Grid services framework

• US 20050021594A1
• Filed: 12/05/2003
• Published: 01/27/2005
• Est. Priority Date: 02/04/2002
• Status: Active Grant

• Current Assignee: Cloud Software Group
• Original Assignee: DataSynapse, Inc. (Cloud Software Group)
• Inventors: Lee, Peter, Bernardin, James
• Granted Patent: US 7,130,891 B2
• Time in Patent Office: Days
• US Class Current: 709/200
• CPC Class Codes: G06F 2209/503 Resource availability
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: Y10S 707/99932 Access augmentation or optimizing

Abstract

A service-oriented framework allows client applications to access computational services hosted on a distributed computing grid. Services facilitate remote, parallel execution of code in a way that is scalable, fault-tolerant, dynamic and language-independent. Services can be written in a variety of languages, and do not need to be compiled or linked with vendor-supplied code. A client written in one language can invoke a Service written in another. A benefit of the invention over traditional approaches is that it virtualizes the Service. Rather than send a request directly to the remote machine hosting the Service, a client request is sent to a manager, which enqueues it until an Engine is available. The first Engine to dequeue the request hosts the Service. This mechanism, in which a single virtual Service instance (the client-side object) is implemented by one or more physical instances (Engine processes), provides for fault tolerance and essentially unlimited scalability.

Claims

1. A method for providing computational services to a client using a grid-based distributed computing system, the system including a plurality of engines and at least one grid manager, the method comprising:

deploying executable code corresponding one or more service(s) such that the engines can access the executable code;

registering the service(s) with the manager;

creating instance(s) of the service(s) that may be invoked by the client; and

,

using one or more of the instance(s) to invoke one or more of the registered service(s), wherein invoking a service comprises;

communicating a service request to the manager;

using the manager to assign the service request to an available engine; and

,

executing code corresponding to the requested service on the assigned engine.

2. A method, as defined in claim 1, wherein deploying executable code comprises storing the executable code on a shared file system accessible to the engines.

3. A method, as defined in claim 1, wherein deploying executable code comprises using a file update mechanism provided by the manager to distribute the executable code to the engines.

4. A method, as defined in claim 3, wherein using a file update mechanism to distribute the executable code to the engines comprises distributing the executable code to the engines when the manager is idle.

5. A method, as defined in claim 1, wherein creating an instance of a service comprises invoking a service creation method.

6. A method, as defined in claim 1, wherein communicating a service request to the manager comprises invoking a synchronous invocation method.

7. A method, as defined in claim 1, wherein communicating a service request to the manager comprises invoking an asynchronous invocation method.

8. A method, as defined in claim 1, wherein communicating a service request to the manager comprises synchronously invoking a client-side proxy that corresponds to the service request.

9. A method, as defined in claim 1, wherein communicating a service request to the manager comprises asynchronously invoking a client-side proxy that corresponds to the service request.

10. A method, as defined in claim 1, wherein communicating a service request to the manager comprises communicating a synchronous service request via a client-side wsdl proxy.

11. A method, as defined in claim 1, wherein communicating a service request to the manager comprises communicating an asynchronous service request via a client-side wsdl proxy.

12. A method, as defined in claim 1, wherein using the manager to assign the service request to an available engine comprises applying an adaptive scheduling algorithm to select an engine.

13. A method, as defined in claim 12, wherein applying an adaptive scheduling algorithm to select an engine comprises selecting among competing engines requesting work from the manager.

14. A method, as defined in claim 12, wherein applying an adaptive scheduling algorithm to select an engine comprises computing a neediness score for service(s) with pending request(s).

15. A method, as defined in claim 14, wherein each neediness score is computed, at least in part, from two or more of:

(i) a priority weight for the service, (ii) an amount of time that engines have spent running task(s) associated with the service, and (iii) an amount of time that the request for the service has spent waiting to be assigned.

16. A method, as defined in claim 14, wherein applying an adaptive scheduling algorithm to select an engine further comprises computing affinity scores reflecting the affinities of available engines to requested service(s).

17. A method, as defined in claim 16, wherein each affinity score is computed, at least in part, from information representing the amount of the requested service'"'"'s state that is already present on the engine-in-question.

18. A method, as defined in claim 14, wherein requested service(s) are first ranked according to neediness score(s), then assigned to engine(s) according to affinity score(s).

19. A method for providing a client with a flexible, service-oriented computing environment using a computing grid, the method comprising:

invoking a grid manager to create client-side instances of stateless services;

invoking the grid manager to create client-side instances of stateful services;

using one or more of the client-side instances to invoke stateless services on one or more manager-assigned engines; and

,

using one or more of the client-side instances to invoke stateful services on one or more manager-assigned engines.

20. A method, as defined in claim 19, wherein invoking a grid manager to create client-side instances of stateless services comprises invoking a service creation method.

21. A method, as defined in claim 19, wherein invoking a grid manager to create client-side instances of stateless services comprises creating proxy objects associated with the stateless services.

22. A method, as defined in claim 19, wherein invoking a grid manager to create client-side instances of stateful services comprises invoking a service creation method.

23. A method, as defined in claim 19, wherein using one or more of the client-side instances to invoke a stateful service on one or more manager-assigned engines further comprises associating a state update method with the invoked service instance.

24. A method, as defined in claim 23, wherein associating a state update method comprises associating an append state method.

25. A method, as defined in claim 23, wherein associating a state update method comprises associating a set state method.

26. A method, as defined in claim 19, wherein invoking a grid manager to create client-side instances of stateful services comprises creating proxy objects associated with the stateful services.

27. A method, as defined in claim 26, further comprising implementing one or more asynchronous callback interface(s) associated with the proxy objects.

28. A method, as defined in claim 27, wherein implementing an asynchronous callback interface comprises associating a response handling method with the proxy-in-question.

29. A method, as defined in claim 28, wherein implementing an asynchronous callback interface further comprises associating an error handling method with the proxy-in-question.

30. A method, as defined in claim 26, wherein creating proxy objects associated with the stateful services further comprises designating stateful methods.

31. A method, as defined in claim 30, wherein designating a stateful method comprises designating the method as an append state method.

32. A method, as defined in claim 30, wherein designating a stateful method comprises designating the method as a set state method.

33. A method of deploying and invoking a stateful service on a grid computing platform, the method comprising:

deploying service code so as to be accessible to engines in the grid computing platform;

generating a service proxy corresponding to the deployed service code;

configuring the proxy to include state update information for each stateful method associated with the service code; and

,

using the proxy to synchronously and/or asynchronously invoke stateful methods associated with the service code, whereby the proxy'"'"'s state update information is used, as needed, to maintain coherent state information between nodes of the grid computing platform.

34. A method, as defined in claim 33, wherein configuring the proxy to include state update information for each stateful method associated with the service code comprises designating each stateful method as an (i) append state method or a (ii) set state method.