Granted Patent B2

US 11,249,862 · App. 16/867,003 · Granted Feb 15, 2022

Inventors: Aaron Ren Wang (Shanghai, CN); Weiyang Liu (Shanghai, CN); Ming Zhang (Shanghai, CN); Qi Wang (Shanghai, CN); Jerry Zengjie Zhang (Shanghai, CN)

Assignee: EMC IP HOLDING COMPANY LLC

G06F11/1464G06F11/1461G06F11/3476G06K9/6215G06N7/005G06F2201/81

Abstract

Claims (67)

1. A method of managing a backup system, comprising:

extracting, based on historical backup information in a first time period, at least one temporal feature corresponding to the first time period, wherein extracting the at least one temporal feature comprises decomposing temporal data corresponding to the historical backup information, wherein the historical backup information was generated by:

recording the historical backup information of the backup system and a backup log including all backup jobs, and

aggregating the backup log to determine a set of numbers of the backup jobs that have been executed in the first time period;

determining, based on the at least one temporal feature input into a temporal prediction model, a set of predicted numbers of backup jobs expected to be executed by the backup system in a future second time period, wherein the temporal prediction model was trained based on the historical backup information; and

determining a health status of the backup system based on the set of predicted numbers and a set of actual numbers of backup jobs executed by the backup system in the second time period.

2. The method of claim 1 , wherein

decomposing temporal data corresponding to the historical backup information comprises obtaining at least one of the following as the at least one temporal feature:

a trend feature indicating a variation feature of the temporal data,

a periodic feature indicating a cyclic feature of the temporal data, or

a random feature indicating randomness other than the trend feature and the cyclic feature of the temporal data.

3. The method of claim 1 , wherein determining the set of predicted numbers comprises:

determining the set of predicted numbers by processing the at least one temporal feature with the temporal prediction model, the temporal prediction model being built based on prior backup information of the backup system.

4. The method of claim 1 , wherein determining the health status of the backup system comprises:

for a first predicted number in the set of predicted numbers:

determining a number interval associated with the first predicted number, a number within the number interval having a confidence greater than a predetermined threshold;

determining a first actual number corresponding to the first predicted number from the set of actual numbers; and

determining the health status of the backup system by comparing the first actual number with the number interval.

5. The method of claim 4 , wherein determining the health status of the backup system by comparing the first actual number with the number interval comprises:

determining a proportion of actual numbers falling into a corresponding number interval in the set of actual numbers; and

in accordance with a determination that the proportion is less than a predetermined proportion threshold, determining that the backup system is abnormal.

6. The method of claim 4 , wherein determining the health status of the backup system by comparing the first actual number with the number interval comprises:

determining a count of actual numbers falling into a corresponding number interval in the set of actual numbers; and

in accordance with a determination that the count is less than a predetermined count threshold, determining that the backup system is abnormal.

7. The method of claim 1 , further comprising:

in accordance with a determination that the health status indicates abnormality of the backup system, providing an alert of the abnormality.

8. An electronic device, comprising:

at least one processing unit; and

at least one memory coupled to the at least one processing unit and storing instructions executable by the at least one processing unit, the instructions, when executed by the at least one processing unit, causing the device to perform acts comprising:

extracting, based on historical backup information in a first time period, at least one temporal feature corresponding to the first time period, wherein extracting the at least one temporal feature comprises decomposing temporal data corresponding to the historical backup information, wherein the historical backup information was generated by:

recording the historical backup information of a backup system and a backup log including all backup jobs, and

aggregating the backup log to determine a set of numbers of the backup jobs that have been executed in the first time period;

determining, based on the at least one temporal feature input into a temporal prediction model, a set of predicted numbers of backup jobs expected to be executed by the backup system in a future second time period, wherein the temporal prediction model was trained based on the historical backup information; and

determining a health status of the backup system based on the set of predicted numbers and a set of actual numbers of backup jobs executed by the backup system in the second time period.

9. The device of claim 8 , wherein

decomposing temporal data corresponding to the historical backup information comprises obtaining at least one of the following as the at least one temporal feature:

a trend feature indicating a variation feature of the temporal data,

a periodic feature indicating a cyclic feature of the temporal data, or

a random feature indicating randomness other than the trend feature and the cyclic feature of the temporal data.

10. The device of claim 8 , wherein determining the set of predicted numbers comprises:

determining the set of predicted numbers by processing the at least one temporal feature using a temporal prediction model, the temporal prediction model being built based on previous backup information of the backup system.

11. The device of claim 8 , wherein determining the health status of the backup system comprises:

for a first predicted number in the set of predicted numbers:

determining a number interval associated with the first predicted number, numbers within the number interval having a confidence greater than a predetermined threshold;

determining a first actual number corresponding to the first predicted number from the set of actual numbers; and

determining the health status of the backup system by comparing the first actual number with the number interval.

12. The device of claim 11 , wherein determining the health status of the backup system by comparing the first actual number with the number interval comprises:

determining a proportion of actual numbers falling into a corresponding number interval in the set of actual numbers; and

in accordance with a determination that the proportion is less than a predetermined proportion threshold, determining that the backup system is abnormal.

13. The device of claim 11 , wherein determining the health status of the backup system by comparing the first actual number with the number interval comprises:

determining a count of actual numbers falling into a corresponding number interval in the set of actual numbers; and

in accordance with a determination that the count is less than a predetermined count threshold, determining that the backup system is abnormal.

14. The device of claim 8 , the acts further comprising:

in accordance with a determination that the health status indicates abnormality of the backup system, providing an alert of the abnormality.

15. A computer program product being tangibly stored on a non-transient computer storage medium and comprising machine-executable instructions which, when executed by a device, cause the device to perform acts, the acts comprising:

extracting, based on historical backup information in a first time period, at least one temporal feature corresponding to the first time period, wherein extracting the at least one temporal feature comprises decomposing temporal data corresponding to the historical backup information, wherein the historical backup information was generated by:

recording the historical backup information of a backup system and a backup log including all backup jobs, and

aggregating the backup log to determine a set of numbers of the backup jobs that have been executed in the first time period;

determining, based on the at least one temporal feature input into a temporal prediction model, a set of predicted numbers of backup jobs expected to be executed by the backup system in a future second time period, wherein the temporal prediction model was trained based on the historical backup information; and

determining a health status of the backup system based on the set of predicted numbers and a set of actual numbers of backup jobs executed by the backup system in the second time period.

16. The computer program product of claim 15 , wherein

decomposing temporal data corresponding to the historical backup information comprises obtaining at least one of the following as the at least one temporal feature:

a trend feature indicating a variation feature of the temporal data,

a periodic feature indicating a cyclic feature of the temporal data, or

a random feature indicating randomness other than the trend feature and the cyclic feature of the temporal data.

17. The computer program product of claim 15 , wherein determining the set of predicted numbers comprises:

determining the set of predicted numbers by processing the at least one temporal feature with a temporal prediction model, the temporal prediction model being built based on prior backup information of the backup system.

Assignments (9)

Cited By (1)

US 12,711,018