Mechanism for bridging a thread-oriented computing paradigm and a job-oriented computing paradigm
A mechanism is provided for fostering the adoption of grid computing by bridging the thread-oriented paradigm (used in single and multi-processor computers) and the job-oriented paradigm (used in grid computing). This mechanism allows an application developer to write an application in the same manner as if he were writing a thread-oriented program (in terms of creating and joining threads). However, when the application is executed on a client computer and the mechanism is invoked, the mechanism interacts with a resource management interface to cause a distributed resource manager to create, distribute, and manage the execution of jobs on the resource computers of a grid. Thus, rather than actually spawning threads on the client computer, the mechanism causes jobs to be spawned and executed on the grid. By doing this, the mechanism allows the application to take advantage of grid computing without requiring the application to be grid-aware.
1. A method implemented within a client computer to enable the client computer to interact with a computing grid, wherein the computing grid comprises a plurality of resource computers and a distributed resource manager (DRM), wherein the DRM coordinates assignment of jobs to the plurality of resource computers, and wherein the method, comprises:
receiving, from a grid-unaware application executing within the client computer, a request to create a new thread within the client computer to execute a set of code concurrently with the grid-unaware application;
in response to the request to create the new thread, invoking a resource management interface to cause the resource management interface to interact with the DRM to cause the DRM to:
create a new job which entails execution of the set of code in place of the new thread,
assign the new job to a particular resource computer of the plurality of resource computers to be executed thereon, such that the set of code is executed on the particular resource computer instead of on the client computer, and
receive a set of execution results which were generated by the particular resource computer as a result of executing the set of code, when a request from the grid-unaware application to join the new thread is received,
wherein the new thread is not created within the client computer to execute the set of code.
2. The method of claim 1 , wherein the grid-unaware application is unaware that the set of code is not being executed on the client computer.
3. The method of claim 1 , wherein the grid-unaware application is under an impression that a new thread was created within the client computer to execute the set of code, and wherein the method further comprises:
obtaining the set of execution results from the resource management interface; and
providing the set of execution results to the grid-unaware application.
4. The method of claim 1 , wherein the resource management interface resides on the client computer.
5. The method of claim 1 , wherein the resource management interface is an application programming interface (API) configured under a job-oriented paradigm.
6. A method implemented within a client computer to enable the client computer to interact with a computing grid, wherein the computing grid comprises a plurality of resource computers and a distributed resource manager (DRM), wherein the DRM coordinates assignment of jobs to the plurality of resource computers, and wherein the method comprises:
receiving, from a grid-unaware application executing within the client computer, a request to create a new thread within the client computer to execute a set of code concurrently with the grid-unaware application;
in response to the request to create the new thread, interacting with the DRM to cause the DRM to:
create a new job which entails execution of the set of code in place of the new thread, and
assign the new job to a particular resource computer of the plurality of resource computers to be executed thereon, such that the set of code is executed on the particular resource computer instead of on the client computer; and
receive a set of execution results which were generated by the particular resource computer as a result of executing the set of code, when a request from the grid-unaware application to join the new thread is received,
wherein the new thread is not created within the client computer to execute the set of code.
7. The method of claim 6 , wherein the grid-unaware application is unaware that the set of code is not being executed on the client computer.
8. The method of claim 6 , wherein the grid-unaware application is under an impression that a new thread was created within the client computer to execute the set of code, and wherein the method further comprises:
providing the set of execution results to the application.
9. A machine readable medium comprising instructions for causing one or more processors of a client computer to interact with a computer grid, wherein the computing grid comprises a plurality of resource computers and a distributed resource manager (DRM), wherein the DRM coordinates assignment of jobs to the plurality of resource computers, and wherein the machine readable medium comprises:
instructions for causing one or more processors to receive, from a grid-unaware application executing within the client computer, a request to create a new thread within the client computer to execute a set of code concurrently with the grid-unaware application;
instructions for causing one or more processors to invoke, in response to the request to create the new thread, a resource management interface to cause the resource management interface to interact with the DRM to cause the DRM to:
create a new job which entails execution of the set of code in place of the new thread, and
assign the new job to a particular resource computer of the plurality of resource computers to be executed thereon, such that the set of code is executed on the particular resource computer instead of on the client computer; and
receive a set of execution results which were generated by the particular resource computer as a result of executing the set of code, when a request from the grid-unaware application to join the new thread is received,
wherein the new thread is not created within the client computer to execute the set of code.
10. The machine readable medium of claim 9 , wherein the grid-unaware application is unaware that the set of code is not being executed on the client computer.
11. The machine readable medium of claim 9 , wherein the grid-unaware application is under an impression that a new thread was created within the client computer to execute the set of code, and wherein the machine readable medium further comprises:
instructions for causing one or more processors to obtain the set of execution results from the resource management interface; and
instructions for causing one or more processors to provide the set of execution results to the grid-unaware application.
12. The machine readable medium of claim 9 , wherein the resource management interface resides on the client computer.
13. The machine readable medium of claim 9 , wherein the resource management interface is an application programming interface (API) configured under a job-oriented paradigm.
14. A machine readable medium comprising instructions for causing one or more processors of a client computer to interact with a computing grid, wherein the computing grid comprises a plurality of resource computers and a distributed resource manager (DRM), wherein the DRM coordinates assignment of jobs to the plurality of resource computers, and wherein the machine readable medium comprises:
instructions for causing one or more processors to receive, from a grid-unaware application executing within the client computer, a request to create a new thread within the client computer to execute a set of code concurrently with the grid-unaware application;
instructions for causing one or more processors to interact, in response to the request to create the new thread, with the DRM to cause the DRM to:
create a new job which entails execution of the set of code in place of the new thread, and
assign the new job to a particular resource computer of the plurality of resource computers to be executed thereon, such that the set of code is executed on the particular resource computer instead of on the client computer; and
receive a set of execution results which were generated by the particular resource computer as a result of executing the set of code, when a request from the grid-unaware application to join the new thread is received,
wherein the new thread is not created within the client computer to execute the set of code.
15. The machine readable medium of claim 14 , wherein the grid-unaware application is unaware that the set of code is not being executed on the client computer.
16. The machine readable medium of claim 14 , wherein the grid-unaware application is under an impression that a new thread was created within the client computer to execute the set of code, and wherein the machine readable medium further comprises:
instructions for causing one or more processors to provide the set of execution results to the grid-unaware application.