Ransomware infection detection in filesystems
Described herein is a system that detects ransomware infection in filesystems. The system detects ransomware infection by using backup data of machines. The system detects ransomware infection in two stages. In the first stage, the system analyzes a filesystem's behavior. The filesystem's behavior can be obtained by loading the backup data and crawling the filesystem to create a filesystem metadata including information about file operations during a time interval. The filesystem determines a pattern of the file operations and compares the pattern to a normal patter to analyze the filesystem's behavior. If the filesystem's behavior is abnormal, the system proceeds to the second stage to analyze the content of the files to look for signs of encryption in the filesystem. The system combines the analysis of both stages to determine whether the filesystem is infected by ransomware.
1. A method for detecting ransomware infection in filesystems, comprising:
accessing a first set of backup data corresponding to a machine, the first set of backup data indicating changes in a filesystem of the machine during one or more time intervals, the changes including file operations to at least one file in the filesystem;
maintaining a ransomware detection model operative to detect an infection of the filesystem by ransomware, wherein the ransomware detection model uses at least one first model feature that is based at least in part on the accessed first set of backup data associated with the machine including the changes and file operations over the one or more time intervals, and wherein the ransomware detection model further uses a second model feature that is based at least in part on respective entropies of the at least one file in the filesystem;
receiving, by the ransomware detection model, a second set of backup data corresponding to the machine, the second set of backup data indicating second changes in the filesystem of the machine during a second time interval subsequent to the one or more time intervals, the second changes including file operations to one or more files in the filesystem;
applying the second set of backup data to the ransomware detection model to screen for abnormal behavior of the filesystem, wherein the ransomware detection model is operative to assess the second changes and entropies of the one or more files in the filesystem associated with the second time interval, wherein the ransomware detection model is further operative to calculate a set of abnormal scores indicating a respective probability that the filesystem behaves abnormally and to detect the abnormal behavior based at least in part on a combined abnormal score of the set of abnormal scores;
based at least in part on the ransomware detection model detecting the abnormal behavior, generating a user interface for presenting a determination that the filesystem is infected; and
providing the user interface for display.
2. The method of claim 1 wherein the machine is a virtual machine.
3. The method of claim 2 wherein the first and second sets of backup data are snapshots of the virtual machine.
4. The method of claim 1 further comprising:
developing the ransomware detection model at least in part by training one or more machine learning models using training data, the training data including at least one of filesystem data of filesystems owned by multiple users, filesystem data of filesystems owned by a user, and filesystem data of a filesystem owned by the user.
5. The method of claim 1 further comprising:
developing the ransomware detection model at least in part by training one or more machine learning models using training data, the training data including filesystem data of the machine.
6. The method of claim 1 wherein the changes in the filesystem, the second changes in the filesystem, or both are recorded in filesystem metadata, the filesystem metadata including a list of entries corresponding to data changes in the filesystem.
7. The method of claim 1 wherein the ransomware detection model is operative to apply a set of machine learning models to the changes, the second changes, or both, the set of machine learning models trained to determine a pattern of the file operations and to compare the pattern of the file operations to a normal pattern of file operations based on features representing a behavior of the filesystem.
8. The method of claim 1 wherein the ransomware detection model is operative to determine whether files in the filesystem are encrypted by extracting entropy features from the files and calculating an encryption score representing a likelihood of the files being encrypted using the entropy features, the entropy features measuring randomness of the files.
9. A non-transitory computer-readable storage medium storing computer program instructions executable by a processor, the computer program instructions configured to cause the processor to perform:
accessing a first set of backup data corresponding to a machine, the first set of backup data indicating changes in a filesystem of the machine during one or more time intervals, the changes including file operations to at least one file in the filesystem;
maintaining a ransomware detection model operative to detect an infection of the filesystem by ransomware, wherein the ransomware detection model uses at least one model feature that is based at least in part on the accessed first set of backup data associated with the machine including the changes and file operations over the one or more time intervals, and wherein the ransomware detection model further uses a second model feature that is based at least in part on respective entropies of the at least one file in the filesystem;
receiving, by the ransomware detection model, a second set of backup data corresponding to the machine, the second set of backup data indicating second changes in the filesystem of the machine during a second time interval subsequent to the one or more time intervals, the second changes including file operations to one or more files in the filesystem;
applying the second set of backup data to the ransomware detection model to screen for abnormal behavior of the filesystem, wherein the ransomware detection model is operative to assess the second changes and entropies of the one or more files in the filesystem associated with the second time interval, wherein the ransomware detection model is further operative to calculate a set of abnormal scores indicating a respective probability that the filesystem behaves abnormally and to detect the abnormal behavior based at least in part on a combined abnormal score of the set of abnormal scores;
based at least in part on the ransomware detection model detecting the abnormal behavior, generating a user interface for presenting a determination that the filesystem is infected; and
providing the user interface for display.
10. The non-transitory computer-readable storage medium of claim 9 wherein the machine is a virtual machine.
11. The non-transitory computer-readable storage medium of claim 10 wherein the first and second sets of backup data are snapshots of the virtual machine.
12. The non-transitory computer-readable storage medium of claim 9 wherein the computer program instructions are configured to cause the processor to perform:
developing the ransomware detection model at least in part by training one or more machine learning models using training data, the training data including at least one of filesystem data of filesystems owned by multiple users, filesystem data of filesystems owned by a user, and filesystem data of a filesystem owned by the user.
13. The non-transitory computer-readable storage medium of claim 9 wherein the computer program instructions are configured to cause the processor to perform:
developing the ransomware detection model at least in part by training one or more machine learning models using training data, the training data including filesystem data of the machine.
14. The non-transitory computer-readable storage medium of claim 9 wherein the computer program instructions are configured to cause the processor to record the changes in the filesystem, the second changes in the filesystem, or both, in filesystem metadata, the filesystem metadata including a list of entries corresponding to data changes in the filesystem.
15. The non-transitory computer-readable storage medium of claim 9 wherein the ransomware detection model is operative to apply a set of machine learning models to the changes, the second changes, or both, the set of machine learning models trained to determine a pattern of the file operations and to compare the pattern of the file operations to a normal pattern of file operations based on features representing a behavior of the filesystem.
16. The non-transitory computer-readable storage medium of claim 15 wherein the features include a ratio of modify operations to file operations and a modify operation is a type of the file operations, and wherein detecting the abnormal behavior comprises extracting the features from the changes.
17. The non-transitory computer-readable storage medium of claim 15 , wherein each machine learning model of the set of machine learning models is configured to calculate an abnormal score of the set of abnormal scores.
18. The non-transitory computer-readable storage medium of claim 17 , wherein detecting the abnormal behavior comprises generating the combined abnormal score by combining the abnormal scores calculated by the set of machine learning models.
19. The non-transitory computer-readable storage medium of claim 9 wherein the ransomware detection model is operative to determine whether files in the filesystem are encrypted by extracting entropy features from the files and calculating an encryption score representing a likelihood of the files being encrypted using the entropy features, the entropy features measuring randomness of the files.
20. The non-transitory computer-readable storage medium of claim 9 wherein the ransomware detection model is operative to assess the entropies of the one or more files of the filesystem responsive to the detection of the abnormal behavior.