Similarity hash for android executables
A method of generating a similarity hash for an executable includes extracting a plurality of characteristics for one or more classes in the executable, and transforming the plurality of characteristics into a set of one or more class fingerprint strings corresponding to the one or more classes. The set of class fingerprint strings is transformed into a hash string using minwise hashing, such that a difference between hash strings for different executables is representative of the degree of difference between the executables. The hash of a target executable is compared with hashes of known malicious executables to determine whether the target executable is likely malicious.
1. A method of generating a similarity hash for an executable, comprising:
extracting a plurality of characteristics for one or more classes in the executable, wherein the class characteristics are derived from class definitions corresponding to classes in the executable;
transforming the plurality of characteristics into a set of one or more class fingerprint strings wherein each of the one or more class fingerprint strings corresponds to one of the one or more classes; and
transforming the set of class fingerprint strings into a hash string using minwise hashing, such that a difference between hash strings for different executables is representative of the degree of difference between the executables.
2. The method of generating a similarity hash for an executable of claim 1 , wherein the class characteristics are derived from class definition data comprising one or more of access flags, superclass name, implemented interfaces, annotations, class data, and static values.
3. The method of generating a similarity hash for an executable of claim 2 , wherein the one or more fingerprint strings comprises one or more string characters representing each of the one or more class definition data types.
4. The method of generating a similarity hash for an executable of claim 1 , wherein the executable comprises a Dalvik (DEX) executable.
5. The method of generating a similarity hash for an executable of claim 1 , wherein transforming the set of class fingerprint strings into a hash string using minwise hashing comprises using weighted b-bit minwise hashing.
6. The method of generating a similarity hash for an executable of claim 1 , wherein the executable is a Java executable.
7. The method of generating a similarity hash for an executable of claim 1 , further comprising comparing the hash string with a second hash string from a second executable to determine the similarity between the executable and the second executable.
8. The method of generating a similarity hash for an executable of claim 7 , wherein comparing the hash string with the hash string from a second executable comprises calculating a Hamming distance between the hash string and the second hash string.
9. The method of generating a similarity hash for an executable of claim 7 , further comprising determining whether the executable and the second executable are likely related based on the determined similarity between the executable and the second executable.
10. A computerized security device, comprising:
a processor and a memory;
a database of hashes derived from and representing one or more known malicious executables; and
an executable malware detection module comprising instructions operable when executed on the processor to extract a plurality of characteristics for one or more classes in a target executable, wherein the class characteristics are derived from class definitions corresponding to classes in the executable, transform the plurality of characteristics into a set of one or more class fingerprint strings wherein each of the one or more class fingerprint strings corresponds to one of the one or more classes, and transform the set of class fingerprint strings into a hash string using minwise hashing such that a difference between hash strings for the target executable and the hashes in the database represent a degree of difference between the target executable and the executables represented by the database hashes.
11. The computerized security device of claim 10 , wherein the class characteristics are derived from class definition data comprising one or more of access flags, superclass name, implemented interfaces, annotations, class data, and static values.
12. The computerized security device of claim 11 , wherein the one or more fingerprint strings comprises one or more string characters representing each of the one or more class definition data types.
13. The computerized security device of claim 10 , wherein the executable comprises at least one of a Dalvik (DEX) executable.
14. The computerized security device of claim 10 , wherein transforming the set of class fingerprint strings into a hash string using minwise hashing comprises using weighted b-bit minwise hashing.
15. The computerized security device of claim 10 , wherein the executable is a Java executable.
16. The computerized security device of claim 10 , the executable malware detection module further operable to compare the hash string with a second hash string from a second executable to determine the similarity between the target executable and the executables represented by the hash strings in the database representing known malicious executables.
17. The computerized security device of claim 16 , wherein comparing the hash string with the hash string from a second executable comprises calculating a Hamming distance between the hash string and the second hash string.
18. The computerized security device of claim 16 , the executable malware detection module further operable to determine whether the executable and the second executable are likely related based on the determined similarity between the target executable and the executables represented by the hashes in the database representing known malicious executables.