Data placement method based on health scores
Embodiments described herein relate to techniques for placing backup data based on health scores. The techniques may include: obtaining data items associated with a first data domain restorer; obtaining data items associated with a second data domain restorer; making a prediction that the first data domain restorer is operating normally; making a prediction that the second data domain restorer is operating normally; assigning a confidence value to the first prediction; making a classification of the first data domain restorer in a first group based on the confidence value; assigning a confidence value to the second prediction; making a classification of the second data domain restorer in a second group based on the confidence value; and performing a data backup to the first data domain restorer from a first computing device based on the classification and a first service level required for the first computing device.
1. A method for placing backup data based on health scores, the method comprising:
obtaining a first plurality of data items associated with a first data domain restorer;
obtaining a second plurality of data items associated with a second data domain restorer;
making a first prediction that the first data domain restorer is operating normally;
making a second prediction that the second data domain restorer is operating normally;
assigning a first confidence value to the first prediction;
making a first classification of the first data domain restorer in a first group based on the first confidence value;
assigning a second confidence value to the second prediction;
making a second classification of the second data domain restorer in a second group based on the second confidence value; and
performing a first data backup to the first data domain restorer from a first computing device based on the first classification and a first service level required for the first computing device.
2. The method of claim 1 , further comprising:
performing a second data backup to the second data domain restorer from a second computing device based on the second classification and a second service level required for the second computing device.
3. The method of claim 1 , further comprising:
obtaining a third plurality of data items associated with a third data domain restorer;
making a third prediction that the first data domain restorer is failed; and
removing, based on the third prediction, the third data domain restorer from a set of data domain restorers to be used for data backups.
4. The method of claim 1 , wherein the first confidence value and the second confidence value are part of a set of confidence values in a ranked list of confidence values.
5. The method of claim 1 , wherein the first data domain restorer and the second data domain restorer are part of a deduplication cluster.
6. The method of claim 1 , wherein the first prediction and the second prediction are made using a machine learning algorithm.
7. The method of claim 6 , wherein the machine learning algorithm is a nearest neighbor algorithm.
8. The method of claim 1 , wherein assigning the first confidence value comprises performing a conformal prediction analysis.
9. The method of claim 8 , wherein performing a conformal prediction analysis comprises:
assigning a failed label to the first data domain restorer;
performing a first comparison of the failed label using the first plurality of data items and a plurality of other data items for a plurality of other data domain restorers previously predicted to have failed labels to obtain a first non-conformity value;
assigning a normal label to the first data domain restorer; and
performing a second comparison of the normal label using the first plurality of data items and the plurality of other data items for the plurality of other data domain restorers previously predicted to have normal labels to obtain a second non-conformity value,
wherein the first confidence value is based on the second non-conformity value.
10. The method of claim 1 , wherein the first group comprises a first plurality of data domain restorers with confidence values above a threshold value and the second group comprises a second plurality of data domain restorers with confidence values below the threshold value.
11. The method of claim 10 , wherein the threshold value is associated with a predicted amount of down time.
12. A non-transitory computer readable medium comprising computer readable program code, which when executed by a computer processor enables the computer processor to perform a method for placing backup data based on health scores, the method comprising:
obtaining a first plurality of data items associated with a first data domain restorer;
obtaining a second plurality of data items associated with a second data domain restorer;
making a first prediction that the first data domain restorer is operating normally;
making a second prediction that the second data domain restorer is operating normally;
assigning a first confidence value to the first prediction;
making a first classification of the first data domain restorer in a first group based on the first confidence value;
assigning a second confidence value to the second prediction;
making a second classification of the second data domain restorer in a second group based on the second confidence value; and
performing a first data backup to the first data domain restorer from a first computing device based on the first classification and a first service level required for the first computing device.
13. The non-transitory computer readable medium of claim 12 , wherein the method further comprises:
performing a second data backup to the second data domain restorer from a second computing device based on the second classification and a second service level required for the second computing device.
14. The non-transitory computer readable medium of claim 12 , wherein assigning the first confidence value comprises performing a conformal prediction analysis comprising:
assigning a failed label to the first data domain restorer;
performing a first comparison of the failed label using the first plurality of data items and a plurality of other data items for a plurality of other data domain restorers previously predicted to have failed labels to obtain a first non-conformity value;
assigning a normal label to the first data domain restorer; and
performing a second comparison of the normal label using the first plurality of data items and the plurality of other data items for the plurality of other data domain restorers previously predicted to have normal labels to obtain a second non-conformity value,
wherein the first confidence value is based on the second non-conformity value.
15. The non-transitory computer readable medium of claim 12 , further comprising:
obtaining a third plurality of data items associated with a third data domain restorer;
making a third prediction that the first data domain restorer is failed; and
removing, based on the third prediction, the third data domain restorer from a set of data domain restorers to be used for data backups.
16. The non-transitory computer readable medium of claim 12 , wherein the first prediction and the second prediction are made using a machine learning algorithm.
17. The non-transitory computer readable medium of claim 16 , wherein the machine learning algorithm is a nearest neighbor algorithm.
18. The non-transitory computer readable medium of claim 12 , wherein the first group comprises a first plurality of data domain restorers with confidence values above a threshold value and the second group comprises a second plurality of data domain restorers with confidence values below the threshold value.
19. The non-transitory computer readable medium of claim 18 , wherein the threshold value is associated with a predicted amount of down time.
20. A system for placing backup data based on health scores, the system comprising:
a first data domain restorer comprising a first processor, first memory, and a first persistent storage device;
a second data domain restorer comprising a second processor, second memory, and a second persistent storage device; and
a data placement controller comprising a third processor, third memory, and a third persistent storage device, and configured to:
obtain a first plurality of data items associated with the first data domain restorer;
obtain a second plurality of data items associated with the second data domain restorer;
make a first prediction that the first data domain restorer is operating normally;
make a second prediction that the second data domain restorer is operating normally;
assign a first confidence value to the first prediction;
make a first classification of the first data domain restorer in a first group based on the first confidence value;
assign a second confidence value to the second prediction;
make a second classification of the second data domain restorer in a second group based on the second confidence value; and
perform a first data backup to the first data domain restorer from a first computing device based on the first classification and a first service level required for the first computing device.