Automatically optimize performance of package execution

• US 7,904,894 B2
• Filed: 03/29/2006
• Issued: 03/08/2011
• Est. Priority Date: 03/29/2006
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Garden, Euan P., Haselden, J. Kirk, Entin, Michael
• Primary Examiner(s): DAS, CHAMELI
• Application Number:
  US11/392,342
• Publication Number:
  US 20070234240A1
• Time in Patent Office: 1,805 Days
• Field of Search: 717126-128, 717130-131, 717/151, 718/102, 718/103, 718/106
• US Class Current: 717/151
• CPC Class Codes: G06F 11/3476 Data logging G06F11/14, G06F11/2205 take precedence

Abstract

Various technologies and techniques are disclosed that automatically optimize package execution performance. A profiling phase executes each task in a control flow package and measures performance metrics, such as task execution length, task memory usage, task correlation to CPU versus input/output operations, network bandwidth, and running applications. An optimization phase optimizes subsequent executions of the package by using the performance metrics to make appropriate adjustments to the package. The profiling phase and optimizing phase are repeated over a period of time as appropriate to improve the package'"'"'s performance without requiring a user to make adjustments to the tasks in the package.

Claims

1. A method for automatically optimizing scheduling of tasks in an extract-transform-load (ETL) package by a control flow application comprising the steps of:

the control flow application running an automated process for the ETL package to optimize scheduling of a plurality of tasks included in the ETL package, wherein each of the plurality of tasks is an ETL task operable to be executed sequentially or in parallel with the other tasks, the automated process comprising;

the control flow application running a profiling phase to execute each of the plurality of tasks in the ETL package at least one time and measure a plurality of performance metrics; and

the control flow application running an optimization phase to optimize subsequent executions of the plurality of tasks in the ETL package, wherein, while running the optimization phase, the control flow application uses at least a portion of the performance metrics to make at least one adjustment to the ETL package by adjusting execution of at least one of the plurality of tasks.

2. The method of claim 1, wherein the automated process allows performance of the ETL package to improve over a period of time without requiring a user to make adjustments to the tasks.

3. The method of claim 1, wherein the automated process is initiated programmatically.

4. The method of claim 1, wherein the automated process is initiated by a user.

5. The method of claim 1, wherein the plurality of performance metrics is selected from the group consisting of task execution length, task memory usage, task correlation to CPU versus input/output operations, network bandwidth, and running applications.

6. The method of claim 1, wherein the at least one adjustment includes adjusting the ETL package so at least one longer executing parallel task of the plurality of tasks is started earlier.

7. The method of claim 1, wherein the at least one adjustment includes adjusting the ETL package so a first task of the plurality of tasks that consumes a significant amount of memory is not run in parallel with a second task of the plurality of tasks that consumes the significant amount of memory.

8. The method of claim 1, wherein the at least one adjustment includes adjusting the ETL package so a number of simultaneous CPU bound tasks of the plurality of tasks does not exceed a quantity of physical processors present on a computer executing the ETL package.

9. A computer-readable storage medium having computer-executable instructions for causing a computer to perform the steps recited in claim 1.

10. A computer-readable storage medium having computer-executable instructions for causing a computer to perform steps comprising:

providing a control flow application operable to execute an extract-transform-load (ETL) control flow package;

providing a package creation feature that allows a user to create an ETL control flow package with a plurality of tasks;

providing an automatic performance improvement process that initiates a profiling phase and an optimization phase to optimize scheduling of one or more of the plurality of tasks in the ETL control flow package, wherein during the optimization phase the control flow application adjusts execution of at least one of the plurality of tasks of the ETL control flow package; and

providing a monitoring feature to monitor the performance of execution of the plurality of tasks in the ETL control flow package over a period of time and adjust an execution plan of the plurality of tasks in the ETL control flow package according to a set of data gathered using the monitoring feature.

11. The computer-readable storage medium of claim 10, wherein the profiling phase is operable to measure a plurality of performance metrics for the ETL control flow package.

12. The computer-readable storage medium of claim 10, wherein at least one of the plurality of performance metrics is selected from the group consisting of task execution length, task memory usage, task correlation to CPU versus input/output operations, network bandwidth, and running applications.

13. The computer-readable storage medium of claim 10, wherein the optimizing phase is operable to perform at least one adjustment to the ETL control flow package based on a set of profiling data gathered in the profiling phase.

14. The computer-readable storage medium of claim 13, wherein the at least one adjustment includes adjusting the ETL control flow package so at least one longer executing parallel task of the plurality of tasks is started earlier.

15. The computer-readable storage medium of claim 13, wherein the at least one adjustment includes adjusting the ETL control flow package so a first task of the plurality of tasks that consumes a significant amount of memory is not run in parallel with a second task of the plurality of tasks that consumes the significant amount of memory.

16. The computer-readable storage medium of claim 13, wherein the at least one adjustment includes adjusting the ETL control flow package so a number of simultaneous CPU bound tasks of the plurality of tasks does not exceed a quantity of physical processors present on a computer executing the ETL control flow package.

17. The computer-readable storage medium of claim 10, wherein each task of the plurality of tasks is selected from the group consisting of a sequential task and a parallel task.

18. A method for automatically optimizing package execution performance comprising the steps of:

a control flow application running a profiling phase to execute each of a plurality of tasks in an extract-transform-load (ETL) package at least one time and measure a plurality of performance metrics, wherein at least one of the plurality of performance metrics is selected from the group consisting of task execution length, task memory usage, task correlation to CPU versus input/output operations, network bandwidth, and running applications;

the control flow application running an optimization phase to optimize subsequent executions of the ETL package, wherein, while running the optimization phase, the control flow application uses at least a portion of the performance metrics to make at least one adjustment to the ETL package by adjusting execution of at least one of the plurality of tasks; and

wherein the profiling phase and optimizing phase are repeated over a period of time to improve a performance of the ETL package without requiring a user to make adjustments to execution of the tasks in the ETL package.

19. The method of claim 18, wherein the profiling phase and the optimizing phase run a first time on a first computer to improve the performance of the ETL package for the first computer, and wherein the profiling phase and the optimizing phase run a second time on a second computer to improve the performance of the ETL package for the second computer.

20. A computer-readable storage medium having computer-executable instructions for causing a computer to perform the steps recited in claim 18.