Enterprise management system and method which indicates chaotic behavior in system resource usage for more accurate modeling and prediction

• US 6,564,174 B1
• Filed: 09/29/1999
• Issued: 05/13/2003
• Est. Priority Date: 09/29/1999
• Status: Expired due to Term

• Current Assignee: BMC Software Incorporated (KKR & Co., Inc.)
• Original Assignee: BMC Software Incorporated (KKR & Co., Inc.)
• Inventors: Fiorini, Pierre, Ding, Yiping
• Primary Examiner(s): Barlow, John
• Assistant Examiner(s): VO, HIEN XUAN
• Application Number:
  US09/408,085
• Time in Patent Office: 1,322 Days
• Field of Search: 702/186, 702/187, 702/188, 709/223, 709/224, 709/225, 709/226, 370/229, 370/235, 370/254, 370/241, 370/252, 706/25
• US Class Current: 702/186
• CPC Class Codes: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling

Abstract

A set of algorithms for detecting a renewal power-tail behavior which often relates to chaotic system activities in one or more computer system resources of a distributed computing environment, i.e., an enterprise. Analysis and/or prediction software receives a set of metric data points from agent software on one or more computer systems. The analysis and/or prediction software performs three analytic tests relating to distinctive properties of power-tail distributions: a first test to determine whether the set of data points exhibits large deviations from the mean, a second test to determine whether the set of data points exhibits a high variance, and a third test to determine whether the set of the largest data points exhibits properties consistent with large values in a tail portion of a power-tail distribution. The tests can be performed in any order, and in other embodiments, fewer than three can be performed. Each test has two possible results: successful if the test indicates a likelihood of power-tail behavior, or unsuccessful if it indicates that power-tail behavior is unlikely. The results of the three tests are combined to determine the overall likelihood of a power-tail distribution. If all three tests are successful, then power-tail behavior is likely. If all three tests are unsuccessful, then power-tail behavior is unlikely. If the results are mixed, then typically more data or analysis is needed. The results are used for modeling and/or altering the configuration of the enterprise.

Claims

1. A method for monitoring the performance of an enterprise, wherein the enterprise comprises one or more computer systems, the method comprising:

receiving a set of data points from said one or more computer systems, wherein said set of data points represents a series of measurements of one or more system resources of said one or more computer systems;

determining whether there is a power-tail behavior in said set of data points, comprising;

performing two or more analytic tests on said set of data points;

combining results of said analytic tests to determine a likelihood of said power-tail behavior;

modeling said enterprise based on said determining whether there is a power-tail behavior in said set of data points.

2. The method of claim 1, further comprising:

altering a configuration of said enterprise based on said determining whether there is a power-tail behavior in said set of data points.

3. The method of claim 1,

wherein said modeling said enterprise based on said determining whether there is power-tail behavior in said set of data points further comprises:

modifying a model of said enterprise.

4. The method of claim 1,

wherein said modeling said enterprise based on said determining whether there is a power-tail behavior in said set of data points further comprises:

displaying said likelihood of said power-tail behavior.

5. The method of claim 1,

wherein performing said analytic tests on said set of data points further comprises performing a first analytic test, a second analytic test, and a third analytic test on said set of data points;

wherein said combining results of said analytic tests to determine a likelihood of said power-tail behavior further comprises combining results of said first analytic test, said second analytic test, and said third analytic test to determine said likelihood of said power-tail behavior.

6. The method of claim 5,

wherein performing said first analytic test on said set of data points further comprises determining whether said set of data points exhibits large deviations from a mean of said data points.

7. The method of claim 5,

wherein performing said second analytic test on said set of data points further comprises determining whether said set of data points exhibits a high variance.

8. The method of claim 5,

wherein performing said third analytic test on said set of data points further comprises determining whether a set of largest data points of said set of data points exhibits properties consistent with large values in a tail portion of a power-tail distribution.

9. The method of claim 5,

wherein performing said first analytic test on said set of data points further comprises:

determining a substantially largest data point of said set of data points;

computing a probability P_D that a random variable X is greater than or equal to said substantially largest data point;

computing a probability P_E that said random variable X is greater than or equal to an expected value of a largest order statistic of an exponential distribution;

determining a tolerance factor;

determining whether said probability P_D is less than or equal to said tolerance factor;

determining whether said probability P_D is less than or equal to said probability P_E.

10. The method of claim 9,

wherein said determining whether said probability P_D is less than or equal to said tolerance factor further comprises determining that said probability P_D is less than or equal to said tolerance factor;

wherein said determining whether said probability P_D is less than or equal to said probability P_E further comprises determining that said probability P_D is less than or equal to said probability P_E;

wherein performing said first analytic test further comprises;

determining that said first analytic test is successful.

11. The method of claim 9,

wherein said determining whether said probability P_D is less than or equal to said tolerance factor further comprises determining that said probability P_D is less than or equal to said tolerance factor;

wherein said determining whether said probability P_D is less than or equal to said probability P_E further comprises determining that said probability P_D is not less than or equal to said probability P_E;

wherein performing said first analytic test further comprises;

determining that said first analytic test is unsuccessful.

12. The method of claim 9,

wherein said determining whether said probability P_D is less than or equal to said tolerance factor further comprises determining that said probability P_D is not less than or equal to said tolerance factor;

wherein said determining whether said probability P_D is less than or equal to said probability P_E further comprises determining that said probability P_D is less than or equal to said probability P_E;

wherein performing said first analytic test further comprises;

determining that said first analytic test is unsuccessful.

13. The method of claim 9,

wherein said determining whether said probability P_D is less than or equal to said tolerance factor further comprises determining that said probability P_D is not less than or equal to said tolerance factor;

wherein said determining whether said probability P_D is less than or equal to said probability P_E further comprises determining that said probability P_D is not less than or equal to said probability P_E;

wherein performing said first analytic test further comprises;

determining that said first analytic test is unsuccessful.

14. The method of claim 5,

wherein performing said second analytic test on said set of data points further comprises:

computing a power-tail variance index for a power-tail distribution with a power-tail index α

of 2;

computing a variance of said set of data points;

determining whether said variance of said set of data points is greater than r equal to said power-tail variance index.

15. The method of claim 14,

wherein said determining whether said variance of said set of data points is greater than or equal to said power-tail variance index further comprises determining that said variance of said set of data points is greater than or equal to said power-tail variance index;

wherein performing said second analytic test further comprises;

determining that said second analytic test is successful.

16. The method of claim 14,

wherein said determining whether said variance of said set of data points is greater than or equal to said power-tail variance index further comprises determining that said variance of said set of data points is not greater than or equal to said power-tail variance index;

wherein performing said second analytic test further comprises;

determining that said second analytic test is unsuccessful.

17. The method of claim 5,

wherein performing said third analytic test on said set of data points further comprises:

normalizing said set of data points such that an expected value of said set of data points is 1;

determining a substantially largest data point of said set of data points;

estimating a power-tail index α

of said set of data points;

determining whether said power-tail index α

of said set of data points is less than 2.

18. The method of claim 17,

wherein said determining whether said power-tail index α

of said set of data points is less than 2 further comprises determining that said power-tail index α

of said set of data points is less than 2;

wherein performing said third analytic test further comprises;

determining that said third analytic test is successful.

19. The method of claim 17,

wherein said determining whether said power-tail index α

of said set of data points is less than 2 further comprises determining that said power-tail index α

of said set of data points is not less than 2;

wherein performing said third analytic test further comprises;

determining that said third analytic test is unsuccessful.

20. The method of claim 5,

wherein said determining whether there is a power-tail behavior in said set of data points further comprises:

determining that said first analytic test is unsuccessful;

determining that said second analytic test is unsuccessful;

determining that said third analytic test is unsuccessful;

wherein said combining results of said first analytic test, said second analytic test, and said third analytic test to determine said likelihood of said power-tail behavior further comprises;

concluding that said power-tail behavior is unlikely.

21. The method of claim 5,

wherein said determining whether there is a power-tail behavior in said set of data points further comprises:

determining that said first analytic test is successful;

determining that said second analytic test is successful;

determining that said third analytic test is successful;

wherein said combining results of said first analytic test, said second analytic test, and said third analytic test to determine said likelihood of said power-tail behavior further comprises;

concluding that said power-tail behavior is likely.

22. The method of claim 5,

wherein said determining whether there is a power-tail behavior in said set of data points further comprises:

determining that two of said three analytic tests are unsuccessful;

determining that one of said three analytic tests is successful;

wherein said combining results of said first analytic test, said second analytic test, and said third analytic test to determine said likelihood of said power-tail behavior further comprises;

concluding that said power-tail behavior is possible.

23. The method of claim 5,

wherein said determining whether there is a power-tail behavior in said set of data points further comprises:

determining that one of said three analytic tests is unsuccessful;

determining that two of said three analytic tests are successful;

wherein said combining results of said first analytic test, said second analytic test, and said third analytic test to determine said likelihood of said power-tail behavior further comprises;

concluding that said power-tail behavior is possible.

24. The method of claim 1,

wherein said data points are independent and identically distributed;

wherein said determining whether there is a power-tail behavior in said set of data points further comprises determining whether there is a renewal power-tail behavior in said set of data points.

25. A system for monitoring the performance of an enterprise, wherein the enterprise comprises one or more computer systems, the system comprising:

one or more agent computer systems;

a CPU;

a memory coupled to said CPU, wherein said memory stores one or more computer programs executable by the CPU, wherein said computer programs are executable to;

receive a set of data points from said one or more agent computer systems, wherein said set of data points represents a series of measurements of one or more system resources of said one or more agent computer systems;

determine whether there is a power-tail behavior in said set of data points, wherein said computer programs are further executable to;

perform two or more analytic tests on said set of data points;

combine results of said analytic tests to determine a likelihood of said power-tail behavior;

model said enterprise based on said determining whether there is a power-tail behavior in said set of data points.

26. The system of claim 25, wherein said computer programs are further executable to:

alter a configuration of said enterprise based on said determining whether there is a power-tail behavior in said set of data points.

27. The system of claim 25,

wherein in said modeling said enterprise based on said determining whether there is a power-tail behavior in said set of data points, said computer programs are further executable to:

modify a model of said enterprise.

28. The system of claim 25,

wherein in said modeling said enterprise based on said determining whether there is a power-tail behavior in said set of data points, said computer programs are further executable to:

display said likelihood of said power-tail behavior.

29. The system of claim 25,

wherein in performing said analytic tests on said set of data points, said computer programs are further executable to perform a first analytic test, a second analytic test, and a third analytic test on said set of data points;

wherein in said combining results of said analytic tests to determine a likelihood of said power-tail behavior, said computer programs are further executable to combine results of said first analytic test, said second analytic test, and said third analytic test to determine said likelihood of said power-tail behavior.

30. The system of claim 29,

wherein in performing said first analytic test on said set of data points, said computer programs are further executable to determine whether said set of data points exhibits large deviations from a mean of said data points.

31. The system of claim 29,

wherein in performing said second analytic test on said set of data points, said computer programs are further executable to determine whether said set of data points exhibits a high variance.

32. The system of claim 29,

wherein in performing said third analytic test on said set of data points, said computer programs are further executable to determine whether a set of largest data points of said set of data points exhibits properties consistent with large values in a tail portion of a power-tail distribution.

33. The system of claim 29,

wherein in performing said first analytic test on said set of data points, said computer programs are further executable to:

determine a substantially largest data point of said set of data points;

compute a probability P_D that a random variable X is greater than or equal to said substantially largest data point;

compute a probability P_E that said random variable X is greater than or equal to an expected value of a largest order statistic of an exponential distribution;

determine a tolerance factor;

determine whether said probability P_D is less than or equal to said tolerance factor;

determine whether said probability P_D is less than or equal to said probability P_E.

34. The system of claim 33,

wherein in said determining whether said probability P_D is less than or equal to said tolerance factor, said computer programs are further executable to determine that said probability P_D is less than or equal to said tolerance factor;

wherein in said determining whether said probability P_D is less than or equal to said probability P_E, said computer programs are further executable to determine that said probability P_D is less than or equal to said probability P_E;

wherein in performing said first analytic test, said computer programs are further executable to;

determine that said first analytic test is successful.

35. The system of claim 33,

wherein in said determining whether said probability P_D is less than or equal to said tolerance factor, said computer programs are further executable to determine that said probability P_D is less than or equal to said tolerance factor;

wherein said determining whether said probability P_D is less than or equal to said probability P_E, said computer programs are further executable to determine that said probability P_D is not less than or equal to said probability P_E;

wherein in performing said first analytic test, said computer programs are further executable to;

determine that said first analytic test is unsuccessful.

36. The system of claim 33,

wherein in said determining whether said probability P_D is less than or equal to said tolerance factor, said computer programs are further executable to determine that said probability P_D is not less than or equal to said tolerance factor;

wherein in said determining whether said probability P_D is less than or equal to said probability P_E, said computer programs are further executable to determine that said probability P_D is less than or equal to said probability P_E;

wherein in performing said first analytic test, said computer programs are further executable to;

determine that said first analytic test is unsuccessful.

37. The system of claim 33,

wherein in said determining whether said probability P_D is less than or equal to said tolerance factor, said computer programs are further executable to determine that said probability P_D is not less than or equal to said tolerance factor;

wherein in said determining whether said probability P_D is less than or equal to said probability P_E, said computer programs are further executable to determine that said probability P_D is not less than or equal to said probability P_E;

wherein in performing said first analytic test, said computer programs are further executable to;

determine that said first analytic test is unsuccessful.

38. The system of claim 29,

wherein in performing said second analytic test on said set of data points, said computer programs are further executable to:

compute a power-tail variance index for a power-tail distribution with a power-tail index α

of 2;

compute a variance of said set of data points;

determine whether said variance of said set of data points is greater than or equal to said power-tail variance index.

39. The system of claim 38,

wherein in said determining whether said variance of said set of data points is greater than or equal to said power-tail variance index, said computer programs are further executable to determine that said variance of said set of data points is greater than or equal to said power-tail variance index;

wherein in performing said second analytic test, said computer programs are further executable to;

determine that said second analytic test is successful.

40. The system of claim 38,

wherein in said determining whether said variance of said set of data points is greater than or equal to said power-tail variance index, said computer programs are further executable to determine that said variance of said set of data points is not greater than or equal to said power-tail variance index;

wherein in performing said second analytic test, said computer programs are further executable to;

determine that said second analytic test is unsuccessful.

41. The system of claim 29,

wherein in performing said third analytic test on said set of data points, said computer programs are further executable to:

normalize said set of data points such that an expected value of said set of data points is 1;

determine a substantially largest data point of said set of data points;

estimate a power-tail index α

of said set of data points;

determine whether said power-tail index α

of said set of data points is less than 2.

42. The system of claim 41,

wherein in said determining whether said power-tail index α

of said set of data points is less than 2, said computer programs are further executable to determine that said power-tail index α

of said set of data points is less than 2;

wherein in performing said third analytic test, said computer programs are further executable to;

determine that said third analytic test is successful.

43. The system of claim 41,

wherein in said determining whether said power-tail index α

of said set of data points is less than 2, said computer programs are further executable to determine that said power-tail index α

of said set of data points is not less than 2;

wherein in performing said third analytic test, said computer programs are further executable to;

determine that said third analytic test is unsuccessful.

44. The system of claim 29,

wherein in said determining whether there is a power-tail behavior in said set of data points, said computer programs are further executable to:

determine that said first analytic test is unsuccessful;

determine that said second analytic test is unsuccessful;

determine that said third analytic test is unsuccessful;

wherein in said combining results of said first analytic test, said second analytic test, and said third analytic test to determine said likelihood of said power-tail behavior, said computer programs are further executable to;

conclude that said power-tail behavior is unlikely.

45. The system of claim 29,

wherein in said determining whether there is a power-tail behavior in said set of data points, said computer programs are further executable to:

determine that said first analytic test is successful;

determine that said second analytic test is successful;

determine that said third analytic test is successful;

wherein in said combining results of said first analytic test, said second analytic test, and said third analytic test to determine said likelihood of said power-tail behavior, said computer programs are further executable to;

conclude that said power-tail behavior is likely.

46. The system of claim 29,

wherein in said determining whether there is a power-tail behavior in said set of data points, said computer programs are further executable to:

determine that two of said three analytic tests are unsuccessful;

determine that one of said three analytic tests is successful;

wherein in said combining results of said first analytic test, said second analytic test, and said third analytic test to determine said likelihood of said power-tail behavior, said computer programs are further executable to;

conclude that said power-tail behavior is possible.

47. The system of claim 29,

wherein in said determining whether there is a power-tail behavior in said set of data points, said computer programs are further executable to:

determine that one of said three analytic tests is unsuccessful;

determine that two of said three analytic tests are successful;

wherein in said combining results of said first analytic test, said second analytic test, and said third analytic test to determine said likelihood of said power-tail behavior, said computer programs are further executable to;

conclude that said power-tail behavior is possible.

48. The system of claim 25,

wherein said data points are independent and identically distributed;

wherein in said determining whether there is a power-tail behavior in said set of data points, said computer programs are further executable to determine whether there is a renewal power-tail behavior in said set of data points.

49. A carrier medium comprising program instructions for monitoring the performance of an enterprise, wherein the enterprise comprises one or more computer systems, wherein said program instructions are executable to implement:

receiving a set of data points from said one or more computer systems, wherein said set of data points represents a series of measurements of one or more system resources of said one or more computer systems;

determining whether there is a power-tail behavior in said set of data points, comprising;

performing two or more analytic tests on said set of data points;

combining results of said analytic tests to determine a likelihood of said power-tail behavior; and

modeling said enterprise based on said determining whether there is a power-tail behavior in said set of data points.

50. The carrier medium of claim 49, wherein said program instructions are further executable to implement:

altering a configuration of said enterprise based on said determining whether there is a power-tail behavior in said set of data points.

51. The carrier medium of claim 49,

wherein said modeling said enterprise based on said determining whether there is a power-tail behavior in said set of data points further comprises:

modifying a model of said enterprise.

52. The carrier medium of claim 49,

wherein said modeling said enterprise based on said determining whether there is a power-tail behavior in said set of data points further comprises:

displaying said likelihood of said power-tail behavior.

53. The carrier medium of claim 49,

wherein performing said analytic tests on said set of data points further comprises performing a first analytic test, a second analytic test, and a third analytic test on said set of data points;

wherein said combining results of said analytic tests to determine a likelihood of said power-tail behavior further comprises combining results of said first analytic test, said second analytic test, and said third analytic test to determine said likelihood of said power-tail behavior.

54. The carrier medium of claim 53,

wherein performing said first analytic test on said set of data points further comprises determining whether said set of data points exhibits large deviations from a mean of said data points.

55. The carrier medium of claim 53,

wherein performing said second analytic test on said set of data points further comprises determining whether said set of data points exhibits a high variance.

56. The carrier medium of claim 53,

wherein performing said third analytic test on said set of data points further comprises determining whether a set of largest data points of said set of data points exhibits properties consistent with large values in a tail portion of a power-tail distribution.

57. The carrier medium of claim 53,

wherein performing said first analytic test on said set of data points further comprises:

determining a substantially largest data point of said set of data points;

computing a probability P_D that a random variable X is greater than or equal to said substantially largest data point;

computing a probability P_E that said random variable X is greater than or equal to an expected value of a largest order statistic of an exponential distribution;

determining a tolerance factor;

determining whether said probability P_D is less than or equal to said tolerance factor;

determining whether said probability P_D is less than or equal to said probability P_E.

58. The carrier medium of claim 57,

wherein said determining whether said probability P_D is less than or equal to said tolerance factor further comprises determining that said probability P_D is less than or equal to said tolerance factor;

wherein said determining whether said probability P_D is less than or equal to said probability P_E further comprises determining that said probability P_D is less than or equal to said probability P_E;

wherein performing said first analytic test further comprises;

determining that said first analytic test is successful.

59. The carrier medium of claim 57,

wherein said determining whether said probability P_D is less than or equal to said tolerance factor further comprises determining that said probability P_D is less than or equal to said tolerance factor;

wherein said determining whether said probability P_D is less than or equal to said probability P_E further comprises determining that said probability P_D is not less than or equal to said probability P_E;

wherein performing said first analytic test further comprises;

determining that said first analytic test is unsuccessful.

60. The carrier medium of claim 57,

wherein said determining whether said probability P_D is less than or equal to said tolerance factor further comprises determining that said probability P_D is not less than or equal to said tolerance factor;

wherein said determining whether said probability P_D is less than or equal to said probability P_E further comprises determining that said probability P_D is less than or equal to said probability P_E;

wherein performing said first analytic test further comprises;

determining that said first analytic test is unsuccessful.

61. The carrier medium of claim 57,

wherein said determining whether said probability P_D is less than or equal to said tolerance factor further comprises determining that said probability P_D is not less than or equal to said tolerance factor;

wherein said determining whether said probability P_D is less than or equal to said probability P_E further comprises determining that said probability P_D is not less than or equal to said probability P_E;

wherein performing said first analytic test further comprises;

determining that said first analytic test is unsuccessful.

62. The carrier medium of claim 53,

wherein performing said second analytic test on said set of data points further comprises:

computing a power-tail variance index for a power-tail distribution with a power-tail index α

of 2;

computing a variance of said set of data points;

determining whether said variance of said set of data points is greater than or equal to said power-tail variance index.

63. The carrier medium of claim 62,

wherein said determining whether said variance of said set of data points is greater than or equal to said power-tail variance index further comprises determining that said variance of said set of data points is greater than or equal to said power-tail variance index;

wherein performing said second analytic test further comprises;

determining that said second analytic test is successful.

64. The carrier medium of claim 62,

wherein said determining whether said variance of said set of data points is greater than or equal to said power-tail variance index further comprises determining that said variance of said set of data points is not greater than or equal to said power-tail variance index;

wherein performing said second analytic test further comprises;

determining that said second analytic test is unsuccessful.

65. The carrier medium of claim 53,

wherein performing said third analytic test on said set of data points further comprises:

normalizing said set of data points such that an expected value of said set of data points is 1;

determining a substantially largest data point of said set of data points;

estimating a power-tail index α

of said set of data points;

determining whether said power-tail index α

of said set of data points is less than 2.

66. The carrier medium of claim 65,

wherein said determining whether said power-tail index α

of said set of data points is less than 2 further comprises determining that said power-tail index α

of said set of data points is less than 2;

wherein performing said third analytic test further comprises;

determining that said third analytic test is successful.

67. The carrier medium of claim 65,

wherein said determining whether said power-tail index α

of said set of data points is less than 2 further comprises determining that said power-tail index α

of said set of data points is not less than 2;

wherein performing said third analytic test further comprises;

determining that said third analytic test is unsuccessful.

68. The carrier medium of claim 53,

wherein said determining whether there is a power-tail behavior in said set of data points further comprises:

determining that said first analytic test is unsuccessful;

determining that said second analytic test is unsuccessful;

determining that said third analytic test is unsuccessful;

wherein said combining results of said first analytic test, said second analytic test, and said third analytic test to determine said likelihood of said power-tail behavior further comprises;

concluding that said power-tail behavior is unlikely.

69. The carrier medium of claim 53,

wherein said determining whether there is a power-tail behavior in said set of data points further comprises:

determining that said first analytic test is successful;

determining that said second analytic test is successful;

determining that said third analytic test is successful;

wherein said combining results of said first analytic test, said second analytic test, and said third analytic test to determine said likelihood of said power-tail behavior further comprises;

concluding that said power-tail behavior is likely.

70. The carrier medium of claim 53,

wherein said determining whether there is a power-tail behavior in said set of data points further comprises:

determining that two of said three analytic tests are unsuccessful;

determining that one of said three analytic tests is successful;

wherein said combining results of said first analytic test, said second analytic test, and said third analytic test to determine said likelihood of said power-tail behavior further comprises;

concluding that said power-tail behavior is possible.

71. The carrier medium of claim 53,

wherein said determining whether there is a power-tail behavior in said set of data points further comprises:

determining that one of said three analytic tests is unsuccessful;

determining that two of said three analytic tests are successful;

wherein said combining results of said first analytic test, said second analytic test, and said third analytic test to determine said likelihood of said power-tail behavior further comprises;

concluding that said power-tail behavior is possible.

72. The carrier medium of claim 49,

wherein said data points are independent and identically distributed;

wherein said determining whether there is a power-tail behavior in said set of data points further comprises determining whether there is a renewal power-tail behavior in said set of data points.

73. The carrier medium of claim 49,

wherein said carrier medium is a storage medium.