Task-based modeling for parallel data integration

Patent metadata

Field	Value
Publication number	US-9477511-B2
Application number	US-201313966903-A
Country	US
Kind code	B2
Filing date	Aug 14, 2013
Priority date	Aug 14, 2013
Publication date	Oct 25, 2016
Grant date	Oct 25, 2016

Abstract

System, method, and computer program product to perform an operation for task-based modeling for parallel data integration, by determining, for a data flow, a set of processing units, each of the set of processing units defining one or more data processing operations to process the data flow, generating a set of tasks to represent the set of processing units, each task in the set of tasks comprising one or more of the data processing operations of the set of processing units, optimizing the set of tasks based on a set of characteristics of the data flow, and generating a composite execution plan based on the optimized set of tasks to process the data flow in a distributed computing environment.

First claim

What is claimed is:

1. A system, comprising:
one or more computer processors; and
a memory containing a program which when executed by the one or more computer processors performs an operation, the operation comprising:
determining, for a data flow, a set of processing units, each processing unit of the set of processing units defining one or more data processing operations to process the data flow, wherein the determination of the set of processing units comprises:
determining a partition source to associate with each processing unit,
inserting one or more processing units into the set to automatically repartition the partition source associated with at least one of the processing units, and
optimizing virtual datasets associated with the dataflow;

generating a set of tasks to represent the set of processing units, each task in the set of tasks comprising one or more of the data processing operations of the set of processing units;
optimizing the set of tasks based on a set of characteristics of the data flow comprising one or more user defined requirements;
determining an execution order for the set of tasks based on the one or more user defined requirements;
generating a composite execution plan based on the optimized set of tasks to process the data flow in a distributed computing environment and the execution order.

2. The system of claim 1 , wherein the set of characteristics of the data flow further comprise: (i) a checkpointing requirement of the data flow, (ii) a processing requirement for each of the processing units, and (iii) analyzing one or more keys for each of the processing units.

3. The system of claim 2 , wherein analyzing the one or more keys for each of the processing units comprises:
identifying at least one key-based processing unit in the set of processing units;
identifying an upstream processing unit from the at least one key-based processing unit, wherein the upstream processing unit is not used processing any tasks of the set of tasks;
defining a map task including the upstream processing unit;
defining a reduce task using the at least one key-based processing unit; and
upon determining that the upstream processing unit and the at least one key-based processing unit have the same key requirements, combining the upstream processing unit and the at least one key-based processing unit.

4. The system of claim 1 , wherein the one or more user defined requirements comprise: (i) a number of physical partitions for processing the data flow, (ii) a number of tasks on one or more logical partitions, (iii) a maximum number of tasks across all partitions.

5. The system of claim 2 , wherein the processing requirement for each of the processing units comprises at least one of: (i) a memory requirement of each processing unit, (ii) a CPU requirement of each processing unit, and (iii) an I/O requirement of each processing unit.

6. The system of claim 1 , wherein the composite execution plan is further generated based on a determined set of processing units.

7. The system of claim 6 , wherein determining the set of processing units is based on combining at least two processing units based on one or more processing requirements of the at least two processing units.

8. A computer program product, comprising:
a computer-readable storage medium having computer-readable program code embodied therewith, the computer-readable program code comprising:
computer-readable program code configured to determine, for a data flow, a set of processing units, each processing unit of the set of processing units defining one or more data processing operations to process the data flow, wherein the determination of the set of processing units comprises:
determining a partition source to associate with each processing unit,
inserting one or more processing units into the set to automatically repartition the partition source associated with at least one of the processing units, and
optimizing virtual datasets associated with the dataflow;

computer-readable program code configured to generate a set of tasks to represent the set of processing units, each task in the set of tasks comprising one or more of the data processing operations of the set of processing units;
computer-readable program code configured to optimize the set of tasks based on a set of characteristics of the data flow comprising one or more user defined requirements;
computer-readable program code configured to determine an execution order for the set of tasks based on the one or more user defined requirements;
computer-readable program code configured to generate a composite execution plan based on the optimized set of tasks to process the data flow in a distributed computing environment and the execution order.

9. The computer program product of claim 8 , wherein the set of characteristics of the data flow further comprise: (i) a checkpointing requirement of the data flow, (ii) a processing requirement for each of the processing units, and (iii) analyzing one or more keys for each of the processing units.

10. The computer program product of claim 9 , wherein analyzing the one or more keys for each of the processing units comprises:
identifying at least one key-based processing unit in the set of processing units;
identifying an upstream processing unit from the at least one key-based processing unit, wherein the upstream processing unit is not used processing any tasks of the set of tasks;
defining a map task including the upstream processing unit;
defining a reduce task using the at least one key-based processing unit; and upon determining that the upstream processing unit and the at least one key-based processing unit have the same key requirements, combining the upstream processing unit and the at least one key-based processing unit.

11. The computer program product of claim 8 , wherein the one or more user defined requirements comprise: (i) a number of physical partitions for processing the data flow, (ii) a number of tasks on one or more logical partitions, (iii) a maximum number of tasks across all partitions.

12. The computer program product of claim 9 , wherein the processing requirement for each of the processing units comprises at least one of: (i) a memory requirement of each processing unit, (ii) a CPU requirement of each processing unit, and (iii) an I/O requirement of each processing unit.

13. The computer program product of claim 8 , wherein the composite execution plan is further generated based on a determined set of processing units, wherein determining the set of processing units is based on combining at least two processing units based on one or more processing requirements of the at least two processing units.

Assignees

• IBM

Inventors

• Jacobson Eric A
• Li Yong
• Mudambi Shyam R
• Pu Xiaoyan

Classifications

• G06F9/46 Primary

  Multiprogramming arrangements · CPC title

• G06F9/5066 Primary

  Algorithms for mapping a plurality of inter-dependent sub-tasks onto a plurality of physical CPUs (mappping at compile time, see G06F8/451) · CPC title