Increased performance of grid applications

• US 20060130066A1
• Filed: 12/13/2004
• Published: 06/15/2006
• Est. Priority Date: 12/13/2004
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: SAP AG (SAP SE)
• Inventors: Bozak, Erol, Gebhart, Alexander
• Granted Patent: US 7,627,655 B2
• Time in Patent Office: Days
• US Class Current: 718/104
• CPC Class Codes: G06F 8/456 Parallelism detection

Abstract

A method includes identifying a component within a non-distributed grid-enabled software application that can be isolated and independently executed during an execution of the software application on a first compute resource within a grid network, replacing the isolatable component in the software application with a proxy, and executing the software application on the first compute resource with the isolated component executing on a second compute resource.

Claims

1. A method comprising:

identifying a component within a non-distributed grid-enabled software application that can be isolated and independently executed during an execution of the software application on a first compute resource within a grid network;

replacing the isolatable component in the software application with a proxy; and

executing the software application on the first compute resource with the isolated component executing on a second compute resource in the grid network.

2. The method of claim 1 wherein the software application comprises a Java application having a plurality of classes, the isolated component representing a first class within the Java application.

3. The method of claim 2 wherein the proxy communicates between the Java application on the first compute resource and a Java Virtual Machine (JVM) on the second compute resource.

4. The method of claim 2 wherein identifying comprises:

receiving a list of components;

executing a byte code analyzer to inspect byte code representing components in the list; and

determining which of the components contained in the list can be separated from the software application.

5. The method of claim 4 wherein determining comprises selecting a critical component.

6. The method of claim 1 wherein replacing comprises distributing the isolated component to the second compute resource prior to the execution of the application.

7. The method of claim 1 wherein executing comprises invoking the isolatable component on the second compute resource.

8. The method of claim 1 wherein executing further comprises executing the isolatable component on the first compute resource if a network connection between the first compute resource and the second compute resource is lost.

9. The method of claim 1 wherein executing comprises executing the isolatable component on the first compute resource if an error occurs during execution of the isolatable component on the second compute resource.

10. A method comprising:

in a grid network, distributing an isolatable component of a non-distributed grid-enabled software application residing on a first compute resource to a second compute resource; and

replacing the isolatable component in the non-distributed software application with a proxy that communicates with the isolatable component on the second compute resource.

11. The method of claim 10 further comprising executing the non-distributed software application on the first compute resource.

12. The method of claim 11 wherein executing comprises executing the isolatable component on the second compute resource as directed by the proxy.

13. The method of claim 10 wherein distributing comprises sending the isolatable component to the second compute resource prior to an execution of the non-distributed software application on the first compute resource.

14. The method of claim 10 wherein distributing comprises sending the isolatable component to the second compute resource during an execution of the non-distributed software application on the first compute resource.

15. The method of claim 10 wherein the non-distributed software application is a Java application and the isolatable component is a Java class.

16. A computer program product, tangibly embodied in an information carrier, for increased performance of a non-distributed grid-enabled software application, the computer program product being operable to cause data processing apparatus to:

identify a component within the non-distributed software application that can be isolated and independently executed during an execution of the software application on a first compute resource within a grid network;

replace the isolatable component in the software application with a proxy; and

execute the software application on the first compute resource with the isolated component executing on a second compute resource in the grid network.

17. The product of claim 16 wherein the software application comprises a Java application having a plurality of classes, the isolated component representing a first class with the Java application.

18. The product of claim 17 wherein the proxy communicates between the Java application on the first compute resource and a Java Virtual Machine (JVM) on the second compute resource.

19. The product of claim 17 wherein identifying comprises:

receiving a list of components;

executing a byte code analyzer to inspect byte code representing components in the list; and

determining which of the components contained in the list can be separated from the software application.

20. The product of claim 19 wherein determining comprises selecting a critical component.

21. The product of claim 16 wherein replacing comprises distributing the isolated component to the second compute resource prior to the execution of the application.

22. The product of claim 16 wherein executing comprises invoking the isolatable component on the second compute resource.

23. The product of claim 16 wherein executing further comprises executing the isolatable component on the first compute resource if a network connection between the first compute resource and the second compute resource is lost.

24. The product of claim 16 wherein executing comprises executing the isolatable component on the first compute resource if an error occurs during execution of the isolatable component on the second compute resource.

25. An apparatus comprising:

means for identifying a component within a non-distributed grid-enabled software application that can be isolated and independently executed during an execution of the software application on a first compute resource within a grid network;

means for replacing the isolatable component in the software application with a proxy in the grid network; and

means for executing the software application on the first compute resource with the isolated component executing on a second compute resource in the grid network.

26. The system of claim 25 wherein the software application comprises a Java application having a plurality of classes, the isolated component representing a first class with the Java application.

27. The system of claim 26 wherein the proxy communicates between the Java application on the first compute resource and a Java Virtual Machine (JVM) on the second compute resource.

28. The system of claim 26 wherein the means for identifying comprises:

receiving a list of components;

executing a byte code analyzer to inspect byte code representing components in the list; and

determining which of the components contained in the list can be separated from the software application.

29. The system of claim 28 wherein determining comprises selecting a critical component.

30. The system of claim 25 wherein means for replacing comprises distributing the isolated component to the second compute resource prior to the execution of the application.

31. The system of claim 25 wherein means for executing comprises invoking the isolatable component on the second compute resource.

32. The system of claim 25 wherein means for executing further comprises executing the isolatable component on the first compute resource if a network connection between the first compute resource and the second compute resource is lost.

33. The system of claim 25 wherein means for executing comprises executing the isolatable component on the first compute resource if an error occurs during execution of the isolatable component on the second compute resource.