System and method for providing timer affinity through notifications within a session-based server deployment

• US 9,723,048 B2
• Filed: 10/29/2008
• Issued: 08/01/2017
• Est. Priority Date: 10/29/2008
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Cosmadopoulos, Ioannis, Kulkarni, Mihir Arvind, Larson, Johan Gustav
• Primary Examiner(s): Murray, Daniel C.
• Application Number:
  US12/260,846
• Publication Number:
  US 20100106842A1
• Time in Patent Office: 3,198 Days
• Field of Search: 709228
• CPC Class Codes: H04L 65/80 Responding to QoS
• CPC Class Codes: H04L 67/5682 Policies or rules for updating, deleting or replacing the stored data
• CPC Class Codes: H04L 69/28 Timers or timing mechanisms used in protocols

Abstract

The session-based server deployment can be comprised of an engine tier that is designed for high throughput processing of communications and a state tier that maintains state information for access by the engine tier. The engine tier can include a plurality of engine nodes that process incoming messages. The state nodes can contain call state, including various timers. When processing a call, an engine can set various timers in the state tier. Upon expiration of a timer, the state node can transmit a notification assigning the timer to the corresponding engine node which handled previous communications for the call. This affinity between the timer and its respective engine can provide numerous benefits, such as reduced latency and more localized data access.

Claims

1. A telecommunication system that provides timer affinity through notifications, said telecommunication system comprising:

an engine tier comprising a plurality of computing devices and a plurality of engine nodes executing on said plurality of computer devices that process a plurality of communications associated with a plurality of sessions;

a state tier comprising a plurality of state nodes that store a plurality of session states associated with said plurality of sessions;

a plurality of timer objects associated with the plurality of sessions, wherein the plurality of timer objects are initiated by the plurality of engine nodes upon having processed said plurality of communications, and stored on said state tier; and

a notification thread deployed and running on each state node of said plurality of state nodes, wherein

the notification thread periodically looks for expired timer objects from among said plurality of timer objects stored on the state node,

the notification thread upon detecting an expired timer object transmits a notification, independent of engine tier polling, to a particular engine node of said plurality of engine nodes that is determined to have initiated the expired timer object; and

the notification thread thereby creates affinity between the expired timer object and the particular engine node.

2. The telecommunication system of claim 1, wherein:

each of said plurality of engine nodes has a unique identifier which is associated with timer objects initiated by each of said plurality of engine nodes; and

wherein the notification thread utilizes the unique identifier associated with the expired timer object to transmit the notification to the particular engine node of said plurality of engine nodes.

3. The telecommunication system of claim 1, further comprising:

a plurality of local near-caches in the engine tier associated with each of said plurality of engine nodes, wherein each local near-cache of the plurality of local near-caches temporarily maintains call state information associated with sessions processed by an associated engine node.

4. The telecommunication system of claim 3, wherein processing the expired timer object on the particular engine node increases a probability that call state information for a session associated with the expired timer object is available in a local near-cache associated with the particular engine node.

5. The telecommunication system of claim 1, wherein the engine tier further includes:

a hot standby deployment of two or more engine nodes, wherein at least one but not all of said two or more engine nodes is redundant and does not process any messages.

6. The telecommunication system of claim 1, further comprising:

a partition client residing on the state tier, said partition client monitoring health information of the plurality of engine nodes, wherein, if the partition client determines that an engine node has failed, the partition client reassigns timer objects associated with the failed engine node to a different engine node.

7. The telecommunication system of claim 1, wherein engine tier and the state tier comprise a session initiation protocol (SIP) application server.

8. The telecommunication system of claim 1, wherein:

the plurality of state nodes of the state tier are configured to reduce polling of timer objects by the engine nodes by determining whether there exist any timer objects associated with the particular engine node from among said plurality of timer objects; and

setting a flag indicating that polling is optional for said particular engine at a next polling interval if there are no existing timer objects associated with said particular engine.

9. The telecommunications system of claim 8, wherein:

said particular engine polls the state tier after a predetermined number of polling intervals regardless of the flag.

10. A method for providing timer affinity within a session-based server deployment operating on an engine tier comprising a plurality of engine nodes executing on a plurality of computing devices and a state tier comprising a plurality of state nodes, said method comprising:

processing a plurality of communications associated with a plurality of sessions using the plurality of engine nodes;

storing a plurality of session states associated with the plurality of sessions on the state tier;

initiating a plurality of timer objects associated with the plurality of sessions, wherein said plurality of timer objects is initiated by the plurality of engine nodes upon having processed said plurality of communications, wherein the plurality of timer objects are stored on said state tier;

running a notification thread on each state node of said plurality of state nodes;

periodically looking using the notification thread for expired timer objects from among said plurality of timer objects stored on said each state node;

transmitting a notification from the notification thread to a particular engine node of said plurality of engine nodes, independently of state tier polling by the particular engine node, that is determined to have initiated an expired timer object upon detecting the expired timer object; and

thereby using the notification thread to create affinity between the expired timer object and the particular engine node.

11. The method of claim 10, further comprising:

associating each of said plurality of engine nodes with a unique identifier which is associated with timer objects initiated by said each of said plurality of engine nodes; and

utilizing the unique identifier associated with the expired timer object to enable the notification thread to transmit the notification to the particular engine node of said plurality of engine nodes.

12. The method of claim 10, further comprising:

maintaining a plurality of local near-caches located in the engine tier associated with each of said plurality of engine nodes, wherein each local near-cache temporarily maintains call state information associated with sessions processed by an associated engine node.

13. The method of claim 12, wherein processing the expired timer object on the particular engine node increases a probability that the call state information for a session associated with the expired timer object is available in a local near-cache associated with the particular engine node.

14. The method of claim 10, wherein the plurality of engine nodes further include:

a hot standby deployment of two or more engine nodes, wherein at least one but not all of said two or more engine nodes is redundant and does not process any messages.

15. The method of claim 10, further comprising:

providing a partition client residing on the state tier wherein said partition client monitors health information of the plurality of engine nodes, wherein, if the partition client determines that an engine node has failed, the partition client reassigns timer objects associated with the failed engine node to a different engine node.

16. The method of claim 10, wherein the engine tier and the state tier comprise a session initiated protocol (SIP) application server.

17. The method of claim 10, further comprising:

determining by each of the plurality of state nodes whether there exist any expired timer objects associated with the particular engine node from among said plurality of timer objects; and

setting a flag indicating that polling is optional for said particular engine at a next polling interval if there are no existing expired timer objects associated with said particular engine.

18. The method of claim 17, further comprising:

said particular engine polling the state tier after a predetermine number of polling intervals regardless of the flag.

19. A non-transitory computer-readable medium storing instructions for providing timer affinity in a session-based server deployment operating on an engine tier comprising a plurality of engine nodes executing on a plurality of computing devices and a state tier comprising a plurality of state nodes, the instructions, when executed by one or more processors, cause the one or more processors to carry out steps comprising:

processing a plurality of communications associated with a plurality of sessions using the plurality of engine nodes;

storing a plurality of session states associated with the plurality of sessions on the state tier;

initiating a plurality of timer objects associated with the plurality of sessions, wherein said plurality of timer objects is initiated by the plurality of engine nodes upon having processed said plurality of communications, wherein the plurality of timer objects are stored on said state tier;

running a notification thread on each state node of said plurality of state nodes;

periodically looking using the notification thread for expired timer objects from among said plurality of timer objects stored on said each state node;

transmitting a notification from the notification thread to a particular engine node of said plurality of engine nodes, independently from the particular engine node polling the state tier, determined to have initiated an expired timer object upon detecting the expired timer object; and

thereby using the notification thread to create affinity between the expired timer object and the particular engine node.

20. The non-transitory computer-readable medium of claim 19, comprising further instructions, which when executed by one or more processors, cause the one or more processors to carry out steps further comprising:

determining by each of the plurality of state nodes whether there exist any expired timer objects associated with the particular engine node from among said plurality of timer objects; and

setting a flag indicating that polling is optional for said particular engine if there are no existing expired timer objects associated with said particular engine.