Method and system for automatically interpreting computer system performance measurements

• US 7,016,809 B1
• Filed: 09/26/2000
• Issued: 03/21/2006
• Est. Priority Date: 09/26/2000
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Garg, Neeraj, Gotwals, Jacob K.
• Primary Examiner(s): Hoff, Marc S.
• Assistant Examiner(s): Kim, Paul L
• Application Number:
  US09/669,604
• Time in Patent Office: 2,002 Days
• Field of Search: 702/183, 702/185, 702/186, 714/25, 714/30, 714/26, 714/48, 700/26
• US Class Current: 702/185
• CPC Class Codes: G06F 11/2257 using expert systems
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G06F 11/3447 Performance evaluation by modeling
• CPC Class Codes: G06F 11/3612 by runtime analysis performance monitoring G06F11/3466
• CPC Class Codes: G06F 2201/88 Monitoring involving counting

Abstract

An improvement in applying and interpreting encoded knowledge about computer system performance engineering and in enabling a user to more readily understand and apply that knowledge to improve computer system performance. Textual and graphical depictions of insights, advice, behaviors, and explanations can be presented, all related to improving computer system performance. Fuzzy logic can be employed to accommodate incomplete measurement data and to present results in a qualitative format, presenting relevant data as being high, low, substantial, etc., rather than simply presenting the values of the associated counters leading to those determinations. Intermediate values used to generate the results can also be processed with qualitative rather than quantitative values.

Claims

1. A method, comprising:

identifying one or more behavioral properties based on measured behavioral data in a computer system, each behavioral property including a behavioral description describing the respective behavioral property and a value representing a degree of relevancy of the respective behavioral property with respect to the measured behavioral data;

generating an analysis based on the one or more behavioral properties, the analysis including an analysis description indicating a result expected to happen based on a combination of the identified behavioral properties; and

determining an advice associated with the analysis, the advice suggesting what should be done in response to the expected result described by the analysis.

2. The method of claim 1, further comprising encoding knowledge of computer subsystem behavioral properties based on behavioral data to provide a basis for identifying the one or more behavioral properties.

3. The method of claim 2, wherein the encoding knowledge of computer system behavioral properties comprises:

associating one or more of the behavioral properties to each behavior, each of the behavioral properties including a description regarding the respective behavioral property;

associating one or more property values to each of the behavioral properties; and

calculating a behavior value based on the measured behavioral data using a mathematical formula to determine a relevancy of the behavioral properties.

4. The method of claim 3, further comprising comparing the behavior value with the one or more property values to determine which of one or more behavioral properties corresponding to the property values are relevant, wherein a proximity of the behavior value with respect to a behavioral property indicates a degree whether the respective behavioral property should be relevant.

5. The method of claim 1, further comprising encoding knowledge of analysis based on the behavioral properties to provide a basis for determining the analysis.

6. The method of claim 5, wherein the encoding knowledge of analysis comprises:

associating one or more of the analysis properties to each analysis, each of the analysis properties including an analysis description regarding the respective analysis; and

generating a weight list of advices describing actions that should be taken, wherein the advice is generated based on the determined analysis.

7. The method of claim 6, further comprising:

generating a list of constraints describing one or more conditions under which the respective analysis should be excluded from consideration; and

generating a list of weighted behavioral properties identifying the one or more conditions under which the respective analysis is relevant, wherein the advice is determined when the respective analysis is relevant.

8. The method of claim 7, wherein a weight of each behavioral properties indicates a relationship with other behavioral properties.

9. The method of claim 7, wherein the advice includes an advice description describing actions that should be taken and a list of constraints describing one or more conditions under which the respective advice should not be given.

10. The method of claim 7, further comprising determining a relevancy of the analysis based on contribution of each behavior associated with the analysis based on a proximity of the measured behavioral data with respect to one or more behavioral properties associated with the measured behavioral data.

11. The method of claim 1, further comprising generating an explanation of the analysis using in part the behavioral descriptions of the one or more behavioral properties when the corresponding behavioral properties are relevant based on the values associated with the respective behavioral properties.

12. The method of claim 1, further comprising generating an explanation of the advice using in part the behavioral descriptions and the analysis description.

13. The method of claim 1, wherein said identifying one or more behavioral properties includes identifying behavior whose value is outside a predefined range.

14. The method of claim 1, further comprising providing the analysis and the advice to a user, including displaying the analysis description and the advice.

15. A machine-readable medium having stored thereon instructions, which when executed by a processor can cause said processor to perform:

identifying one or more behavioral properties based on measured behavioral data in a computer system, each behavioral property including a behavioral description describing the respective behavioral property and a value representing a degree of relevancy of the respective behavioral property with respect to the measured behavioral data;

generating an analysis based on the one or more behavioral properties, the analysis including an analysis description indicating a result expected to happen based on a combination of the identified behavioral properties; and

determining an advice associated with the analysis, the advice suggesting what should be done in response to the expected result described by the analysis.

16. The medium of claim 15, further comprising encoding knowledge of computer subsystem behavioral properties based on behavioral data to provide a basis for identifying the one or more behavioral properties.

17. The medium of claim 15, further comprising encoding knowledge of analysis based on the behavioral properties to provide a basis for determining the analysis.

18. The medium of claim 15, further comprising generating an explanation of the analysis using in part the behavioral descriptions of the one or more behavioral properties when the corresponding behavioral properties are relevant based on the values associated with the respective behavioral properties.

19. The medium of claim 15, further comprising generating an explanation of the advice using in part the behavioral descriptions and the analysis description.

20. The medium of claim 15, wherein said identifying one or more behavioral properties includes identifying behavior whose value is outside a predefined range.

21. The medium of claim 15, further comprising providing the analysis and the advice to a user including displaying the analysis description describing a problem about to happen and the advice indicating what should be done with respect to the problem.

22. A system, comprising:

a measurement interface to receive measured values of computer system performance parameters;

a measurement interpreter coupled to the measurement interface to receive the measured values from the measurement interface and to provide behavioral property data, the behavioral property data including a description for describing the respective behavioral property;

a behavior interpreter coupled to the measurement interpreter to receive the behavioral property data and to provide analysis and explanations, each of the analysis and explanations including a description describing the respective analysis and the explanation; and

a user interface coupled to the behavior interpreter to present the analysis and explanations to a user using in part the descriptions of the behavioral property data, analysis, and explanations.

23. The system of claim 22, wherein the measurement interface includes a measurement adapter.

24. The system of claim 22, wherein the measurement interpreter includes a behavior knowledge interpreter.

25. The system of claim 22, wherein the measurement interpreter includes a compiler.

26. The system of claim 22, wherein the behavior interpreter includes an analysis generator.

27. The system of claim 26, wherein the analysis generator includes an advice generator.