AUTOMATIC IDENTIFICATION OF BOTTLENECKS USING RULE-BASED EXPERT KNOWLEDGE

• US 20120054472A1
• Filed: 09/01/2010
• Published: 03/01/2012
• Est. Priority Date: 03/25/2010
• Status: Active Grant

• Current Assignee: DoorDash, Inc.
• Original Assignee: International Business Machines Corporation
• Inventors: Altman, Erik R., Arnold, Matthew R., Mitchell, Nicholas M.
• Granted Patent: US 8,627,317 B2
• Time in Patent Office: Days
• US Class Current: 712/219
• CPC Class Codes: G06F 11/3419 by assessing time
• CPC Class Codes: G06F 11/3476 Data logging G06F11/14, G06F11/2205 take precedence

Abstract

Execution states of tasks are inferred from collection of information associated with runtime execution of a computer system. Collection of information may include infrequent samples of executing tasks, the samples which may provide inaccurate executing states. One or more tasks may be aggregated by one or more execution states for determining execution time, idle time, or system policy violations, or combinations thereof.

Claims

1. A method for automatically identifying computer system performance bottlenecks and/or excess idle time, comprising:

collecting information associated with runtime execution of a computer system;

inferring, using a processor, one or more execution states of one or more tasks within the system.

2. The method of claim 1, wherein the collected information includes infrequent samples of executing tasks or the collected information specifies inaccurate execution states or combinations thereof.

3. The method of claim 1, wherein the collected information includes full traces of executing tasks.

4. The method of claim 1, wherein the step of inferring applies rule-based expert knowledge to infer the one or more execution states.

5. The method of claim 4, wherein the rule-based expert knowledge maps names of tasks to an execution state.

6. The method of claim 4, wherein the rule-based expert knowledge maps names of internal context of tasks to an execution state.

7. The method of claim 1, further including aggregating said one or more tasks by said one or more execution states for determining execution time, idle time, or system policy violations, or combinations thereof.

8. The method of claim 1, wherein the inferred execution states are further, hierarchically, subcategorized into one or more categories based on common execution tasks.

9. The method of claim 8, wherein the one or more categories include one or more of:

a category indicating that an executing task is delayed by waiting on one or more requests the executing task has made to an external service provider,

a category indicating that an executing task is delayed by waiting on one or more requests the t executing task has made to local disks,

a category indicating that an executing task is delayed by waiting on data from a client machine

a category indicating that an executing task is ready and waiting for one or more new requests to the service that the executing task provides,

a category indicating that an executing task is waiting for one or more sub-tasks that are executing in one or more other executing tasks to complete,

a category indicating that an executing task is currently executing work,

a category indicating that an executing task is unable to proceed because machine resources are being used by other system facilities,

a category indicating that an executing task is unable to proceed because machine resources are being used by another set of tasks unrelated to the executing task, or

combinations thereof.

10. The method of claim 1, wherein the collected information include one or more samples of thread stacks, process stacks, or resource utilization, or combinations thereof.

11. The method of claim 1, wherein the one or more tasks include a thread, a process or combinations thereof.

12. The method of claim 1, wherein the automatically categorizing is performed by applying one or more rules based on method names found in the samples, the sample including at least thread stacks.

13. The method of claim 1, wherein the information is collected without restarting an application or altering its command-line invocation or combinations thereof.

14. A system for automatically identifying computer system performance bottlenecks using rule-based expert knowledge, comprising:

a processor;

a module operable to collect information associated with runtime execution of a computer system and infer one or more execution states of one or more tasks within the system.

15. The system of claim 14, further including:

an application interface operable to provide the summarized execution activity to a user.

16. The system of claim 15, further including:

a user interface operable to provide the summarized execution activity to the user.

17. The system of claim 14, wherein the collected information includes infrequent samples of executing tasks.

18. The system of claim 14, wherein the collected information includes full traces of executing tasks.

19. The system of claim 14, wherein the step of inferring applies rule-based expert knowledge to infer the one or more execution states.

20. A computer readable storage medium storing a program of instructions executable by a machine to perform a method of automatically identifying computer system performance bottlenecks using rule-based expert knowledge, comprising:

collecting information associated with runtime execution of a computer system; and

inferring one or more execution states of one or more tasks within the system.

21. The computer readable storage medium of claim 20, wherein the collected information includes infrequent samples of executing tasks.

22. The computer readable storage medium of claim 20, wherein the collected information includes full traces of executing tasks.

23. The computer readable storage medium of claim 20, wherein the step of inferring applies rule-based expert knowledge to infer the one or more execution states.

24. The computer readable storage medium of claim 20, wherein the rule-based expert knowledge maps names of tasks to an execution state.

25. The computer readable storage medium of claim 20, further including:

aggregating said one or more tasks by said one or more execution states for determining execution time, idle time, or system policy violations, or combinations thereof.