Performance problem detection in arrays of similar hardware

• US 9,715,418 B2
• Filed: 12/02/2014
• Issued: 07/25/2017
• Est. Priority Date: 12/02/2014
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Barajas, Felipe A.
• Primary Examiner(s): Leibovich, Yair
• Application Number:
  US14/558,581
• Publication Number:
  US 20160154688A1
• Time in Patent Office: 966 Days
• Field of Search: 714 37
• CPC Class Codes: G06F 11/0727 in a storage system, e.g. in a DASD or network based storage system drivers for digital recording or reproducing units G06F3/06; circuits for error detection or correction within digital recording or reproducing units G11B20/18; for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS], H04L67/1097
• CPC Class Codes: G06F 11/0751 Error or fault detection not based on redundancy power supply failures G06F1/30; network fault management H04L41/06
• CPC Class Codes: G06F 11/34 Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G06F 11/3452 Performance evaluation by statistical analysis

Abstract

Embodiments described herein include detecting under-performing devices using a correlational matrix in which devices are compared with one another in order to establish which devices are performing significantly different, e.g. under-performing, than the rest. According to one embodiment, a method includes generating a vector for each of a plurality of hardware devices. Using a hardware processor, a correlational matrix representing projections of the vectors against each other is created. A determination is made as to which hardware device is least like the other hardware devices based on the correlational matrix. A result of the determining is output.

Claims

1. A computer-implemented method, comprising:

performing the following sub-process for a plurality of iterations;

generating a vector for each of a plurality of hardware devices,

creating, using a hardware processor, a correlational matrix representing projections of the vectors against each other,

determining which hardware device is least like the other hardware devices based on the correlational matrix, and

outputting a result of the determining;

identifying the hardware device that is most often least like the other hardware devices based on the results;

comparing a number of times the hardware device found most often to be least like the other hardware devices to a quorum; and

determining that the hardware device found most often to be least like the other hardware devices is underperforming in response to determining that the quorum has been met.

2. The method of claim 1, wherein the vector for each of the hardware devices is based on at least two performance metrics of the hardware device, wherein the number of performance metrics gathered for a particular device determines the number of dimensions of the vector that represents that device.

3. The method of claim 2, comprising calculating a statistical value for each of the performance metrics, the statistical value being derived from values from all of the hardware devices for the respective performance metric, wherein the performance metrics of the hardware device include a latest performance metric of the hardware device.

4. The method of claim 3, wherein each of the vectors is based on a deviation of the associated performance metric from the statistical value.

5. The method of claim 1, comprising calculating the projections of the vectors against each other.

6. The method of claim 5, wherein the projections of the vectors against each other are calculated using a trigonometric function applied to pairs of the vectors.

7. The method of claim 1, wherein the sub-process is performed in response to initiation of troubleshooting of the hardware devices.

8. The method of claim 7, wherein a number of iterations that the sub-process is performed is computed in response to initiation of the troubleshooting.

9. A computer program product for detecting an underperforming hardware device, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions readable and/or executable by a processor to cause the processor to:

perform the following sub-process for a plurality of iterations;

generate, by the processor, a vector for each of a plurality of hardware devices,

create, by the processor, a correlational matrix representing projections of the vectors against each other,

determine, by the processor, which hardware device is least like the other hardware devices based on the correlational matrix, and

output, by the processor, a result of the determining;

identify, by the processor, the hardware device that is most often least like the other hardware devices based on the results;

compare, by the processor, a number of times the hardware device found most often to be least like the other hardware devices to a quorum; and

determine, by the processor, that the hardware device found most often to be least like the other hardware devices is underperforming in response to determining that the quorum has been met.

10. The computer program product of claim 9, wherein the vector for each of the hardware devices is based on at least two performance metrics of the hardware device, wherein the performance metrics of the hardware device include a latest performance metric of the hardware device.

11. The computer program product of claim 10, comprising program instructions readable and/or executable by the processor to calculate a statistical value for each of the performance metrics, the statistical value being derived from values from all of the hardware devices for the respective performance metric, wherein the number of performance metrics gathered for a particular device determines the number of dimensions of the vector that represents that device.

12. The computer program product of claim 11, wherein each of the vectors is based on a deviation of the associated performance metric from the statistical value.

13. The computer program product of claim 9, comprising program instructions readable and/or executable by the processor to calculate the projections of the vectors against each other.

14. The computer program product of claim 13, wherein the projections of the vectors against each other are calculated using a trigonometric function applied to pairs of the vectors.

15. The computer program product of claim 9, wherein the sub-process is performed in response to initiation of troubleshooting of the hardware devices.

16. The computer program product of claim 15, wherein a number of iterations that the sub-process is performed is computed in response to initiation of the troubleshooting.

17. A system, comprising:

a hardware processor and logic integrated with and/or executable by the processor, the logic being configured to;

perform the following sub-process for a plurality of iterations;

generate a vector for each of a plurality of hardware devices,

create a correlational matrix representing projections of the vectors against each other,

determine which hardware device is least like the other hardware devices based on the correlational matrix, and

output a result of the determining;

identify the hardware device that is most often least like the other hardware devices based on the results;

compare a number of times the hardware device found most often to be least like the other hardware devices to a quorum;

determine that the hardware device found most often to be least like the other hardware devices is underperforming in response to determining that the quorum has been met; and

further monitor the hardware device found most often to be least like the other hardware devices in response to determining that the quorum has not been met,

wherein a number of iterations that the sub-process is performed is predefined,

wherein the quorum is defined as (predetermined number of iterations/2)+1,

wherein the result of the determining is output to a watch list,

wherein the result of the determining is output to a user interface,

wherein a pre-determined period of time elapses between each iteration of the sub-process.

18. The system of claim 17, wherein the vector for each of the hardware devices is based on at least three performance metrics of the hardware device, wherein the number of performance metrics gathered for a particular device determines the number of dimensions of the vector that represents that device.

19. The system of claim 18, comprising logic configured to calculate a statistical value for each of the performance metrics, the statistical value being derived from values from all of the hardware devices for the respective performance metric.

20. The system of claim 19, wherein each of the vectors is based on a deviation of the associated performance metric from the statistical value, wherein the sub-process is performed in response to troubleshooting of the hardware devices being initiated, wherein the number of iterations that the sub-process is performed is computed in response to initiation of the troubleshooting.