Methods and systems for dynamic parallel processing

• US 8,108,521 B2
• Filed: 02/03/2006
• Issued: 01/31/2012
• Est. Priority Date: 02/04/2005
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: SAP AG (SAP SE)
• Inventors: Greven, Boris, Westendorf, Frank
• Primary Examiner(s): Moise, Emmanuel L
• Assistant Examiner(s): HUQ, FARZANA B
• Application Number:
  US11/346,306
• Publication Number:
  US 20060195336A1
• Time in Patent Office: 2,188 Days
• Field of Search: 709/225, 709/229, 709/246
• US Class Current: 709/226
• CPC Class Codes: G06F 9/5038 considering the execution order of a plurality of tasks, e.g. taking priority or time dependency constraints into consideration scheduling strategies G06F9/4881 and subgroups
• CPC Class Codes: G06Q 10/06316 Sequencing of tasks or work

Abstract

Systems and methods are disclosed for dynamic parallel processing of a plurality of subprocesses of a process initiated by a parallel processing framework. The plurality of subprocesses may include processing of one or more data objects in a computer system including a plurality of servers for processing the plurality of data objects by the plurality of subprocesses. The method may include building work packages from the plurality of data objects according to a process-specific scheme, and dynamically distributing the work packages of the plurality of data objects among the plurality of servers according to a process-specific execution order.

Claims

1. A computer-implemented method comprising:

performing dynamic parallel processing of a plurality of subprocesses of a process initiated by a parallel processing framework, each of the plurality of subprocesses including processing of one or more data objects in a computer system comprising a plurality of servers for processing the plurality of data objects by the plurality of subprocesses;

building, by using a processor of the computer system, work packages from the plurality of data objects according to a process-specific scheme and by using a wildcard to associate a range of the data objects to one of the work packages, wherein a size of one of the work packages is chosen based on a predefined granularity, the predefined granularity causing all available servers from the plurality of servers to execute at least one work package during runtime when the work packages are procurable;

setting a process specific execution order of the work packages, the work packages including a first work package and a second work package;

dynamically distributing the work packages among the plurality of servers according to the process-specific execution order, the plurality of servers including a first server and a second server, wherein processing of the work packages is performed by the plurality of servers in a semantically arbitrary order based on predefined process-specific parameters; and

generating a graphical display indicating that the first work package is assigned to the first server and that the second work package is assigned to the second server, wherein the graphical display includes a name of the first work package, a name of the second package, a name of the first server, a name of the second server, and a status indicating that the first work package is currently being processed by the first server.

2. The method according to claim 1, wherein the scheme is defined by the framework, such that the work packages can be executed in the semantically arbitrary order.

3. The method according to claim 1, wherein the execution order is defined by the framework.

4. The method according to claim 1, wherein the process-specific scheme accounts for blocking rules relating to the plurality of data objects, the blockings being activated at a beginning of the corresponding process.

5. The method according to claim 1, further comprising:

composing rules for building and rules for distributing the work packages; and

defining the process-specific scheme and the process-specific execution order, wherein the parallel processing framework provides a display for dynamically showing a status of the process during processing.

6. The method according to claim 1, wherein each work package is built up of one or more blocks, each block having a specific size.

7. The method according to claim 1, wherein the work packages are distributed among the plurality of servers.

8. The method according to claim 1, wherein the work packages are commit controlled.

9. The method according to claim 1, wherein each work package is processed on a server via an asynchronous remote function call.

10. The method according to claim 1, wherein each work package is processed according to the predefined granularity.

11. The method according to claim 1, wherein the process-specific scheme to build the work packages and the process-specific execution order of the work packages are substituted by functional modules to be executed.

12. A computer system for dynamic parallel processing of a plurality of subprocesses of a process, the plurality of subprocesses including processing of one or more data objects, comprising:

a computing unit providing a parallel processing framework, each of the plurality of subprocesses including processing of one or more data objects;

a plurality of servers for processing the plurality of data objects by the plurality of subprocesses; and

an adapter unit interconnecting the computing unit and the plurality of servers, the adapter unit being arranged and constructed to;

automatically build work packages of the plurality of data objects based on predefinable rules and by using a wildcard to associate a range of the data objects to one of the work packages, wherein a size of one of the work packages is chosen based on a predefined granularity, the predefined granularity causing all available servers from the plurality of servers to execute at least one work package during runtime when the work packages are procurable;

set a process specific execution order of the work packages, the work packages including a first work package and a second work package;

dynamically distribute the work packages of the plurality of data objects among the plurality of servers according to the process-specific execution order, the plurality of servers including a first server and a second server, wherein processing of the work packages is performed by the plurality of servers in a semantically arbitrary order based on predefined process-specific parameters; and

generate a graphical display indicating that the first work package is assigned to the first server and that the second work package is assigned to the second server, wherein the graphical display includes a name of the first work package, a name of the second package, a name of the first server, a name of the second server, and a status indicating that the first work package is currently being processed by the first server.

13. The computer system according to claim 12, wherein the rules are predefinable by the parallel processing framework, such that the work packages can be executed in the semantically arbitrary order.

14. The computer system according to claim 12, wherein the process-specific execution order can be defined by the parallel processing framework.

15. The computer system according to claim 12, wherein the adapter unit is chosen as a process-specific adapter unit.

16. The computer system according to claim 12, wherein the work packages can be distributed automatically according to a process-specific execution order predefined by the parallel processing framework.

17. The computer system according to claim 12, wherein each work package comprises one or more blocks, each block having a specific size.

18. The computer system according to claim 12, wherein the work packages can be commit controlled.

19. The computer system according to claim 18, wherein one or more of the work packages are divided into one or more commit blocks.

20. The computer system according to claim 12, wherein each work package can be processed on a server via an asynchronous remote function call.

21. A non-transitory computer readable storage medium comprising program code instructions which, when executed on a processor, causes the processor to perform a method, the method being executed by the processor comprising:

performing dynamic parallel processing of a plurality of subprocesses of a process initiated by a parallel processing framework, the plurality of subprocesses including processing of one or more data objects in a computer system comprising a plurality of servers for processing the plurality of data objects by the plurality of subprocesses;

building work packages from the plurality of data objects according to a process-specific scheme and by using a wildcard to associate a range of the data objects to one of the work packages, wherein a size of one of the work packages is chosen based on a predefined granularity, the predefined granularity causing all available servers from the plurality of servers to execute at least one work package during runtime when the work packages are procurable;

setting a process specific execution order of the work packages, the work packages including a first work package and a second work package;

dynamically distributing the work packages among the plurality of servers according to the process-specific execution order, the plurality of servers including a first server and a second server, wherein processing of the work packages is performed by the plurality of servers in a semantically arbitrary order based on predefined process-specific parameters; and

generating a graphical display indicating that the first work package is assigned to the first server and that the second work package is assigned to the second server, wherein the graphical display includes a name of the first work package, a name of the second package, a name of the first server, a name of the second server, and a status indicating that the first work package is currently being processed by the first server.

22. The non-transitory computer readable storage medium according to claim 21, wherein the program code comprises a computer implemented framework.

23. The non-transitory computer readable storage medium according to claim 22, further comprising a display for dynamically showing a status of the process during processing.

24. The non-transitory computer readable storage medium according to claim 22, further comprising means for defining functional modules substituting a process-specific scheme and a process-specific execution order when being executed.