System and method for supporting cooperative concurrency in a middleware machine environment

• US 9,378,045 B2
• Filed: 02/28/2013
• Issued: 06/28/2016
• Est. Priority Date: 02/28/2013
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Otenko, Oleksandr
• Primary Examiner(s): Nguyen, Van
• Application Number:
  US13/781,475
• Publication Number:
  US 20140245312A1
• Time in Patent Office: 1,216 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 9/46 Multiprogramming arrangements
• CPC Class Codes: G06F 9/524 Deadlock detection or avoidance
• CPC Class Codes: G06F 9/526 Mutual exclusion algorithms
• CPC Class Codes: G06F 9/546 Message passing systems or structures, e.g. queues

Abstract

A system and method can support cooperative concurrency in a priority queue. The priority queue, which includes a calendar ring and a fast lane, can detect one or more threads that contend to claim one or more requests in the priority queue. Then, a victim thread can place a request in the fast lane in the priority queue, and release a contending thread, which proceeds to consume the request in the fast lane.

Claims

1. A method for supporting cooperative concurrency in a priority queue of a middleware system, the priority queue including a calendar ring area that queues requests according to target response time and a lane area that queues requests moved from the calendar ring area to the lane area due to contention between consumers, the method comprising, with the middleware system, performing the method of:

detecting that a first thread that is associated with a first consumer and a second thread that is associated with a second consumer contend to claim a request in the priority queue; and

handling the contention without locks to expose the contention, the handling including;

placing, by the first thread, the request from the calendar ring area that is under contention into the lane area of the priority queue;

releasing, by the first thread, the second thread; and

consuming the request in the lane area, by the second thread that is associated with the second consumer.

2. The method according to claim 1, wherein the requests stored in the calendar ring area have a target response time less than a pre-configured quality of service (QoS).

3. The method according to claim 2, further comprising:

specifying the pre-configured quality of service (QoS) to be two seconds.

4. The method according to claim 2, further comprising:

using a bit map updated in synchronization with the calendar ring area for faster scanning.

5. The method according to claim 2, further comprising:

associating a list area of the priority queue with the calendar ring area, wherein the list area operates to store requests with target response time longer than the pre-configured quality of service (QoS).

6. The method according to claim 5, further comprising:

promoting a request in the list area into the calendar ring area, when no request can be found in the calendar ring area.

7. The method according to claim 2, further comprising:

using a list of requests to reduce contention on the calendar ring area when workload is intensive, wherein the list of requests contains one or more requests that land on a same calendar entry.

8. The method according to claim 1, further comprising:

using a sequencer to order the contending threads.

9. The method according to claim 1, further comprising:

proceeding, by the first thread, to claim another request in the calendar ring area.

10. The method according to claim 1, wherein:

supporting continuation-passing when the priority queue is empty.

11. A system for supporting cooperative concurrency in a priority queue, comprising:

one or more microprocessors of a middleware system;

a priority queue that runs on the one or more microprocessors, wherein the priority queue includes a calendar ring area that queues requests according to target response time and a lane area that queues requests moved from the calendar ring area to the lane area due to contention between consumers, and the priority queue operates to;

detect that a first thread that is associated with a first consumer and a second thread that is associated with a second consumer contend to claim a request in the priority queue; and

handle the contention without locks to expose the contention, the handling including;

placing, by the first thread, the request from the calendar ring area that is under contention into the lane area of the priority queue;

releasing, by the first thread, the second thread; and

consuming the request in the lane area, by the second thread that is associated with the second consumer.

12. The system according to claim 11, wherein requests stored in the calendar ring area have a target response time less than a pre-configured quality of service (QoS).

13. The system according to claim 12, wherein:

the pre-configured quality of service (QoS) is specified to be two seconds.

14. The system according to claim 12, wherein:

a bit map updated in synchronization with the calendar ring area is used for faster scanning.

15. The system according to claim 12, wherein:

the calendar ring area is associated with a list area of the priority queue, wherein the list area operates to store requests without the pre-configured quality of service (QoS).

16. The system according to claim 15, wherein:

a request in the outlier area is promoted into the calendar ring area, when no request can be found in the calendar ring.

17. The system according to claim 12, wherein:

a list of requests is used to reduce contention on the calendar ring area when workload is intensive, wherein the list of requests contains one or more requests that land on a same calendar entry.

18. The system according to claim 11, wherein:

a sequencer is used to order the contending threads.

19. The system according to claim 11, wherein:

the first thread proceeds to claim another request in the calendar ring area.

20. A non-transitory machine readable storage medium having instructions stored thereon that when executed cause a middleware system having a priority queue, the priority queue including a calendar ring area that queues requests according to target response time and a lane area that queues requests moved from the calendar ring area to the lane area due to contention between consumers, to perform the method comprising:

detecting that a first thread that is associated with a first consumer and a second thread that is associated with a second consumer contend to claim a request in the priority queue; and

handling the contention without locks to expose the contention, the handling including;

placing, by the first thread, the request from the calendar ring area that is under contention into the lane area of the priority queue;

releasing, by the first thread, the second thread; and

consuming the request in the lane area, by the second thread that is associated with the second consumer.