System and method for providing a linearizable request manager

• US 8,930,584 B2
• Filed: 08/09/2012
• Issued: 01/06/2015
• Est. Priority Date: 08/09/2012
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Otenko, Oleksandr, Agarwal, Prashant
• Primary Examiner(s): Park, Ilwoo
• Application Number:
  US13/570,434
• Publication Number:
  US 20140047140A1
• Time in Patent Office: 880 Days
• US Class Current: 710/5
• CPC Class Codes: G06F 2209/5013 Request control
• CPC Class Codes: G06F 2209/5021 Priority
• CPC Class Codes: G06F 9/5027 the resource being a machine, e.g. CPUs, Servers, Terminals

Abstract

Described herein are systems and methods for improving concurrency of a request manager for use in an application server or other environment. A request manager receives a request, and upon receiving the request the request manager associates a token with the request. A reference to the request is enqueued in each of a plurality of queues, wherein each queue stores a local copy of the token. A first reference to the request is dequeued from a particular queue, wherein when the first reference to the request is dequeued, the token is modified to create a modified token. Thereafter the request is processed. When other references to the request are dequeued from other queues, the other references to the request are discarded.

Claims

1. A method for improving concurrency of a request manager for use in an application server, comprising:

providing a request manager to execute on one or more microprocessors;

receiving a request by the request manager, wherein the request manager associates a token with the request;

enqueuing a reference to the request in each of a plurality of queues, wherein each queue stores a local copy of the token;

dequeuing a first reference to the request from a particular queue, wherein when the first reference to the request is dequeued, the request is processed, and the token is modified to create a modified token; and

wherein when other references to the request are dequeued from other queues, the other references to the request are discarded.

2. The method of claim 1, further comprising:

providing a priority queue, a maximum constraint queue, and a minimum constraint queue.

3. The method of claim 1, wherein dequeuing the first reference to the request further includes:

comparing the token associated with the request to the local copy of the token, wherein when the token associated with the request matches the local copy of the token, the request is claimed by the particular queue.

4. The method of claim 1, wherein dequeuing the request from other queues further includes:

dequeuing the request by a different queue;

comparing the modified token associated with the request in the different queue to the local copy of the token;

determining, based on the comparison, that the modified token associated with the request does not match the local copy of the token; and

discarding the request from the different queue.

5. The method of claim 1, further comprising:

providing a first in first out (FIFO) queue, wherein the FIFO queue receives a plurality of requests;

providing a priority queue to receive the plurality of requests from the FIFO queue, wherein the priority queue receives requests from the FIFO queue one request at a time;

acquiring a first lock on the priority queue;

enqueuing one of the plurality of request in the priority queue; and

releasing the first lock on the priority queue.

6. The method of claim 5, further comprising:

acquiring a second lock on the priority queue;

dequeuing the one of the plurality of requests from the priority; and

releasing the second lock on the priority queue.

7. The method of claim 5, further comprising:

providing a dedicated thread to enqueue the plurality of requests from the FIFO queue in the priority queue, wherein a number of idle worker threads remove the plurality of requests from the priority queue for processing.

8. The method of claim 1, further comprising:

associating a function with the request, wherein the function, when executed by the queue, causes a change in a state of the queue;

executing the function by the queue when the request is enqueued; and

modifying the state of the queue based on the function.

9. The method of claim 8, further comprising:

attempting to claim the token associated with the request by a worker thread; and

reverting the change of the state of the queue caused by executing the function if the claim to the token fails;

otherwise, executing the request by the worker thread.

10. A system for improving concurrency of a request manager, comprising:

one or more microprocessors;

a request manager that executes on said one or more microprocessors, wherein the request manager receives a request, and wherein the request manager associates a token with the request;

a plurality of queues, wherein the request manager enqueues a reference to the request in each of the plurality of queues, and wherein each queue stores a local copy of the token;

a plurality of threads that execute the requests from the plurality of queues, wherein one of the plurality of threads dequeues a first reference to the request from a particular queue, wherein when the first reference to the request is dequeued, the request is processed, and the token is modified to create a modified token; and

wherein when other references to the request are dequeued from other queues, the other references to the request are discarded.

11. The system of claim 10, wherein the plurality of queues is one of a priority queue, a maximum constraint queue, and a minimum constraint queue.

12. The system of claim 10, wherein when the first reference to the request is dequeued, the token associated with the request is compared to the local copy of the token, and wherein when the token associated with the request matches the local copy of the token, the request is claimed by the particular queue.

13. The system of claim 10, wherein when a second reference to the request is dequeued by a different queue, the different queue

compares the modified token associated with the request in the different queue to the local copy of the token;

determines, based on the comparison, that the modified token associated with the request does not match the local copy of the token; and

discards the request from the different queue.

14. The system of claim 10, further comprising:

a first in first out (FIFO) queue, wherein the FIFO queue receives a plurality of requests;

a priority queue to receive the plurality of requests from the FIFO queue, wherein the priority queue receives requests from the FIFO queue one request at a time, and wherein when a dedicated thread attempts to enqueue the request in the priority queue from the FIFO queue, the dedicated thread

acquires a first lock on the priority queue, and

enqueues one of the plurality of request in the priority queue; and

wherein when the request is enqueued, the first lock on the priority queue is released.

15. The system of claim 14, wherein when a worker thread attempts to dequeue the request from the priority queue, the worker thread:

acquires a second lock on the priority queue;

dequeues the one of the plurality of requests from the priority; and

wherein when the one of the plurality of requests request is dequeued, the second lock on the priority queue is released.

16. The system of claim 14, further comprising:

a dedicated thread to enqueue the plurality of requests from the FIFO queue into the priority queue, wherein a number of idle worker threads remove the plurality of requests from the priority queue.

17. The system of claim 10, wherein the request manager further:

associates a function with the request, wherein the function, when executed by the queue, causes a change in a state of the queue; and

enqueues the function with the request, wherein the function is executed by the queue when the request is enqueued, and wherein when the request is executed, a state of the queue is modified based on the function.

18. The system of claim 17, wherein a worker thread attempts to claim the token associated with the request, wherein if the claim to the token fails, the change of the state of the queue reverts back;

otherwise, the request is executed by the worker thread.

19. A non-transitory computer readable storage medium storing one or more sequences of instructions for improving concurrency of a request manager for use in an application server, wherein said instructions, when executed by one or more processors, cause the one or more processors to execute steps comprising:

providing a request manager to execute on one or more microprocessors;

receiving a request by the request manager, wherein the request manager associates a token with the request;

enqueuing a reference to the request in each of a plurality of queues, wherein each queue stores a local copy of the token;

dequeuing a first reference to the request from a particular queue, wherein when the first reference to the request is dequeued, the request is processed, and the token is modified to create a modified token; and

wherein when other references to the request are dequeued from other queues, the other references to the request are discarded.

20. The non-transitory computer readable storage medium of claim 19, further comprising:

comparing the token associated with the request to the local copy of the token, wherein when the token associated with the request matches the local copy of the token, the request is claimed by the particular queue.