Method and system for risk score based asset data protection using a conformal framework
A method and system for risk score based asset data protection using a conformal framework. Concerning Internet-of-Things (IoT) environments, there is a growing challenge to ensure the availability and continuity of applications executing, as well as relevant application data maintained, on IoT devices, in order to mitigate potential operational failures and/or interruptions. The disclosed method and system address this growing challenge through risk-score assessments of the applications and their respective relevant application data (collectively referred to as assets), as well as the potential degradation of the underlying IoT device hardware. The aforementioned risk-score assessments employ clustering to identify, and a conformal framework to rank, high-risk (for failure) assets, which may subsequently be subjected to backup operations (overriding existing policies) to guarantee asset protection.
1. A method for asset data protection, comprising:
identifying a set of assets hosted across a host device cluster comprising a set of host devices;
ranking, in descending order and to obtain a ranked subset of assets, a subset of the set of assets based on a risk score calculated for each asset in the subset of the set of assets;
identifying at least one failure-soon asset from the ranked subset of assets; and
performing, while overriding any existing data protection policy configured thereto, a backup operation targeting the at least one failure-soon asset,
wherein the risk score calculated for each asset in the subset of the set of assets is provided using a conformal framework.
2. The method of claim 1 , further comprising:
prior to ranking the subset of the set of assets:
collecting performance metrics for each asset in the set of assets; and
assigning, based on the performance metrics, each asset in the set of assets to one selected from a group consisting of a high-risk class and a low-risk class,
wherein each asset in the ranked subset of assets is a member of the high-risk class.
3. The method of claim 2 , wherein assignment of each asset in the set of assets, to one selected from the group consisting of the high-risk class and the low-risk class, resulted from machine learning classification performed through cluster analysis.
4. The method of claim 2 , wherein the conformal framework associates a confidence value with each assignment mapping an asset in the subset of the set of assets to the high-risk class.
5. The method of claim 4 , wherein the risk score, calculated for each asset in the subset of the set of assets, comprises the confidence value associated with the assignment mapping the asset to the high-risk class.
6. The method of claim 2 , wherein performance metrics for a respective host device, in the set of host devices, on which each asset in the set of assets resides, are also collected.
7. The method of claim 2 , wherein the at least one failure-soon asset is identified from assets of the high-risk class based on a risk cut-off threshold.
8. The method of claim 7 , wherein the risk score of each failure-soon asset of the at least one failure-soon asset at least matches the risk cut-off threshold.
9. The method of claim 1 , wherein performing the backup operation targeting the at least one failure-soon asset comprises storing at least one selected from a group consisting of configuration information, computer code, and data, pertinent to the at least one failure-soon asset onto a backup storage system.
10. The method of claim 1 , wherein each host device in the set of host devices is an Internet-of-Things (IoT) device.
11. A non-transitory computer readable medium (CRM) comprising computer readable program code, which when executed by a computer processor, enables the computer processor to:
identify a set of assets hosted across a host device cluster comprising a set of host devices;
rank, in descending order and to obtain a ranked subset of assets, a subset of the set of assets based on a risk score calculated for each asset in the subset of the set of assets;
identify at least one failure-soon asset from the ranked subset of assets; and
perform, while overriding any existing data protection policy configured thereto, a backup operation targeting the at least one failure-soon asset,
wherein the risk score calculated for each asset in the subset of the set of assets is provided using a conformal framework.
12. The non-transitory CRM of claim 11 , further comprising computer readable program code, which when executed by the computer processor, further enables the computer processor to:
prior to ranking the subset of the set of assets:
collect performance metrics for each asset in the set of assets; and
assign, based on the performance metrics, each asset in the set of assets to one selected from a group consisting of a high-risk class and a low-risk class,
wherein each asset in the ranked subset of assets is a member of the high-risk class.
13. The non-transitory CRM of claim 12 , wherein assignment of each asset in the set of assets, to one selected from the group consisting of the high-risk class and the low-risk class, resulted from machine learning classification performed through cluster analysis.
14. The non-transitory CRM of claim 12 , wherein the conformal framework associates a confidence value with each assignment mapping an asset in the subset of the set of assets to the high-risk class.
15. The non-transitory CRM of claim 13 , wherein the risk score, calculated for each asset in the subset of the set of assets, comprises the confidence value associated with the assignment mapping the asset to the high-risk class.
16. The non-transitory CRM of claim 12 , wherein performance metrics for a respective host device, in the set of host devices, on which each asset in the set of assets resides, are also collected.
17. The non-transitory CRM of claim 12 , wherein the at least one failure-soon asset is identified from assets of the high-risk class based on a risk cut-off threshold.
18. The non-transitory CRM of claim 17 , wherein the risk score of each failure-soon asset of the at least one failure-soon asset at least matches the risk cut-off threshold.
19. The non-transitory CRM of claim 11 , wherein performing the backup operation targeting the at least one failure-soon asset comprises storing at least one selected from a group consisting of configuration information, computer code, and data, pertinent to the at least one failure-soon asset onto a backup storage system.
20. The non-transitory CRM of claim 11 , wherein each host device in the set of host devices is an Internet-of-Things (IoT) device.