Detecting polymorphic threats

• US 7,739,740 B1
• Filed: 09/22/2005
• Issued: 06/15/2010
• Est. Priority Date: 09/22/2005
• Status: Active Grant

• Current Assignee: CA, Inc. (d/b/a CA Technologies) (Broadcom, Inc.)
• Original Assignee: Symantec Corporation (NortonLifeLock Inc.)
• Inventors: Wilhelm, Jeffrey, Nachenberg, Carey
• Primary Examiner(s): Henning; Matthew T
• Application Number:
  US11/233,195
• Time in Patent Office: 1,727 Days
• Field of Search: 713/188
• US Class Current: 726/25
• CPC Class Codes: G06F 21/566 Dynamic detection, i.e. detection performed at run-time, e.g. emulation, suspicious activities

Abstract

A polymorphic threat manager monitors an incoming email stream, and identifies incoming email messages to which executable files are attached. The polymorphic threat manager characterizes incoming executable files according to at least one metric. For example, the polymorphic threat manager can decompose an executable file into fragments, hash some or all of these, and use the hashes as characterization metrics. The polymorphic threat manager subsequently de-obfuscates executable files, and creates corresponding characterization metrics for the de-obfuscated images. The characterizations of executable files before and after de-obfuscation are compared, and if they differ sufficiently, the polymorphic threat manager determines that the file in question is polymorphic. The characterization metrics of such an executable file after de-obfuscation can be used as a signature for that file.

Claims

1. A computer implemented method for managing polymorphic malicious code, the method comprising the steps of:

using a computer to perform steps comprising;

monitoring an incoming email stream;

identifying an incoming email message to which an executable file is attached;

characterizing the executable file according to at least one metric to produce a before characterization;

de-obfuscating the executable file according to at least one technique, the at least one technique comprising running the executable file in an emulator having a memory and dumping an image of the memory to produce the de-obfuscated executable file;

characterizing the de-obfuscated executable file according to the at least one metric to produce an after characterization;

comparing the before characterization of the executable file with the after characterization of the de-obfuscated executable file; and

determining whether the executable file is polymorphic based on the results of the comparing step, wherein the determining comprises concluding that the executable file is polymorphic responsive to the before characterization being different from the after characterization.

2. The method of claim 1 wherein characterizing the executable file further comprises:

hashing at least a portion of the executable file.

3. The method of claim 2 further comprising:

decomposing the executable file into fragments; and

hashing at least some of the fragments of the executable file.

4. The method of claim 1 further comprising:

comparing the before characterization of the executable file to signatures of known legitimate executable files; and

responsive to the before characterization matching a signature of a known legitimate executable file, not further processing the executable file.

5. The method of claim 1 wherein characterizing the executable file further comprises running the executable file and further performing at least one action from a group of actions consisting of:

tracking instruction usage;

recording a control flow graph of at least one section;

noting a size of at least a part of at least one section;

noting a range of at least a part of at least one section;

noting entropy of at least a part of at least one section;

detecting a transformation of code;

detecting a transformation of data; and

detecting the execution of at least one instruction.

6. The method of claim 1 wherein de-obfuscating the executable file further comprises performing at least one action from a group of actions consisting of:

unpacking the executable file;

examining the memory image dump; and

canonicalizing instructions in the executable file.

7. The method of claim 1 further comprising:

comparing the after characterization of the de-obfuscated executable file to a characterization of a de-obfuscated state of a known malicious polymorphic entity; and

responsive to the after characterization being similar to the characterization of the de-obfuscated state of the known malicious polymorphic entity, determining that the executable file comprises that malicious polymorphic entity.

8. The method of claim 1 further comprising:

storing the after characterization of the de-obfuscated executable file if the file is found to be polymorphic.

9. The method of claim 8 further comprising:

comparing the after characterization of the de-obfuscated executable file to a stored characterization of a stored de-obfuscated executable file known to be polymorphic; and

responsive to the after characterization being similar to the stored characterization of the de-obfuscated executable file known to be polymorphic, concluding that the two executable files comprise different forms of a single polymorphic executable file.

10. The method of claim 1 further comprising:

transmitting to a centralized component a portion of the after characterization of the de-obfuscated executable file if the executable file is found to be polymorphic.

11. The method of claim 1, wherein the at least one technique dumps the image of the memory in response to detecting decryption of the executable file.

12. The method of claim 1, wherein the at least one technique dumps the image of the memory after a fixed amount of time.

13. The method of claim 1, wherein the at least one technique dumps the image of the memory after running the executable file in the emulator removes compression and/or encryption in the executable file.

14. A computer readable storage medium containing an executable computer program product for managing polymorphic malicious code, the computer program product comprising:

program code for monitoring an incoming email stream;

program code for identifying an incoming email message to which an executable file is attached;

program code for characterizing the executable file according to at least one metric to produce a before characterization;

program code for de-obfuscating the executable file according to at least one technique, the at least one technique comprising running the executable file in an emulator having a memory and dumping an image of the memory to produce the de-obfuscated executable file;

program code for characterizing the de-obfuscated executable file according to the at least one metric to produce an after characterization;

program code for comparing the before characterization of the executable file with the after characterization of the de-obfuscated executable; and

program code for determining whether the executable file is polymorphic based on the results of the comparing step, wherein the determining comprises concluding that the file is polymorphic responsive to the before characterization being different from the after characterization.

15. The computer program product of claim 14 wherein the program code for characterizing the executable file further comprises:

program code for hashing at least a portion of the executable file.

16. The computer program product of claim 15 further comprising:

program code for decomposing the executable file into fragments; and

program code for hashing at least some of the fragments of the executable file.

17. The computer program product of claim 14 wherein the program code for characterizing the executable file further comprises program code for running the executable file and further performing at least one action from a group of actions consisting of:

tracking instruction usage;

recording a control flow graph of at least one section;

noting a size of at least a part of at least one section;

noting a range of at least a part of at least one section;

noting entropy of at least a part of at least one section;

detecting a transformation of code;

detecting a transformation of data; and

detecting the execution of at least one instruction.

18. The computer program product of claim 14 wherein the program code for de-obfuscating the executable file further comprises program code for performing at least one action from a group of actions consisting of:

unpacking the executable file;

examining the memory image dump; and

canonicalizing instructions in the executable file.

19. The computer program product of claim 14 further comprising:

program code for comparing the after characterization of the de-obfuscated executable file to a characterization of a de-obfuscated state of a known malicious polymorphic entity; and

program code for, responsive to the after characterization being similar to the characterization of the de-obfuscated state of the known malicious polymorphic entity, determining that the executable file comprises that malicious polymorphic entity.

20. The computer program product of claim 14 further comprising:

program code for storing the after characterization of the de-obfuscated executable file if the file is found to be polymorphic.

21. A computer system for managing polymorphic malicious code, the computer system product comprising:

a computer readable storage medium storing executable software portions comprising;

a software portion configured to monitor an incoming email stream;

a software portion configured to identify an incoming email message to which an executable file is attached;

a software portion configured to characterize the executable file according to at least one metric to produce a before characteristic;

a software portion configured to de-obfuscate the executable file according to at least one technique, the at least one technique comprising running the executable file in an emulator having a memory and dumping an image of the memory to produce the de-obfuscated executable file;

a software portion configured to characterize the de-obfuscated executable file according to the at least one metric to produce an after characterization;

a software portion configured to compare the before characterization of the executable file with the after characterization of the de-obfuscated executable file; and

a software portion configured to determine whether the executable file is polymorphic based on the results of the comparing step, wherein the determining comprises concluding that the executable file is polymorphic responsive to the before characterization being different from the after characterization.