SYSTEM FOR PARTITIONING BATCH PROCESSES

• US 20140373016A1
• Filed: 05/23/2014
• Published: 12/18/2014
• Est. Priority Date: 06/08/2011
• Status: Active Grant

• Current Assignee: Workday Incorporated
• Original Assignee: Workday Incorporated
• Inventors: Quiroga, Salvador Maiorano, Chan, Kevin, Speer, Christopher, Ruggiero, Jonathan David
• Granted Patent: US 9,495,200 B2
• Time in Patent Office: Days
• US Class Current: 718/101
• CPC Class Codes: G06F 9/46 Multiprogramming arrangements
• CPC Class Codes: G06F 9/5027 the resource being a machine, e.g. CPUs, Servers, Terminals

Abstract

A system for processing a batch job comprises a processor and a memory. The processor is configured to receive a job name for a job submitted to execute, to receive one or more job parameters, and to determine one or more nodes to run the job. The processor is configured to determine one or steps, where for each step: a step is executed on a node using a state of data associated with a start state of the step; and upon completion of executing the step, a result is stored to a durable storage. The durable storage stores the state of data associated with the start state of the step and the completion state of the step and are accessible by other execution processes as associated with either the start state of the step or the completion state of the step. The memory of the system is coupled to the processor and configured to provide processor with instructions.

Claims

1. A system for processing a batch job, comprising:

a processor configured to;

receive a job name for a job submitted to execute;

receive one or more job parameters;

determine one or more nodes to run the job;

determine one or more steps;

for each step;

execute a step of the one or more steps on a node of the one or more nodes using a state of data associated with a start state of the step;

upon completion of executing the step, store a result to a durable storage, wherein the durable storage stores a state of data associated with a completion state of the step, and wherein the state of data associated with the start state of the step and the completion state of the step are accessible by other execution is processes as associated with either the start state of the step or the completion state of the step;

a memory coupled with the processor, wherein the memory is configured to provide the processor with instructions.

2. A system as in claim 1, wherein determining the one or more nodes is based on one or more of the following:

a user requesting the job, a customer size, a customer service level, a priority of the job, a size of the job, a current resource availability, a current number or type of executing jobs, or an overall load on the system.

3. A system as in claim 1, wherein executing the step comprises splitting the execution into one or more partitions.

4. A system as in claim 3, wherein the one or more partitions have a maximum size.

5. A system as in claim 3, wherein the partition is executed in parallel within one node of the one or more nodes.

6. A system as in claim 5, wherein a number of parallel threads executing the partition is based at least in part on a number of physical processing cores associated with the one node of the one or more nodes.

7. A system as in claim 1, wherein the step comprises one of the following:

a calculation step, an update step, an aggregation step, a fork step, or a join step.

8. A method for executing a job, comprising:

receiving a job name for a job submitted to execute;

receiving one or more job parameters;

determining one or more nodes to run the job;

determining one or more steps;

for each step;

executing a step of the one or more steps on a node of the one or more nodes using a state of data associated with a start state of the step; and

upon completion of executing the step, storing a result to a durable storage, is wherein the durable storage stores a state of data associated with a completion state of the step, and wherein the state of data associated with the start state of the step and the completion state of the step are accessible by other execution processes as associated with either the start state of the step or the completion state of the step.

9. A method as in claim 8, wherein determining the one or more nodes is based on one or more of the following:

a user requesting the job, a customer size, a customer service level, a priority of the job, a size of the job, a current resource availability, a current number or type of executing jobs, or an overall load on the system.

10. A method as in claim 8, wherein executing the step comprises splitting the execution into one or more partitions.

11. A method as in claim 10, wherein the one or more partitions have a maximum size.

12. A method as in claim 10, wherein the partition is executed in parallel within one node of the one or more nodes.

13. A method as in claim 12, wherein a number of parallel threads executing the partition is based at least in part on a number of physical processing cores associated with the one node of the one or more nodes.

14. A method as in claim 8, wherein the step comprises one of the following:

a calculation step, an update step, an aggregation step, a fork step, or a join step.

15. A computer program product for executing a job, the computer program product being embodied in a computer readable storage medium and comprising computer instructions for:

receiving a job name for a job submitted to execute;

receiving one or more job parameters;

determining one or more nodes to run the job;

determining one or more steps;

for each step;

executing a step of the one or more steps on a node of the one or more nodes using a state of data associated with a start state of the step; and

is upon completion of executing the step, storing a result to a durable storage, wherein the durable storage stores a state of data associated with a completion state of the step, and wherein the state of data associated with the start state of the step and the completion state of the step are accessible by other execution processes as associated with either the start state of the step or the completion state of the step.

16. The computer program product recited in claim 15, wherein determining the one or more nodes is based on one or more of the following:

a user requesting the job, a customer size, a customer service level, a priority of the job, a size of the job, a current resource availability, a current number or type of executing jobs, or an overall load on the system.

17. The computer program product recited in claim 15, wherein executing the step comprises splitting the execution into one or more partitions.

18. The computer program product recited in claim 17, wherein the one or more partitions have a maximum size.

19. The computer program product recited in claim 17, wherein the partition is executed in parallel within one node of the one or more nodes.

20. The computer program product recited in claim 19, wherein a number of parallel threads executing the partition is based at least in part on a number of physical processing cores associated with the one node of the one or more nodes.

21. The computer program product recited in claim 15, wherein the step comprises one of the following:

a calculation step, an update step, an aggregation step, a fork step, or a join step.