Techniques for establishing a parallel processing framework for a multi-tenant on-demand database system

• US 8,443,366 B1
• Filed: 12/11/2009
• Issued: 05/14/2013
• Est. Priority Date: 12/11/2009
• Status: Active Grant

• Current Assignee: Salesforce.com, Inc.
• Original Assignee: Salesforce.com, Inc.
• Inventors: Yancey, Scott
• Primary Examiner(s): KAWSAR, ABDULLAH AL
• Application Number:
  US12/636,675
• Time in Patent Office: 1,250 Days
• Field of Search: 718 1-106, 709223-226, 707713-721, 707/764, 707769-771
• US Class Current: 718/102
• CPC Class Codes: G06F 2209/5017 Task decomposition
• CPC Class Codes: G06F 9/5061 Partitioning or combining of resources

Abstract

In accordance with embodiments, there are provided techniques for establishing a parallel processing framework in a multi-tenant on-demand database system. These techniques for establishing a parallel processing framework in a multi-tenant on-demand database system may enable embodiments to provide great flexibility to a tenant of the architecture to perform desired functions on content of the database without unduly consuming the resources of the system.

Claims

1. A method of establishing a parallel processing framework for a multi-tenant on-demand database system, the method comprising:

processing computational jobs;

storing data related to said computational jobs in a queue table of a database including a plurality of job types;

identifying, among said data, information concerning a plurality of job classes which maps to a plurality of job types that are associated with the computational jobs to be processed and also identifying one of said computational jobs that may be partitioned as a plurality of tasks and a job type of the one of said computational jobs, such that each of said tasks may be processed independent of the remaining tasks of said plurality of tasks, wherein the plurality job classes defines a plurality of job handlers for processing the computational jobs and initiating a job handler appropriate for processing the one of said computational jobs based on the job type; and

wherein, upon completion of the one of said computational jobs, termination of the job handler of the one of said computational jobs is dependent upon historical information concerning historical computational requirements of a job type, present in the queue table, corresponding to the job type of the one of said computation jobs.

2. The method as recited in claim 1 wherein identifying further comprises identifying information that defines a plurality of partitions each of which is associated with a subset of data on said database, the subset of data associated with one of said plurality of partitions being different from the subset of data associated with the remaining partitions of said plurality of partitions.

3. The method as recited in claim 1 wherein identifying further comprises identifying information that defines a plurality of partitions each of which is associated with a subset of data on said database, the subset of data associated with one of said plurality of partitions being different from the subset of data associated with the remaining partitions of said plurality of partitions, with the subset of data of said partitions being concurrently accessible by a subgroup of the remaining computational jobs.

4. The method as recited in claim 1 wherein processing further comprises processing each of said plurality of tasks as a thread of execution defining multiple threads of execution, with said multiple threads of execution being concurrently executed during said processing.

5. The method as recited in claim 1 wherein processing further comprises initiating a job handler and an enqueue/dequeue processes, in data communication with said job handler, to process said one of said computational jobs.

6. The method as recited in claim 5 further comprising initiating additional instances of said job handler in response to resource utilization of said database.

7. The method as recited in claim 1 wherein processing further comprise executing a plurality of job handlers and enqueuing/dequeuing process with a subset of said plurality of job handlers configured to operate upon a job type that is different than the job type upon which the remaining job handlers of said plurality of job handlers are configured to operate upon.

8. The method as recited in claim 1 wherein processing further comprises executing a plurality of job handlers and enqueuing/dequeuing processes with each of said plurality of job handlers configured to operate upon a job type that is different than the job type upon which the remaining job handlers of said plurality of job handlers are configured to operate upon.

9. A non-transitory machine-readable medium storing one or more sequences of instructions for establishing a parallel processing framework for a multi-tenant on-demand database system, when executed by one or more processors, cause said one or more processors to:

process computational jobs;

store data related to said computational jobs in a queue table of a database including a plurality of job types;

identify, among said data, information concerning a plurality of job classes which maps to a plurality of job types that are associated with the computational jobs to be processed and also identifying one of said computational jobs that may be partitioned as a plurality of tasks and a job type of the one of said computational jobs, such that each of said tasks may be processed independent of the remaining tasks of said plurality of tasks, wherein the plurality job classes defines a plurality of job handlers for processing the computational jobs and initiating a job handler appropriate for processing the one of said computational jobs based on the job type; and

wherein, upon completion of the one of said computational jobs, termination of the job handler of the one of said computational jobs is dependent upon historical information concerning historical computational requirements of a job type, present in the queue table, corresponding to the job type of the one of said computation jobs.

10. The machine-readable medium as recited in claim 9 wherein said instructions for identifying further comprises instructions to identify information that defines a plurality of partitions each of which is associated with one of said plurality of partitions being different from the subset of data associated with the remaining partitions of said plurality of partitions.

11. The machine-readable medium as recited in claim 9 wherein said instructions for identifying further comprises instructions to identify information that defines a plurality of partitions each of which is associated with a subset of data on said database, the subset of data associated with one of said plurality of partitions being different from the subset of data associated with the remaining partitions of said plurality of partitions, with the subsets of data of said partitions being concurrently accessible by a subgroup of the remaining computational jobs.

12. The machine-readable medium as recited in claim 9 wherein said instructions for processing further comprises instructions to process each of said plurality of tasks as a thread of execution defining multiple threads of execution, with said multiple threads of execution being concurrently executed during said processing.

13. The machine-readable medium as recited in claim 9 wherein said instructions for processing further comprises instructions to initiate a job handler and an enqueue/dequeue processes, in data communication with said job handler, to process said one of said computational jobs.

14. The machine-readable medium as recited in claim 9 wherein said instructions for processing further comprises instructions to execute a plurality of job handlers and enqueuing/dequeuing processes with a subset of said plurality of job handlers configured to operate upon a job type that is different than the job type upon which the remaining job handlers of said plurality of job handlers are configured to operate upon.

15. A system for establishing a parallel processing framework in a multi-tenant database, said system comprising:

a processor; and

a memory having a sequence of instructions which, when executed by the processor, cause the processor to;

process computational jobs;

store data related to said computational jobs in a queue table of a database including a plurality of job types;

identify, among said data, information concerning a plurality of job classes which maps to a plurality of job types that are associated with the computational jobs to be processed and also identifying one of said computational jobs that may be partitioned as a plurality of tasks and a job type of the one of said computational jobs, such that each of said tasks may be processed independent of the remaining tasks of said plurality of tasks, wherein the plurality job classes defines a plurality of job handlers for processing the computational jobs and initiating a job handler appropriate for processing the one of said computational jobs based on the job type; and

wherein, upon completion of the one of said computational jobs, termination of the job handler of the one of said computational jobs is dependent upon historical information concerning historical computational requirements of a job type, present in the queue table, corresponding to the job type of the one of said computation jobs.