System and method for probabilistically forecasting health of hardware in a large-scale system
Embodiments described herein relate to a method for probabilistically forecasting the state of hardware components. The method may include obtaining data items corresponding to a hardware component and performing an analysis of the hardware component. The analysis may include making a variety of probability predictions as to whether a label from among a set of possible labels is likely to be the correct label. The set of probabilities from the aforementioned analysis are then analyzed to determine which predicted label has the tightest range, and the prediction with the tightest range for a certain label is displayed to a user in a ranked fashion that includes a quantity of such probability prediction ranges. Such a display may allow an administrator to take action as to which hardware components should be replaced and in what order.
1. A method for probabilistically forecasting a state of hardware components, the method comprising:
obtaining a plurality of data items corresponding to a hardware component;
performing an analysis to determine a hardware failure of the hardware component, wherein the analysis comprises:
making a first prediction that the hardware component should be associated with a first label, wherein the first label indicates that the hardware component is predicted to be in a failed state;
performing, based on the first prediction and using the plurality of data items and the first label, a first comparison of the hardware component to a plurality of other hardware components previously predicted to have the first label to obtain a first percentage of matches;
performing, based on the first prediction and using the plurality of data items and the first label, a second comparison of the hardware component to a second plurality of other hardware components previously predicted to have a second label to obtain a second percentage of matches;
making a second prediction that the hardware component should be associated with a second label, wherein the second label indicates that the hardware component is predicted to be in a normal state;
performing, based on the second prediction and using the plurality of data items and the second label, a third comparison of the hardware component to the plurality of other hardware components previously predicted to have the first label to obtain a third percentage of matches; and
performing, based on the second prediction and using the plurality of data items and the second label, a fourth comparison of the hardware component to the second plurality of other hardware components previously predicted to have the second label to obtain a fourth percentage of matches;
generating a results matrix based on the first percentage of matches, the second percentage of matches, the third percentage of matches, and the fourth percentage of matches, wherein the results matrix represents a prediction of whether the hardware component is likely to be in a failed state within a timeframe;
displaying on a display the results matrix; and
replacing the hardware component based on the analysis.
2. The method of claim 1 , further comprising performing an analysis on the results matrix to make a determination of which column of the results matrix is used in the display presented to a user.
3. The method of claim 2 , wherein the determination is based on the first percentage of matches and the second percentage matches being numerically closer to one another than the third percentage of matches and the fourth percentage of matches.
4. The method of claim 2 , wherein the display further comprises a plurality of other columns selected using a plurality of other results matrix analyses.
5. The method of claim 1 , wherein the hardware component is a storage component.
6. The method of claim 5 , wherein the plurality of data items comprises at least one selected from a group consisting of read errors, seek errors, power cycle data, reallocation event data, temperature data, and cyclic redundancy check error data.
7. The method of claim 1 , wherein the hardware component is one selected from a group consisting of a processor, a battery, and a network component.
8. The method of claim 1 , further comprising:
calculating an accuracy score for an analysis of whether the hardware component is failed or normal; and
making a determination, based on the accuracy score, that the analysis meets an accuracy threshold.
9. 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 probabilistically forecasting a state of hardware components, the method comprising:
obtaining a plurality of data items corresponding to a hardware component;
performing an analysis to determine a hardware failure of the hardware component, wherein the analysis comprises:
making a first prediction that the hardware component should be associated with a first label, wherein the first label indicates that the hardware component is predicted to be in a failed state;
performing, based on the first prediction and using the plurality of data items and the first label, a first comparison of the hardware component to a plurality of other hardware components previously predicted to have the first label to obtain a first percentage of matches;
performing, based on the first prediction and using the plurality of data items and the first label, a second comparison of the hardware component to a second plurality of other hardware components previously predicted to have a second label to obtain a second percentage of matches;
making a second prediction that the hardware component should be associated with a second label, wherein the second label indicates that the hardware component is predicted to be in a normal state;
performing, based on the second prediction and using the plurality of data items and the second label, a third comparison of the hardware component to the plurality of other hardware components previously predicted to have the first label to obtain a third percentage of matches; and
performing, based on the second prediction and using the plurality of data items and the second label, a fourth comparison of the hardware component to the second plurality of other hardware components previously predicted to have the second label to obtain a fourth percentage of matches;
generating a results matrix based on the first percentage of matches, the second percentage of matches, the third percentage of matches, and the fourth percentage of matches, wherein the results matrix represents a prediction of whether the hardware component is likely to be in a failed state within a timeframe;
displaying on a display the results matrix; and
replacing the hardware component based on the analysis.
10. The non-transitory computer readable medium of claim 9 , wherein execution of the computer readable program code further comprises performing an analysis on the results matrix to make a determination of which column of the results matrix is used in the display presented to a user.
11. The non-transitory computer readable of claim 10 , wherein the determination is based on the first percentage of matches and the second percentage matches being numerically closer to one another than the third percentage of matches and the fourth percentage of matches.
12. The non-transitory computer readable of claim 10 , wherein the display further comprises a plurality of other columns selected using a plurality of other results matrix analyses.
13. The non-transitory computer readable medium of claim 9 , wherein the hardware component is a storage component.
14. The non-transitory computer readable medium of claim 13 , wherein the plurality of data items comprises at least one selected from a group consisting of read errors, seek errors, power cycle data, reallocation event data, temperature data, and cyclic redundancy check error data.
15. The non-transitory computer readable medium of claim 9 , wherein the hardware component is one selected from a group consisting of a processor, a battery, and a network component.
16. The non-transitory computer readable medium of claim 9 , wherein execution of the computer readable program code further comprises:
calculating an accuracy score for an analysis of whether the hardware component is failed or normal; and
making a determination, based on the accuracy score, that the analysis meets an accuracy threshold.