Operational leading indicator (OLI) management using in-memory database

• US 9,619,769 B2
• Filed: 04/01/2014
• Issued: 04/11/2017
• Est. Priority Date: 04/01/2014
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: SAP SE
• Inventors: Ploesser, Karsten, Chandra, Piyush, Gill, Keith, Jama, Farah
• Primary Examiner(s): Gart, Matthew
• Assistant Examiner(s): Holzmacher, Derick
• Application Number:
  US14/242,074
• Publication Number:
  US 20150278738A1
• Time in Patent Office: 1,106 Days
• Field of Search: 705 711, 705 35, 705 28, 707602, 370231
• CPC Class Codes: G06F 16/28 Databases characterised by their database models, e.g. relational or object models
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: G06Q 10/06315 Needs-based resource requirements planning or analysis

Abstract

Embodiments leverage capabilities of an in-memory database to manage measurement and modeling of Operational Leading Indicators (OLIs). An OLI template may comprise four model components: ratio calculation; factor calculation; predictive modeling; and cost estimation. Each model component is motivated and explained in terms of information sources, and analytical or statistical modelling tasks used in its definition. Embodiments combine analytical and statistical modelling utilizing in-memory computing, to process large amounts of unmodified source data, calculate cost measures rapidly without preaggregation, and/or run linear regression models on the same data set and in the same memory space without a need for separate hardware. An engine in communication with the in-memory database that comprises a large volume of available data, is configured to receive values for OLI factors as inputs. In response, the engine is configured to process these inputs according to the modeling template to provide corresponding cost measures as outputs.

Claims

1. A computer-implemented method comprising:

extracting performance data of a process from a data object of an application to an in-memory database of a computer system, wherein the performance data is extracted to the in-memory database in real time from unmodified source data without preaggregation;

storing the performance data in the in-memory database;

performing statistical analysis on the performance data to generate a model of at least one operational leading indicator (OLI) of the process, the model comprising calculation of a ratio without an index artifact, the ratio evaluating a first condition comprising a lock indicator of the data object in the application, and a second condition evaluating a threshold absolute balance of the data object over a defined number of posting periods, wherein the statistical analysis is performed using in-memory computing directly on the performance data stored in the in-memory database without requiring separate computer hardware;

storing the model of the OLI in the in-memory database;

causing an in-memory database engine in communication with the model to receive inputs for performance values and performance measures of the process;

causing the in-memory database engine to process an output of the statistical analysis of the model in order to produce a cost measure;

performing, by the computer system, ongoing measurements of operational performance of the process from the output of the statistical analysis of the model;

determining, by the computer system, variations in the operational performance of the process in response to the ongoing measurements; and

implementing, by the computer system, corrective action based on the variations in the operational performance of the process.

2. A method as in claim 1 wherein the in-memory database engine conducts predictive modeling correlating a measure of process inefficiency with the performance data to predict the cost measure.

3. A method as in claim 2 wherein the in-memory database engine further conducts operational modeling identifying variables affecting the process.

4. A method as in claim 1 further comprising causing the in-memory database engine to perform a what-if simulation based upon a new value for performance factors specified by a user.

5. A method as in claim 1 further comprising causing the in-memory database engine to merge the model of the OLI with a model of a related OLI in order to create a composite index.

6. A method as in claim 1 wherein the statistical analysis comprises linear regression.

7. A non-transitory computer readable storage medium embodying a computer program for performing a method, said method comprising:

extracting performance data of a process from a data object of an application to an in-memory database of a computer system, wherein the performance data is extracted to the in-memory database in real time from unmodified source data without preaggregation;

storing the performance data in the in-memory database;

performing statistical analysis on the performance data to generate a model of at least one operational leading indicator (OLI) of the process, the model comprising calculation of a ratio without an index artifact, the ratio evaluating a first condition comprising a lock indicator of the data object in the application, and a second condition evaluating a threshold absolute balance of the data object over a defined number of posting periods, wherein the statistical analysis is performed using in-memory computing directly on the performance data stored in the in-memory database without requiring separate computer hardware;

storing the model of the OLI in the in-memory database;

causing an in-memory database engine in communication with the model to receive inputs for performance values and performance measures of the process;

causing the in-memory database engine to process an output of the statistical analysis of the model in order to produce a cost measure;

performing, by the computer system, ongoing measurements of operational performance of the process from the output of the statistical analysis of the model;

determining, by the computer system, variations in the operational performance of the process in response to the ongoing measurements; and

implementing, by the computer system, corrective action based on the variations in the operational performance of the process.

8. A non-transitory computer readable storage medium as in claim 7 wherein the in-memory database engine conducts predictive modeling correlating a measure of process inefficiency with the performance data to predict the cost measure.

9. A non-transitory computer readable storage medium as in claim 8 wherein the in-memory database engine further conducts operational modeling identifying variables affecting the process.

10. A non-transitory computer readable storage medium as in claim 7 wherein the method further comprises causing the in-memory database engine to perform a what-if simulation based upon a new value for performance factors specified by a user.

11. A non-transitory computer readable storage medium as in claim 7 wherein the method further comprises causing the in-memory database engine to merge the model of the OLI with a model of a related OLI in order to create a composite index.

12. A non-transitory computer readable storage medium as in claim 7 wherein the statistical analysis comprises linear regression.

13. A computer system comprising:

one or more processors;

a software program, executable on said computer system, the software program configured to;

extract performance data of a process from a data object of an application to an in-memory database of the computer system, wherein the performance data is extracted to the in-memory database in real time from unmodified source data without preaggregation;

store the performance data in the in-memory database;

perform statistical analysis on the performance data to generate a model of at least one operational leading indicator (OLI) of the process, the model comprising calculation of a ratio without an index artifact, the ratio evaluating a first condition comprising a lock indicator of the data object in the application, and a second condition evaluating a threshold absolute balance of the data object over a defined number of posting periods, wherein the statistical analysis is performed using in-memory computing directly on the performance data stored in the in-memory database without requiring separate computer hardware;

store the model of the OLI in the in-memory database;

cause an in-memory database engine in communication with the model to receive inputs for performance values and performance measures of the process;

cause the in-memory database engine to process an output of the statistical analysis of the model in order to produce a cost measure;

performing ongoing measurements of operational performance of the process from the output of the statistical analysis of the model;

determining variations in the operational performance of the process in response to the ongoing measurements; and

implementing corrective action based on the variations in the operational performance of the process.

14. A computer system as in claim 13 wherein the in-memory database engine conducts predictive modeling correlating a measure of process inefficiency with the performance data to predict the cost measure.

15. A computer system as in claim 14 wherein the in-memory database engine is caused to further conduct operational modeling identifying variables affecting the process.

16. A computer system as in claim 13 wherein the software program is further configured to cause the in-memory database engine to perform a what-if simulation based upon a new value for performance factors specified by a user.

17. A computer system as in claim 13 wherein the software program is further configured to cause the in-memory database engine to merge the model of the OLI with a model of a related OLI in order to create a composite index.