CONFIGURABLE MASS DATA PORTIONING FOR PARALLEL PROCESSING

• US 20130061026A1
• Filed: 09/05/2011
• Published: 03/07/2013
• Est. Priority Date: 09/05/2011
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: Artur Kaufmann, Georg Lang
• Inventors: KAUFMANN, ARTUR, LANG, GEORG
• Granted Patent: US 8,875,137 B2
• Time in Patent Office: Days
• US Class Current: 712/220
• CPC Class Codes: G06F 9/4843 by program, e.g. task dispatcher, supervisor, operating system

Abstract

A configurable mass data portioning for parallel processing is described herein. One or more operation attributes are selected to participate in parallelization criteria. The values of the selected operation attributes for a number of operations are submitted to a specified algorithm using to provide parallelization values corresponding to the operations. The parallelization values are applied to group the operations in comparable portions for parallel execution without conflicts.

Claims

1. A computer system processing business transactions in parallel, the system comprising:

a memory to store computer executable instructions; and

a processor coupled to said memory and operable to execute said instructions to generate

a data structure to receive and store at least one operation attribute selected for parallelization criteria,

an operation manager to receive a plurality of operations and to compute a plurality of parallelization values based on said at least one operation attribute, wherein at least one parallelization value of said plurality of parallelization values corresponds to at least one operation of said plurality of operations,

an execution process comprising a plurality of process threads, wherein the execution process executes in parallel a plurality of portions of said operations utilizing separate process threads of said plurality of process threads, and

at least one consolidating document aggregating results of said execution in parallel of said plurality of portions.

2. The system of claim 1, wherein said data structure comprises:

at least one field corresponding to at least one qualitative operation attribute to receive and store at least one value of said at least one qualitative operation attribute; and

at least one field corresponding to said parallelization criteria to receive and store said at least one operation attribute corresponding to said at least one value to said at least one qualitative operation attribute.

3. The system of claim 1, wherein said operation manager comprises:

an algorithm to derive at least one parallelization value of said plurality of parallelization values corresponding to an operation of said plurality of operations based on at least one value of said at least one operation attribute selected for parallelization criteria.

4. The system of claim 1, wherein said execution process comprises:

a locking mechanism to provide simultaneous access to a shared resource for said plurality of process threads based on said at least one operation attribute selected for parallelization criteria.

5. A non-transitory computer-readable medium storing instructions, which when executed cause a computer system to:

receive at least one operation attribute selected to participate in parallelization criteria;

for a plurality of operations, compute a plurality of parallelization values based on said at least one operation attribute, wherein at least one parallelization value of said plurality of parallelization values corresponds to at least one operation of said plurality of operations;

split said plurality of operations into a plurality of portions based on said plurality of parallelization values; and

execute in parallel a first portion of operations of said plurality of portions and a second portion of operations of said plurality of portions.

6. The computer-readable medium of claim 5, wherein to receive said at least one operation attribute comprises:

receiving at least one value of at least one qualitative operation attribute; and

storing correspondence between said at least one value of said at least one qualitative operation attribute and said at least one operation attribute selected to participate in said parallelization criteria.

7. The computer-readable medium of claim 5, wherein to compute said plurality of parallelization values comprises:

from an operation of said plurality of operations, extracting at least one value of said at least one operation attribute selected to participate in said parallelization criteria; and

submitting said at least one value of said at least one operation attribute to a specified algorithm to derive at least one corresponding parallelization value of said plurality of parallelization values.

8. The computer-readable medium of claim 5, wherein to compute said plurality of parallelization values comprises:

calculating a first result of a hash function with at least one input parameter set to at least one value of said at least one operation attribute for a first operation of said plurality of operations;

calculating a second result of said hash function with said at least one input parameter set to at least one value of said at least one operation attribute for a second operation of said plurality of operations, wherein said second result; and

order said first operation and said second operation based on said first result and said second result of said hash function.

9. The computer-readable medium of claim 5, wherein to split said plurality of operations comprises:

grouping said plurality of parallelization values in a plurality of intervals; and

portioning said plurality of operations by said plurality of value intervals.

10. The computer-readable medium of claim 5, wherein to execute in parallel said first portion of operations and said second portion of operations comprises:

accessing a shared resource by parallel process threads based on locks applied on said at least one operation attribute selected to participate in said parallelization criteria.

11. The computer-readable medium of claim 5 storing instructions, which when executed cause the computer system further to:

generate aggregated mass activity report of the simultaneous execution of said first portion of operations and said second portion of operations.

12. The computer-readable medium of claim 11, wherein generating said aggregated mass activity result comprises:

consolidating a first result of said simultaneous execution of said first portion of operations with a second result of said simultaneous execution of said second portion of operations.

13. A computer implemented method for processing business transactions in parallel, the method comprising:

receiving at least one operation attribute selected for parallelization criteria;

for a plurality of operations, computing a plurality of parallelization values based on said at least one operation attribute, wherein at least one parallelization value of said plurality of parallelization values corresponds to at least one operation of said plurality of operations;

breaking said plurality of parallelization values to a plurality of intervals;

splitting said plurality of operations into a plurality of portions by said plurality of value intervals; and

executing simultaneously a first portion of operations of said plurality of portions by a first parallel computer processes thread and a second portion of operations of said plurality of portions by a second parallel computer process thread.

14. The method of claim 13, wherein receiving said at least one operation attribute comprises:

receiving at least one value of at least one qualitative operation attribute; and

extracting said at least one operation attribute selected for said parallelization criteria from a data structure storing a correspondence between said at least one qualitative operation attribute and a plurality of operation attributes selected for said parallelization criteria.

15. The method of claim 13, wherein computing said plurality of parallelization values comprises:

from an operation of said plurality of operations, extracting at least one value of said at least one operation attribute selected for said parallelization criteria; and

submitting said at least one value of said at least one operation attribute to a specified algorithm to derive at least one corresponding parallelization value of said plurality of parallelization values.

16. The method of claim 13, wherein computing said plurality of parallelization values comprises:

calculating a first result of a hash function with at least one input parameter set to at least one value of said at least one operation attribute for a first operation of said plurality of operations; and

calculating a second result of said hash function with said at least one input parameter set to at least one value of said at least one operation attribute for a second operation of said plurality of operations, wherein said second result is similar to said first result when a simultaneous execution of said first operation and said second operation will cause conflicts.

17. The method of claim 13, wherein splitting said plurality of operations comprises:

indexing said plurality of operations according to said plurality of parallelization values.

18. The method of claim 13, wherein executing simultaneously said first portion of operations and said second portion of operations comprises:

locking a shared resource based on said at least one operation attribute selected for parallelization criteria.

19. The method of claim 13 further comprises:

generating aggregated mass activity result based on the simultaneous execution of said first portion of operations and said second portion of operations.

20. The method of claim 19, wherein generating said aggregated mass activity result comprises:

consolidating a first result of said simultaneous execution of said first portion of operations with a second result of said simultaneous execution of said second portion of operations.