Automation and optimization of data recovery after a ransomware attack
In the face of ransomware attacks, which can be increasingly difficult to effectively prevent, a solution can be considered to be the minimization of the cost and time taken to recover data and, hence business activities. Embodiments perform a restore operation that include automatically identifying the most recent healthy backup, from which data should be restored, and the prioritizing of the order in which data should be restored.
1. A method comprising:
identifying a plurality of candidate backups from backups in a data storage system;
extracting characteristics from each of the candidate backups;
automatically analyzing the characteristics of the candidate backups to identify which of the candidate backups are infected with an infection, wherein analyzing the characteristics of the candidate backups includes comparing the characteristics of each of the plurality of candidate backups with each other;
assigning a classification for each of the plurality of candidate backups based on the analyzed characteristics;
automatically identifying a healthy backup from the plurality of candidate backups that does not include the infection for performing a restore operation based on the classification assigned to each of the candidate backups; and
performing the restore operation by restoring production data from the healthy backup to storage devices, wherein the restore operation includes restoring a first portion within the production data prior to a second portion within the production data, wherein the first portion within the production data has a higher priority than the second portion within the production data.
2. The method of claim 1 , wherein the candidate backups include full backups, incremental backups, or snapshots and wherein the infection is at least one of ransomware, a virus, a worm, malware, or combination thereof.
3. The method of claim 1 , wherein the characteristics of each of the candidate backups include file change information, file metadata, timestamps, and file extensions.
4. The method of claim 1 , further comprising mining the candidate backups to identify patterns indicative of the infection or using features associated with infected peers to identify the infection.
5. The method of claim 4 , wherein the patterns which indicate the infection include an unexpected presence of a file in a specific location.
6. The method of claim 1 , further comprising using data augmentation to identify the healthy backup or the infection.
7. The method of claim 6 , wherein the data augmentation includes historical data obtained from a historical database, change rates for files, typical file sizes, file extensions, client behavior models, machine learning.
8. The method of claim 1 , further comprising detecting an attack and performing preventative measures that include at least one of stopping new backups from being generated, or preventing access to the backups in the data storage system.
9. The method of claim 1 , further comprising determining which of the candidate backups include the infection based on change rates across the candidate backups or deviations in the characteristics from a historical norm.
10. A non-transitory computer readable medium comprising computer executable instructions that, when executed, perform a method comprising:
identifying a plurality of candidate backups stored in a data storage system;
extracting characteristics from each of the candidate backups;
automatically analyzing the characteristics of the candidate backups to identify which of the candidate backups are infected with an infection, wherein analyzing the characteristics of the candidate backups includes comparing the characteristics of each of the plurality of candidate backups with each other;
assigning a classification for each of the plurality of candidate backups based on the analyzed characteristics;
automatically identifying a healthy backup that does not include the infection for performing a restore operation based on the classification assigned to each of the candidate backups; and
performing the restore operation by restoring production data from the healthy backup to production storage devices, wherein the restore operation includes restoring a first portion within the production data prior to a second portion within the production data, wherein the first portion within the production data has a higher priority than the second portion within the production data.
11. The non-transitory computer readable medium of claim 10 , wherein the candidate backups include full backups, incremental backups, or snapshots, wherein the characteristics of each of the candidate backups include file change information, file metadata, timestamps, and file extensions and wherein the infection is at least one of ransomware, a virus, a worm, malware, or combination thereof.
12. The non-transitory computer readable medium of claim 10 , further comprising:
mining the candidate backups to identify patterns indicative of the infection or using features associated with infected peers to identify the infection; and
using data augmentation to identify the healthy backup or the infection.
13. The non-transitory computer readable medium of claim 12 , wherein the data augmentation includes historical data obtained from a historical database, change rates for files, typical file sizes, file extensions, client behavior models, machine learning.
14. The non-transitory computer readable medium of claim 10 , further comprising detecting an attack and performing preventative measures that include at least one of stopping new backups from being generated, or preventing access to the backups in the data storage system.
15. The non-transitory computer readable medium of claim 10 , further comprising determining which of the plurality of candidate backups include the infection based on change rates across the candidate backups or deviations in the characteristics from a historical norm.
16. A method comprising:
detecting a data loss event in a computing system;
performing preventative measures in response to the data loss event;
automatically identifying a healthy backup from a plurality of candidate backups stored in a data storage system, wherein the healthy backup is believed to be free of infection, wherein the healthy backup is identified by analyzing characteristics of each of the plurality of the candidate backups, which includes comparing the characteristics of each of the plurality of candidate backups with each other; and
performing a restore operation from the healthy backup to storage devices, wherein the restore operation includes restoring a first portion within production data prior to a second portion within the production data, wherein the first portion within the production data has a higher priority than the second portion within the production data.
17. The method of claim 16 , wherein performing preventative measures includes stopping new backups from being generated and preventing access to the plurality of candidate backups.
18. The method of claim 16 , further comprising:
extracting features from the plurality of candidate backups;
evaluating the extracted features; and
assigning a score to each of the candidate backups based on the extracted features; and
selecting, as the healthy backup, a candidate backup with a best score from the plurality of candidate backups.
19. The method of claim 18 , further comprising:
identifying the healthy backup with augmented data;
evaluating the candidate backups based on the augmented data; and
assigning a score to each of the candidate backups based on the extracted features and the augmented data.
20. The method of claim 19 , wherein the extracted features include one or more of data extensions, data metadata, filenames, and data timestamps, further comprising:
determining whether the data extensions, data metadata, filenames and data timestamps indicate an infection;
comparing the data extensions from multiple backups in the plurality of candidate backups, comparing the data metadata from the multiple backups, comparing the filenames from the multiple backups, and comparing the data timestamps from the multiple backups;
determining rates of change from the data timestamps;
performing data augmentation with sources of information when identifying the healthy backup, wherein the sources of information include a historical database, a model of normal client behavior, and data from peer systems that have been infected, wherein the model of normal client behavior is based on deltas between successive candidate backups;
comprising generating the historical database or the model of normal client behavior with machine learning algorithms that have knowledge of typical behavior patterns.