Software Platform and System for Grid Computing

• US 20100281166A1
• Filed: 11/07/2008
• Published: 11/04/2010
• Est. Priority Date: 11/09/2007
• Status: Active Grant

• Current Assignee: IP3 2017, Series 200 of Allied Security Trust I (Allied Security Trust)
• Original Assignee: Manjrasoft Pty. Ltd.
• Inventors: Nadiminti, Krishna, Venugopal, Srikumar, Buyya, Rajkumar, Chu, Xingchen
• Granted Patent: US 8,230,070 B2
• Time in Patent Office: Days
• US Class Current: 709/226
• CPC Class Codes: G06F 9/5072 Grid computing

Abstract

A software platform for providing grid computing on a network of computing nodes, comprising a configurable service container executable at the nodes, including message dispatching, communication, network membership and persistence modules, and adapted to host pluggable service modules. When executed at the nodes at least one instance of the container includes a membership service module for maintaining network connectivity between the nodes, at least one instance of the container includes a scheduler service module configured to receive one or more tasks from a client and schedule the tasks on at least one of the nodes, and at least one instance of the container includes an executor service module for receiving one or more tasks from the scheduler service module, executing the tasks so received and returning at least one result to the scheduler service module.

Claims

1. A software platform for providing grid computing on a network of computing nodes in mutual data communication, comprising:

a configurable service container executable at said nodes, said container comprising message dispatching, communication, network membership and persistence modules, and host pluggable service modules;

wherein when executed at said nodes at least one instance of said container includes a membership service module for maintaining network connectivity between said nodes, at least one instance of said container includes a scheduler service module configured to receive one or more tasks from a client and schedule said tasks on at least one of said nodes, and at least one instance of said container includes an executor service module for receiving one or more tasks from said scheduler service module, executing said tasks so received and returning at least one result to said scheduler service module.

2. The platform as claimed in claim 1, wherein said service modules support a selected programming model.

3. The platform as claimed in claim 1, wherein said service modules support a plurality of programming models.

4. The platform as claimed in claim 1, wherein a plurality of said computing nodes are executed on respective processor cores of a single processor.

5. The platform as claimed in claim 1, wherein said container includes security and logging modules.

6. The platform as claimed in claim 1, wherein at least one instance of said container includes more than one of said membership service module, said scheduler module and said executor module.

7. The platform as claimed in claim 1, wherein when executed at said nodes a plurality of instances of said container includes an executor module for executing tasks.

8. The platform as claimed in claim 1, wherein each node comprises a computing device, and wherein a single computing device comprises multiple nodes when the computing device has multiple processors or multiple processor cores.

9. The platform as claimed in claim 1, wherein services provided by said modules and said container are mutually independent.

10. The platform as claimed in claim 1, further comprising an allocation manager service that checks an availability of a computation resource on said nodes in response to a negotiation for said computation resource and reserves said computation resource when said negotiation succeeds.

11. The platform as claimed in claim 10, wherein said negotiation is conducted via a negotiation web service.

12. The platform as claimed in claim 1, further comprising a MapReduce programming model.

13. The platform as claimed in claim 12, wherein said MapReduce programming model is adapted for a .NET platform.

14. A grid of computing nodes in mutual data communication, each of said nodes comprising:

a configurable service container executed at said respective node, said container comprising message dispatching, communication, network membership and persistence modules, and host pluggable service modules;

wherein at least one of said containers includes a membership service module for maintaining network connectivity between said nodes, at least one of said containers includes a scheduler service module configured to receive one or more tasks from a client and schedule said tasks on at least one of said nodes, and at least one of said containers includes an executor service module for receiving one or more tasks from said scheduler service module, executing said tasks so received and returning at least one result to said scheduler service module.

15. The grid as claimed in claim 14, wherein said service modules support a selected programming model.

16. The grid as claimed in claim 14, wherein said service modules support a plurality of programming models.

17. The grid as claimed in claim 14, wherein a plurality of said computing nodes are executed on respective processor cores of a single processor.

18. The grid as claimed in claim 14, wherein said container includes security and logging modules.

19. The grid as claimed in claim 14, wherein at least one instance of said container includes more than one of said membership service module, said scheduler module and said executor module.

20. The grid as claimed in claim 14, wherein when executed at said nodes a plurality of instances of said container includes an executor module for executing tasks.

21. The grid as claimed in claim 14, wherein each node comprises a computing device, and wherein a single computing device comprises multiple nodes when the computing device has multiple processors or multiple processor cores.

22. The grid as claimed in claim 14, wherein services provided by said modules and said container are mutually independent.

23. A grid computing method for providing grid computing on a network of computing nodes in mutual data communication, said method comprising:

executing a configurable service container at said nodes, said container comprising message dispatching, communication, network membership and persistence modules, and being adapted to host pluggable service modules;

maintaining network connectivity between said nodes with a membership service module of at least one instance of said container;

receiving one or more tasks from a client and scheduling said tasks on at least one of said nodes with a scheduler service module of at least one instance of said container; and

receiving one or more tasks from said scheduler service module, executing said tasks so received and returning at least one result to said scheduler service module with an executor service module of at least one instance of said container.

24. The method as claimed in claim 23, further comprising adapting said service modules to support a selected programming model, and executing said selected programming model.

25. The method as claimed in claim 23, further comprising adapting said service modules to support a plurality of programming models and executing said programming models.

26. The method as claimed in claim 23, further comprising adapting said service modules to support at least one parallel programming model and at least one distributed programming model.

27. The method as claimed in claim 23, wherein a plurality of said computing nodes comprise respective processor cores of a single processor.

28. The method as claimed in claim 23, further comprising checking availability of a computation resource on said nodes with an allocation manager service in response to a negotiation for said computation resource and reserving said computation resource with said allocation manager service when said negotiation succeeds.

29. The method as claimed in claim 28, further comprising conducting said negotiation via a negotiation web service.

30. The method as claimed in claim 23, further comprising providing a MapReduce programming model.

31. The method as claimed in claim 30, wherein said MapReduce programming model is adapted for a .NET platform.

32. A grid computing method for performing grid computing on a network of computing nodes in mutual data communication, said method comprising:

executing on each of said nodes a configurable service container executed at said respective node, said container comprising message dispatching, communication, network membership and persistence modules, and being adapted to host pluggable service modules;

wherein at least one of said containers includes a membership service module for maintaining network connectivity between said nodes, at least one of said containers includes a scheduler service module configured to receive one or more tasks from a client and schedule said tasks on at least one of said nodes, and at least one of said containers includes an executor service module for receiving one or more tasks from said scheduler service module, executing said tasks so received and returning at least one result to said scheduler service module.

33. (canceled)

34. (canceled)

35. (canceled)

36. (canceled)