and ignore the others. If one of them is unsalted, then the attacker will target that one, which allows for simultaneous attack of several hashed passwords and usage of precomputed (rainbow) tables. Otherwise, he will use the function which will be most efficient when implemented on his hardware (e.g. if faced with PBKDF2 and Bcrypt, the attacker will target PBKDF2, which is easier to implement on a GPU). As @Polynomial puts it, your hashes will be as strong as the weakest of them.

An extra indirect effect is about your CPU budget. When doing registration, you will hash the user's password with all the hash functions. For the hash functions which are "configurably slow", like bcrypt or PBKDF2, this may incite you to configure them to be "a bit faster" than what you would have done if you had used only one, so as to keep your own costs within tolerable limits. This correspondingly lowers the attacker's efforts, since he has just to crack one hashes, not all of them.

share|improve this answer
    
I think the OP's scenarion was more, that a user reused the same password on several sites, and the attacker got access to those different hashes (not calculating several hashes on the same CPU). But your first part answers the question i think. –  martinstoeckli Jan 23 '13 at 11:58
    
@martinstoeckli right, like reusing the same password on multiple applications in your company, where your sysadmin might be your attacker –  Carlos Campderrós Jan 23 '13 at 13:50

It depends on your password strength and hashing algorithm used for hashing. There are programs uses CPU + GPU to break the hashes using brute force attacks. So if you use common password its easy to crack by using a PC and a graphics card. There are lot of applications available which uses the power of GPU along with CPU. It's not matter if you give two or three types of hashes of a password, the brute force timing is depends on password strength and the hashing algorithm.

If I have 5 different hashes of your password and one of them is very weak, then that password can be recovered quickly by attacking the hash of weakest hashing algorithm.

share|improve this answer
1  
If the three types of hashes differ in strength, then the weakest hash algorithm will prove easier to crack, no matter the strength of the password. –  Kao Jan 23 '13 at 11:49
    
@Kao he spoke not about simple hashes. When you think about hashes don't always assume weak hashes. Consider strong hashes like SHA-1, SHA-256, SHA-512, MD5 etc. Any way I updated my answer with few modification and thank for this suggestion. –  sujeesh Jan 23 '13 at 11:54
2  
@sujeesh - Actually the OP knows that one can brute-force the weakest hash value. I think his question was, whether combining the different hashes may give an advantage. –  martinstoeckli Jan 23 '13 at 12:36
    
@martinstoeckli my answer will be no. a small change in the text will result a very huge change in its hash. so combining hashes won't help. –  sujeesh Jan 23 '13 at 12:38
    
@sujeesh What is a weak hash function in your opinion? –  Henning Klevjer Jan 23 '13 at 12:57

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.