Job-centric scheduling in a grid environment

• US 20050283782A1
• Filed: 06/17/2004
• Published: 12/22/2005
• Est. Priority Date: 06/17/2004
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: Platform Computing Corporation
• Inventors: Lam, Jason, Lu, Bingfeng, Zhong, Jinsong
• Granted Patent: US 7,861,246 B2
• Time in Patent Office: Days
• US Class Current: 718/100
• CPC Class Codes: G06F 9/5072 Grid computing

Abstract

A system for scheduling jobs in a networked computing grid having one or more job service components for creating one or more jobs having certain resource requirements is described. The system includes a plurality of resource manager components communicatively linked with the one or more job service components, each resource manager component being configured to manage a plurality of resources. The system further includes one or more information components communicatively linked with the one or more job service components and the plurality of resource manager components, the one or more information components being configured to aggregate resource data obtained from the resource manager components. The one or more jobs include a scheduling mechanism for querying the one or more information components for resource data, comparing the resource data against the resource requirements and independently scheduling the job to one or more of the plurality of resource manager components. A resource reservation is also provided for reserving resources for the jobs. Also described are related methods for performing job-centric scheduling.

Claims

1. A system for scheduling jobs in a networked computing grid having one or more job service components for creating one or more jobs having certain resource requirements, comprising:

a plurality of resource manager components communicatively linked with the one or more job service components, each resource manager component being configured to manage a plurality of resources; and

one or more information components communicatively linked with the one or more job service components and the plurality of resource manager components, the one or more information components being configured to aggregate resource data obtained from the plurality of resource manager components;

wherein the one or more jobs include a scheduling mechanism for querying the one or more information components for resource data, comparing the resource data against the resource requirements and independently scheduling the one or more jobs to one or more of the plurality of resource manager components.

2. The system of claim 1 further comprising a plurality of resource manager adapter components associated with the plurality of resource manager components, the resource manager adapter components being provided to enable communication between the one or more jobs and the plurality of resource manager components.

3. The system of claim 1 further comprising a metascheduler component communicatively linked with the one or more job service components and the one or more information components, the metascheduler component being configured to provide scheduling suggestions to the one or more jobs.

4. The system of claim 1 further comprising a reservation service component, the reservation service component being configured to receive a reservation request from the one or more jobs and reserve resources according to the reservation request.

5. The system of claim 1 further comprising a persistent service component and an associated database, the persistent service component being configured to download job data to the associated database and recover job data from the associated database.

6. A method for scheduling a job in a networked computing grid, the job having certain resource requirements, the method comprising:

providing a plurality of resources within the networked computing grid; and

enabling the job to schedule itself to one or more of the plurality of resources.

7. The method of claim 6 further comprising providing one or more resource manager components for managing one or more of the plurality of resources.

8. The method of claim 7 further comprising providing one or more information components in communication with the one or more resource manager components, the one or more information components being operable to aggregate resource data obtained from the one or more resource manager components.

9. The method of claim 8 wherein the resource data is obtained by configuring the one or more information components to extract resource data from the one or more resource manager components.

10. The method of claim 8 wherein the resource data is obtained by configuring the one or more resource manager components to submit resource data to the one or more information components.

11. The method of claim 8 wherein the resource data is obtained by configuring the one or more information components to extract resource data from the one or more resource manager components and configuring the one or more resource manager components to submit resource data to the one or more information components.

12. The method of claim 8 wherein enabling the job comprises enabling the job to query the one or more information components and evaluate whether one or more of the plurality of resources is compatible with the resource requirements of the job.

13. The method of claim 12 wherein the job evaluates whether one or more of the plurality of resources is compatible according to a scheduling algorithm.

14. The method of claim 8 further comprising providing a plurality of resource manager adapter components associated with the resource manager components, the resource manager adapter components enabling communication between the job and the plurality of resource manager components.

15. The method of claim 14 wherein one of the resource manager adapter components and one of the associated resource manager components are deployed into a first cluster, wherein the job originates from a second cluster different than the first cluster, and wherein enabling the job comprises enabling the job to communicate with the one of the resource manager components through the one of the resource manager adapter components.

16. The method of claim 14 further comprising providing a metascheduler component, wherein enabling the job comprises enabling the job to query the metascheduler component for scheduling suggestions.

17. The method of claim 14 further comprising providing a reservation service component, wherein enabling the job comprises enabling the job to query the reservation service component to reserve one or more of the plurality of resources.

18. A method for scheduling a job having certain resource requirements within a networked computing grid having a plurality of resources, the method comprising:

providing a plurality of resource manager components for managing the plurality of resources associated with the networked computing grid;

aggregating resource data obtained from the resource manager components;

enabling the job to query the resource data to evaluate whether one or more of the plurality of resources is compatible with the resource requirements of the job; and

scheduling the job to one or more of the plurality of resources determined to be compatible with the resource requirements of the job via communication between the job and one or more of the plurality of resource manager components associated with the one or more of the plurality of resources determined to be compatible.

19. The method of claim 18 further comprising enabling communication between the job and the one or more of the plurality of resource manager components by providing a plurality of resource manager adapter components communicatively linked with the plurality of resource manager components.

20. The method of claim 18 further comprising providing a metascheduler component, and enabling the job to query the metascheduler for scheduling suggestions.

21. The method of claim 18 further comprising providing a reservation service component, and enabling the job to query the reservation service component to reserve one or more of the plurality of resources within the networked computing grid.