System and method for evaluating a pattern of resource demands of a workload

• US 8,543,711 B2
• Filed: 04/30/2007
• Issued: 09/24/2013
• Est. Priority Date: 04/30/2007
• Status: Active Grant

• Current Assignee: Valtrus Innovations Limited (f/k/a Dolya Holdco 9 Limited) (Key Patent Innovations Limited)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Rolia, Jerome, Gmach, Daniel, Cherkasova, Ludmila
• Primary Examiner(s): NANO, SARGON N
• Application Number:
  US11/742,519
• Publication Number:
  US 20080271038A1
• Time in Patent Office: 2,339 Days
• Field of Search: 709/217, 709/224, 709/226, 709/229
• US Class Current: 709/229
• CPC Class Codes: G06F 11/3414 Workload generation, e.g. scripts, playback
• CPC Class Codes: G06F 11/3452 Performance evaluation by statistical analysis
• CPC Class Codes: G06F 2209/5019 Workload prediction
• CPC Class Codes: G06F 2209/508 Monitor
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

A method comprises receiving, by pattern evaluation logic, a plurality of occurrences of a prospective pattern of resource demands in a representative workload. The method further comprises evaluating, by the pattern evaluation logic, the received occurrences of the prospective pattern of resource demands, and determining, by the pattern evaluation logic, based on the evaluation of the received occurrences of the prospective pattern of resource demands, how representative the prospective pattern is of resource demands of the representative workload.

Claims

1. A method comprising:

receiving, by pattern evaluation logic implemented in a computer processor, a plurality of occurrences of a prospective pattern of resource demands in a representative workload;

evaluating, by the pattern evaluation logic, the received occurrences of the prospective pattern of resource demands; and

determining, by the pattern evaluation logic, based on said evaluation of the received occurrences of the prospective pattern of resource demands, how representative the prospective pattern is of resource demands of the representative workload by analyzing a distribution of differences between the occurrences of the prospective pattern of resource demands.

2. The method of claim 1 wherein said determining further comprises:

determining how representative the prospective pattern is of the resource demands of the representative workload based on determined similarity of the distribution of differences between the plurality of occurrences of the prospective pattern.

3. The method of claim 2 wherein said determining further comprises:

determining that the more similar the distribution of differences between the occurrences of the prospective pattern, the more representative the prospective pattern is of the resource demands of the representative workload.

4. The method of claim 1 wherein each occurrence of the prospective pattern comprises a plurality of resource demand measurement intervals.

5. The method of claim 4 wherein said evaluating the received occurrences of the prospective pattern of resource demands comprises:

determining differences between corresponding ones of said plurality of resource demand measurement intervals of a plurality of different occurrences of the prospective pattern.

6. The method of claim 5 wherein said evaluating further comprises:

determining a distribution of the determined differences between the plurality of occurrences of the prospective pattern of resource demands.

7. The method of claim 1, further comprising:

determining a distribution of differences between the occurrences of the prospective pattern of resource demands; and

evaluating said prospective pattern based at least in part on a determined distribution of differences between the identified occurrences of the prospective pattern.

8. The method of claim 7 wherein said determining said distribution of differences comprises:

employing a cumulative distribution function (CDF) for determining the distribution of differences between the identified occurrences of the prospective pattern.

9. The method of claim 8 wherein said evaluating comprises:

comparing the determined distribution of differences between the identified occurrences of the prospective pattern of resource demands.

10. The method of claim 9 wherein said evaluating comprises:

evaluating whether the prospective pattern is representative of the resource demands of the representative workload.

11. The method of claim 10 further comprising:

employing a two-sample Kolmogorov-Smirnov test (KS-test) to compare the CDFs of differences for pairs of occurrences of the prospective pattern of resource demands.

12. The method of claim 11 further comprising:

determining whether the KS-test results exceed a pre-defined threshold, wherein if the pre-defined threshold is exceeded, the prospective pattern is determined as representative of the resource demands of the representative workload.

13. The method of claim 10 further comprising:

employing a Chi-squared based test to compare the CDFs of differences for pairs of occurrences of the prospective pattern of resource demands.

14. The method of claim 7 further comprising:

determining whether a subset of the occurrences of the prospective pattern reflect an anomaly.

15. The method of claim 14 further comprising:

determining, based on occurrences of the prospective pattern that do not reflect an anomaly, a trend in the resource demands of the representative workload.

16. The method of claim 1, in which determining how representative the prospective pattern is of resource demands of the representative work load includes determining a measure of the representativeness by analyzing a distribution of differences between the occurrences of the prospective pattern of resource demands.

17. A method comprising:

receiving, by pattern evaluation logic implemented in a computer processor, a plurality of occurrences of a prospective pattern of resource demands in a representative workload;

evaluating, by the pattern evaluation logic, the received occurrences of the prospective pattern of resource demands; and

determining, by the pattern evaluation logic, based on said evaluation of the received occurrences of the prospective pattern of resource demands, how representative the prospective pattern is of resource demands of the representative workload by;

analyzing a distribution of differences between the plurality of occurrences of the prospective pattern of resource demands, and determining that the more similar the distribution of differences between the occurrences of the prospective pattern, the more representative the prospective pattern is of the resource demands of the representative workload.

18. The method of claim 17, wherein said evaluating the received occurrences of the prospective pattern of resource demands comprises:

determining differences between corresponding ones of said plurality of resource demand measurement intervals of a plurality of different occurrences of the prospective pattern; and

determining a distribution of the determined differences between the plurality of occurrences of the prospective pattern of resource demands.

19. A method comprising:

receiving, by pattern evaluation logic implemented in a computer processor, a plurality of occurrences of a prospective pattern of resource demands in a representative workload;

evaluating, by the pattern evaluation logic, the received occurrences of the prospective pattern of resource demands;

determining, by the pattern evaluation logic, based on said evaluation of the received occurrences of the prospective pattern of resource demands, how representative the prospective pattern is of resource demands of the representative workload;

wherein determining how representative the prospective pattern is of resource demands comprises determining a distribution of differences between the occurrences of the prospective pattern of resource demands using a cumulative distribution function (CDF), comparing the determined distribution of differences between the identified occurrences of the prospective pattern of resource demands, and evaluating whether the prospective pattern is representative of the resource demands of the representative workload by employing a two-sample Kolmogorov-Smirnov test (KS-test) to compare the CDFs of differences for pairs of occurrences of the prospective pattern of resource demands.

20. The method of claim 19, further comprising:

determining whether the KS-test results exceed a pre-defined threshold, wherein if the pre-defined threshold is exceeded, the prospective pattern is determined as representative of the resource demands of the representative workload.

21. The method of claim 20, further comprising:

employing a Chi-squared based test to compare the CDFs of differences for pairs of occurrences of the prospective pattern of resource demands.