Relationships derived from trace data

• US 9,665,474 B2
• Filed: 03/29/2013
• Issued: 05/30/2017
• Est. Priority Date: 03/15/2013
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Li, Ying, Gounares, Alexander G., Garrett, Charles D., Krajec, Russell S.
• Primary Examiner(s): Bullock, Jr., Lewis A
• Assistant Examiner(s): Huynh, Tina
• Application Number:
  US13/853,816
• Publication Number:
  US 20130219057A1
• Time in Patent Office: 1,523 Days
• CPC Class Codes: G06F 11/1471 involving logging of persistent data for recovery
• CPC Class Codes: G06F 11/323 Visualisation of programs or trace data
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G06F 11/3414 Workload generation, e.g. scripts, playback
• CPC Class Codes: G06F 11/3466 Performance evaluation by tracing or monitoring
• CPC Class Codes: G06F 11/3495 for systems
• CPC Class Codes: G06F 11/3612 by runtime analysis performance monitoring G06F11/3466
• CPC Class Codes: G06F 11/3624 by performing operations on the source code, e.g. via a compiler
• CPC Class Codes: G06F 11/3636 by tracing the execution of the program
• CPC Class Codes: G06F 11/3688 for test execution, e.g. scheduling of test suites
• CPC Class Codes: G06F 2201/805 Real-time
• CPC Class Codes: G06F 2201/84 Using snapshots, i.e. a logical point-in-time copy of the data
• CPC Class Codes: G06F 2201/865 Monitoring of software
• CPC Class Codes: G06F 2201/88 Monitoring involving counting
• CPC Class Codes: G06F 8/443 Optimisation
• CPC Class Codes: G06F 8/70 Software maintenance or management
• CPC Class Codes: G06F 9/46 Multiprogramming arrangements
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: H04L 43/00 Arrangements for monitoring or testing data switching networks

Abstract

An analysis system may perform network analysis on data gathered from an executing application. The analysis system may identify relationships between code elements and use tracer data to quantify and classify various code elements. In some cases, the analysis system may operate with only data gathered while tracing an application, while other cases may combine static analysis data with tracing data. The network analysis may identify groups of related code elements through cluster analysis, as well as identify bottlenecks from one to many and many to one relationships. The analysis system may generate visualizations showing the interconnections or relationships within the executing code, along with highlighted elements that may be limiting performance.

Claims

1. A method performed by a computer processor, said method comprising:

receiving trace data gathered by tracing an application;

analyzing said trace data to identify a plurality of code elements, including at least a first code element;

identifying a plurality of directional relationships among said code elements, said directional relationships being identified as occurring between two of said code elements and being identified as having a directionality;

identifying one or more patterns of the code elements, based at least in part on the plurality of directional relationships, the one or more patterns being identified comprising at least one of a star pattern, a pipeline pattern, a parallelism pattern or a cluster pattern, and including at least the first code element;

determining a performance metric for each of said code elements;

determining a characterization for each of said directional relationships;

analyzing said plurality of directional relationships to identify and distinguished between any laggard elements of the plurality of code elements that contribute to a performance bottleneck of the application and any patient elements of the plurality of code elements that do not contribute to a performance bottleneck of the application; and

for at least one laggard element, identifying a laggard optimization setting for increasing resource allocation to or prioritization of the at least one laggard element, while also identifying for at least one patient element, a patient optimization setting for decreasing resource allocation to or prioritization of the at least one patient element, at least one of the laggard optimization setting or the patient optimization setting being operable to improve overall application performance.

2. The method of claim 1, wherein the one or more patterns include an outbound star pattern, said outbound star pattern being defined with relationships from said first code element to each of a first plurality of code elements.

3. The method of claim 1, wherein the one or more patterns include an inbound star pattern, said inbound star pattern being defined with relationships to said first code element from each of a first plurality of code elements.

4. The method of claim 1, wherein the one or more patterns include the star pattern comprising the first code element as a hub node and a plurality of code elements being spoke nodes on said star pattern that are connected to the first code element.

5. The method of claim 4, said first code element being the at least one laggard element which contributes to the performance bottleneck of said application.

6. The method of claim 1, said characterization of each of said directional relationships being determined at least in part from a number of communications.

7. The method of claim 6, said characterization of each of said directional relationships being determined at least in part from a frequency of communications.

8. The method of claim 7, said characterization of each of said directional relationships being determined at least in part from a size of data passed within said communications.

9. The method of claim 1, said first code element being one of a series of code elements in the pipeline pattern.

10. The method of claim 9, said first code element being downstream from a second code element and upstream from a third code element.

11. The method of claim 10, said first code element being the only code element connecting said second code element to said third code element.

12. The method of claim 1, said first code element being identified as the at least one laggard element which contributes to the performance bottleneck when said tracing data indicate that said first code element entered a wait state for a period of time exceeding a predefined threshold.

13. A system comprising:

a processor;

an analyzer operating on said processor, said analyzer that;

receiving trace data gathered by tracing an application;

analyzing said trace data to identify a plurality of code elements, including at least a first code element;

identifying a plurality of directional relationships among said code elements, said directional relationships being identified as occurring between two of said code elements and being identified as having a directionality;

identifying one or more patterns of the code elements, based at least in part on the plurality of directional relationships, the one or more patterns being identified at least one of a star pattern, a pipeline pattern, a parallelism pattern or a cluster pattern, and including at least the first code element;

determining a performance metric for each of said code elements;

determining a characterization for each of said directional relationships;

analyzing said plurality of directional relationships to identify and distinguished between any laggard elements of the plurality of code elements that contribute to a performance bottleneck of the application and any patient elements of the plurality of code elements that do not contribute to a performance bottleneck of the application; and

an optimizer that identifies at least one of;

for at least one laggard element, identifying a laggard optimization setting for increasing resource allocation to or prioritization of the at least one laggard element, or

for at least one patient element, a patient optimization setting for decreasing resource allocation to or prioritization of the at least one patient element, at least one of the laggard optimization setting or the patient optimization setting being operable to improve overall application performance.

14. The system of claim 13, said analyzer that further:

displays a bottleneck annotation on a user interface.

15. The system of claim 14, said analyzer that further:

displays a bottleneck annotation within a graph.

16. The system of claim 13, said trace data gathered on a second processor.

17. The system of claim 13, said trace data gathered on said processor.

18. The system of claim 13, wherein the analyzer identifies said at least one laggard element and said at least one patient element in the plurality of code elements.

19. The system of claim 18, wherein the optimizer identifies the laggard optimization setting and the patient optimization setting.

20. The system of claim 19, wherein the system applies at least one of the laggard optimization setting and the patient optimization setting.