Identifying opportunities to improve multiprocess system performance

• US 8,340,942 B2
• Filed: 04/26/2010
• Issued: 12/25/2012
• Est. Priority Date: 04/26/2010
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Biberstein, Marina, Heilper, Andre
• Primary Examiner(s): TSAI, CAROL S W
• Application Number:
  US12/766,949
• Publication Number:
  US 20110264410A1
• Time in Patent Office: 974 Days
• Field of Search: 702/186, 702/189, 707/999.002, 707/E17.005, 340/539.12, 340/573.1, 706/11, 706/50, 705/2, 705/7.38, 705/36.R, 705/348, 705/7.36, 705/35, 703/2, 700/121, 700/28, 717/130, 711/213, 711/170, 711/150, 713/322
• US Class Current: 702/186
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G06F 11/3447 Performance evaluation by modeling
• CPC Class Codes: G06F 11/3466 Performance evaluation by tracing or monitoring

Abstract

Expert system may utilize multiple metrics to estimate whether a performance improvement activity has a potential to improve performance in respect to a predetermined performance metric. The multiple metrics may be propagated based on whether or not they may be affected by the activity and based on an associated value that may take into account the current value of the metric. An Index Propagation Graph may be further utilized to represent a propagation function which may propagate the potential of improvement of each metric in respect to the predetermined performance metric.

Claims

1. A computer-implemented method for determining an index of a potential of Performance Improvement (PI) in a computerized system by utilizing a PI activity, the method comprising:

receiving a plurality of values of a plurality of performance metrics of the computerized system, the plurality of performance metrics comprises at least a performance index metric;

determining a first portion of the plurality of performance metrics, each metric of the first portion has a potential of PI in response to performing the PI activity;

calculating the index of the potential of PI based on at least the plurality of values of metrics associated with the first portion, wherein the index is calculated in respect to the performance index metric; and

outputting an indication of the index of the potential of PI.

2. The method of claim 1, wherein the plurality of performance metrics are selected from a group consisting of CPU idleness, cycles per instruction and threads per CPU.

3. The method of claim 1, wherein the performance index metric is selected from the group consisting of throughput, latency and machine utilization.

4. The method of claim 1, wherein said receiving the plurality values of the plurality of performance metrics comprises utilizing a profiling tool to determine a value of a metric of the plurality of performance metrics.

5. The method of claim 1, wherein said calculating the index of the potential of PI comprises:

retrieving an index propagation graph, the index propagation graph having a root node associated with the performance index metric, wherein the index propagation graph having a leaf node associated with the PI activity, wherein the index propagation graph having internal nodes associated with metrics comprised by the plurality of metrics; and

calculating the index of the potential of PI based on a route in the index propagation graph.

6. The method of claim 5, wherein an edge in the index propagation graph between a source node and a target node is associated with a function based on a potential of PI in the performance index metric based on improvement of a source metric associated with the source node.

7. The method of claim 6 further comprising:

determining a second portion of the plurality of performance metrics, a second metric of the second portion will not improve in response to performing the PI activity; and

wherein the function is selected based on an affiliation of the source metric to the first portion or the second portion.

8. The method of claim 6, wherein said calculating the index of the potential of PI based on the route in the index propagation graph comprises:

multiplying values of functions associated with edges of the route, wherein the values of the functions are determined based on the plurality of values.

9. The method of claim 1 further comprising calculating multiple indices of the potential of PI in respect to a plurality of PI activities;

and selecting one of the plurality of PI activities based on the multiple indices.

10. The method of claim 9, wherein said selecting one of the plurality of PI activities is based on a comparison between the multiple indices.

11. The method of claim 9 further comprises performing the one of the plurality of PI activities.

12. A computerized apparatus for determining an index of a potential of PI in a computerized system by utilizing a PI activity, the computerized apparatus having a processor, the computerized apparatus comprising:

a receiver configured to receive a plurality of values of a plurality of performance metrics of the computerized system, the plurality of performance metrics comprises at least a performance index metric;

an index calculator configured to calculate the index of the potential of improvement in respect to the performance index metric, said index calculator comprising;

a potentially improved metric determinator configured to determine a first portion of the plurality of performance metrics, each metric of the first portion has a potential of improvement in response to the PI activity; and

an output module configured to output an indication of the index of the potential of PI.

13. The computerized apparatus of claim 12, wherein said receiver is configured to utilize a profiling tool;

the profiling tool is operative to determine a value of a metric of the plurality of performance metrics.

14. The computerized apparatus of claim 12, wherein said index calculator further comprising:

an index propagation graph module;

the index propagation graph module is configured to retain an index propagation graph, wherein the index propagation graph having a root node associated with the performance index metric, wherein the index propagation graph having a leaf node associated with the PI activity, wherein the index propagation graph having internal nodes associated with metrics comprised by the plurality of metrics; and

a route value calculator configured to calculate the index of the potential of PI based on a route in the index propagation graph retained by said index propagation graph module.

15. The computerized apparatus of claim 14 wherein said route value calculator is configured to calculate the index by multiplying values of functions associated with edges of the route, wherein the values of the functions are determined based on the plurality of values.

16. The computerized apparatus of claim 12 wherein said output module is configured to provide indications of indices of the potential of improvement based on multiple PI activities.

17. A computer program product for determining an index of a potential of Performance Improvement (PI) in a computerized system by utilizing a PI activity, the product comprising:

a non-transitory computer readable medium including computer instructions stored therein for causing a computer processor to perform;

a first program instruction for receiving a plurality of values of a plurality of performance metrics of the computerized system, the plurality of performance metrics comprises at least a performance index metric;

a second program instruction for determining a first portion of the plurality of performance metrics, each metric of the first portion has a potential of PI in response to performing the PI activity;

a third program instruction for calculating the index of the potential of PI based on at least the plurality of values of metrics associated with the first portion, wherein the index is calculated in respect to the performance index metric; and

a fourth program instruction for outputting an indication of the index of the potential of PI;

wherein said first, second, third, and fourth program instructions are stored on said computer readable medium.

18. The computer program product of claim 17 wherein the plurality of performance metrics are selected from a group consisting of CPU idleness, cycles per instruction and threads per CPU.

19. The computer program product of claim 17 wherein the performance index metric is selected from the group consisting of throughput, latency and machine utilization.

20. The computer program product of claim 17 wherein said receiving the plurality values of the plurality of performance metrics comprises utilizing a profiling tool to determine a value of a metric of the plurality of performance metrics.