Selecting rules engines for processing abstract rules based on functionality and cost

• US 8,214,351 B2
• Filed: 04/16/2007
• Issued: 07/03/2012
• Est. Priority Date: 04/16/2007
• Status: Active Grant

• Current Assignee: ServiceNow Incorporated
• Original Assignee: International Business Machines Corporation
• Inventors: Dettinger, Richard D., Kulack, Frederick A., Wang, Xueyun S., Wenzel, Shannon E.
• Primary Examiner(s): Trujillo, James
• Assistant Examiner(s): Singh, Amresh
• Application Number:
  US11/735,532
• Publication Number:
  US 20080256047A1
• Time in Patent Office: 1,905 Days
• Field of Search: 707/4, 707/713
• US Class Current: 707/713
• CPC Class Codes: G06F 16/90335 Query processing

Abstract

Embodiments of the invention provide techniques for selecting rule engines for processing abstract rules based on functionality and cost. In general, an abstract rule is analyzed to determine which functions are required to process the rule. The abstract rule is assigned to a rule engine by evaluating metadata describing the functions and costs of the rule engines. The abstract rule is then translated to the format required by the selected rule engine.

Claims

1. A computer-implemented method comprising:

receiving an abstract rule having a conditional statement and a consequential statement, wherein the consequential statement defines a particular recommendation that is returned when the conditional statement is satisfied, and wherein the conditional statement and the consequential statement are defined using logical field definitions defined in an abstraction model that models underlying physical data in a manner making a schema of the physical data transparent to a user of the abstraction model;

determining one or more functions required to evaluate whether a rule predicate of the conditional statement is satisfied;

determining one or more rule engines of a plurality of rule engines that include the required functions;

selecting, from the one or more rule engines determined to include the required functions, a rule engine to process the abstract rule, wherein selecting a rule engine to process the abstract rule comprises determining the rule engine having a lowest cost for processing the required functions; and

executing the selected rule engine to process the abstract rule and produce an output.

2. The method of claim 1, wherein executing the selected rule engine to process the abstract rule and produce an output comprises:

transforming the abstract rule into an executable rule;

processing the executable rule in the selected rule engine; and

if the conditional statement is resolved to true for the processed executable rule, returning the particular recommendation.

3. The method of claim 2, wherein processing the executable rule comprises:

selecting one or more field values from a query result set.

4. The method of claim 2, wherein processing the executable rule comprises:

executing a query against a database;

receiving a result set for the executed query; and

determining, from the result set, one or more inputs against which the conditional statement is resolved, wherein the executable rule includes the one or more inputs.

5. The method of claim 2, wherein transforming the abstract rule into an executable rule comprises:

retrieving a specification of the abstract rule in a first computer-readable language; and

transforming the specification into a language that is accepted by the selected rules engine.

6. The method of claim 1, wherein determining one or more rule engines that include the required functions comprises evaluating metadata describing properties of the plurality of rule engines.

7. The method of claim 6, wherein selecting a rule engine to process the abstract rule further comprises:

excluding, from the one or more rule engines determined to include the required functions, any rule engine having an expected performance that is below a pre-determined minimum.

8. The method of claim 1, wherein the lowest cost is a lowest monetary cost.

9. A computer-implemented method comprising:

receiving an abstract rule having a conditional statement and a consequential statement, wherein the consequential statement defines a particular recommendation that is returned when the conditional statement is satisfied, and wherein the conditional statement and the consequential statement are defined using logical field definitions defined in an abstraction model that models underlying physical data in a manner making a schema of the physical data transparent to a user of the abstraction model;

determining one or more functions required to evaluate whether a rule predicate of the conditional statement is satisfied;

determining one or more rule engines of a plurality of rule engines that include the required functions;

grouping the required functions into two or more groupings;

selecting, from the one or more rule engines determined to include the required functions, a rule engine to process each grouping of required functions; and

executing the selected rule engines to produce an output, wherein each selected rule engine processes the abstract rule according to the corresponding grouping of required functions.

10. The method of claim 9, wherein executing the selected rule engines to produce an output comprises:

transforming each grouping of required functions into an executable sub-rule;

ordering the executable sub-rules in a sequence so that the output resulting from processing each sub-rule is used as an input for processing the next sub-rule, wherein the output resulting from processing the final sub-rule is the same as the output resulting from processing the abstract rule as a whole by a single rule engine;

processing each executable sub-rule in the corresponding rule engine in accordance with the sequence;

passing the output resulting from processing each executable sub-rule as an input for processing the next executable sub-rule; and

if the conditional statement is resolved to true for the output resulting from processing the last executable sub-rule, returning the particular recommendation.

11. The method of claim 9, wherein selecting a rule engine to process each grouping of required functions comprises:

determining the rule engine having a lowest monetary cost for processing the required functions.

12. The method of claim 9, wherein determining one or more rule engines that include the required functions comprises evaluating metadata describing properties of the plurality of rule engines.

13. A computer readable storage medium containing a program which, when executed, performs an operation comprising:

receiving an abstract rule having a conditional statement and a consequential statement, wherein the consequential statement defines a particular recommendation that is returned when the conditional statement is satisfied, and wherein the conditional statement and the consequential statement are defined using logical field definitions defined in an abstraction model that models underlying physical data in a manner making a schema of the physical data transparent to a user of the abstraction model;

determining one or more functions required to evaluate whether a rule predicate of the conditional statement is satisfied;

determining one or more rule engines of a plurality of rule engines that include the required functions;

selecting, from the one or more rule engines determined to include the required functions, a rule engine to process the abstract rule, wherein selecting a rule engine to process the abstract rule comprises determining the rule engine having a lowest cost for processing the required functions; and

executing the selected rule engine to process the abstract rule and produce an output.

14. The computer readable medium of claim 13, wherein executing the selected rule engine to process the rule and produce an output comprises:

transforming the abstract rule into an executable rule;

processing the executable rule in the selected rule engine; and

if the conditional statement is resolved to true for the processed executable rule, returning the particular recommendation.

15. The computer readable medium of claim 14, wherein processing the executable rule comprises:

selecting one or more field values from a query result set.

16. The computer readable medium of claim 14, wherein processing the executable rule comprises:

executing a query against a database;

receiving a result set for the executed query; and

determining, from the result set, one or more inputs against which the conditional statement is resolved, wherein the executable rule includes the one or more inputs.

17. The computer readable medium of claim 14, wherein transforming the abstract rule into an executable rule comprises:

retrieving a specification of the abstract rule in a first computer-readable language; and

transforming the specification into a language that is accepted by the selected rules engine.

18. The computer readable medium of claim 13, wherein determining one or more rule engines that include the required functions comprises evaluating metadata describing properties of the plurality of rule engines.

19. The computer readable medium of claim 18, wherein selecting a rule engine to process the abstract rule further comprises:

excluding, from the one or more rule engines determined to include the required functions, any rule engine having an expected performance that is below a pre-determined minimum.

20. The computer readable storage medium of claim 13, wherein the lowest cost is a lowest monetary cost.

21. A computer readable storage medium containing a program which, when executed, performs an operation comprising:

receiving an abstract rule having a conditional statement and a consequential statement, wherein the consequential statement defines a particular recommendation that is returned when the conditional statement is satisfied, and wherein the conditional statement and the consequential statement are defined using logical field definitions defined in an abstraction model that models underlying physical data in a manner making a schema of the physical data transparent to a user of the abstraction model;

determining one or more functions required to evaluate whether a rule predicate of the conditional statement is satisfied;

determining one or more rule engines of a plurality of rule engines that include the required functions;

grouping the functions required to evaluate the conditional statement into two or more groupings;

selecting, from the one or more rule engines determined to include the required functions, a rule engine to process each grouping of required functions; and

executing the selected rule engines to produce an output, wherein each selected rule engine processes the rule according to the corresponding grouping of required functions.

22. The computer readable medium of claim 21, wherein executing the selected rule engines to produce an output comprises:

transforming each grouping of required functions into an executable sub-rule;

ordering the executable sub-rules in a sequence so that the output resulting from processing each sub-rule is used as an input for processing the next sub-rule, wherein the output resulting from processing the final sub-rule is the same as the output resulting from processing the abstract rule as a whole by a single rule engine;

processing each executable sub-rule in the corresponding rule engine in accordance with the sequence;

passing the output resulting from processing each executable sub-rule as an input for processing the next executable sub-rule; and

if the conditional statement is resolved to true for the output resulting from processing the last executable sub-rule, returning the particular recommendation.

23. The computer readable medium of claim 21, wherein selecting a rule engine to process each grouping of required functions comprises:

determining the rule engine having a lowest monetary cost for processing the grouping of required functions.

24. The computer readable medium of claim 21, wherein determining one or more rule engines that include the required functions comprises evaluating metadata describing properties of the plurality of rule engines.