HIGH PERFORMANT INFORMATION SHARING AND REPLICATION FOR SINGLE-PUBLISHER AND MULTIPLE-SUBSCRIBER CONFIGURATION

• US 20100198920A1
• Filed: 01/30/2009
• Published: 08/05/2010
• Est. Priority Date: 01/30/2009
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Xu, Haobo, Arora, Nimar, Gao, Lei, Wong, Lik, Hoang, Thuvan
• Granted Patent: US 9,230,002 B2
• Time in Patent Office: Days
• US Class Current: 709/206
• CPC Class Codes: G06F 16/273 Asynchronous replication or reconciliation
• CPC Class Codes: G06F 9/544 Buffers; Shared memory; Pipes
• CPC Class Codes: G06F 9/546 Message passing systems or structures, e.g. queues

Abstract

A method for sharing information between a publisher and multiple subscribers is provided. The publisher uses a latch-free, single publisher, multiple subscriber shared queue to share information. Logical change records representing changes made to a database are enqueued in the shared queue as messages in a stream of messages, and subscribers read the logical change records. Subscribers may filter logical change records before sending to apply processes for processing. An identifying property of the source instance of a change encapsulated in a logical change record may be included with each message enqueued.

Claims

1. A method for propagating messages in a multi-node system comprising multiple nodes, comprising:

at said source node, an enqueuing process enqueuing multiple messages into a shared queue without acquiring a latch on said shared queue;

at said source node, each dequeuing process of multiple dequeueing processes dequeuing multiple messages from said shared queue, wherein each dequeuing process dequeues said multiple messages without aquiring a latch on said queue.

2. The method of claim 1, wherein for each dequeuing process of said multiple dequeuing processes, said each dequeuing process communicating a message dequeued from said shared queue to a destination node different than said source node.

3. The method of claim 2, wherein each dequeuing process communicates a dequeued message to a different destination node different than the respective destination node of another dequeuing processes of said plurality of dequeuing processes.

4. The method of claim 1,

wherein for a first dequeuing process of said multiple dequeuing processes:

said first dequeuing process dequeuing a particular message from said shared queue,

that said particular message does not satisfy first one or more criteria,

in response to determining that a particular message does not satisfy one or more criteria, forgoing forwarding said particular message to a destination node; and

wherein for a second dequeuing process of said multiple dequeuing processes;

said second dequeuing process dequeuing said particular message from said shared queue;

determining that said particular message does satisfy second one or more criteria; and

in response to determining that a particular message does not satisfy second one or more criteria, forwarding said particular message to a destination node.

5. The method of claim 1, wherein each apply process applies changes of a message to a database at a logical level.

6. The method of claim 1, wherein each message includes information that identifies an instance that is the source of changes represented by said message.

7. The method of claim 1, wherein the each dequeuing process is guaranteed to completely perform a dequeuing operation on said shared queue within a bounded number of steps whenever there is at least one entry in said shared queue.

19. A volatile or non-volatile medium storing one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 1.

20. A volatile or non-volatile medium storing one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 2.

21. A volatile or non-volatile medium storing one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 3.

22. A volatile or non-volatile medium storing one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 4.

23. A volatile or non-volatile medium storing one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 7.

8. A method for distributing messages in a multi-node database system, said multi-node database system comprising a certain node, said method comprising:

a process on said certain node enqueuing each message of a plurality of messages to a shared queue;

each propagation sender of a plurality of propagation senders dequeuing each message of said plurality of messages from said shared queue;

each propagation sender sending a respective one or more messages of said plurality of messages to a respective apply process on a second node of said multi-node system; and

wherein each respective apply process applies said respective one or more messages to a database.

9. The method of claim 8, further comprising each respective propagation sender of said plurality of propagation senders advancing a respective head pointer to point to a next message after dequeuing a message of said plurality of messages.

10. The method of claim 8, further comprising a propagation sender of said plurality of propagation senders waking said enqueuing process if no other propagation sender has a full logical queue.

11. The method of claim 8, wherein each propagation sender of said plurality of propagation senders maintains a head pointer for reading messages from said shared queue.

12. The memory of claim 8, wherein said enqueuing process waits if any propagation sender has a full logical queue.

13. The method of claim 8, further comprising each apply process of said plurality of apply processes publishing a low watermark, wherein said low watchmarks may be used to determine a global low watermark.

14. The method of claim 8, wherein a logical representation of said shared queue is a set of circular queues.

15. The method of claim 8, wherein the enqueuing process is guaranteed to completely perform an enqueue operation on said shared queue within a bounded number of steps whenever the shared queue is not full.

16. The method of claim 8, wherein said message includes information that identifies an instance at a source node that is the source of changes represented by said message.

17. The method of claim 8, further comprising:

in response to a termination of an instance at a destination node, sending messages representing changes that were made by an instance at a source node to an apply process, wherein prior to termination of the instance at the destination node, the messages containing representing changes that were made by the instance at the source node were not sent to the apply process.

18. The method of claim 8, further comprising:

wherein said multi-node system comprises one or more clustered database systems;

wherein said certain node is logically part of a source clustered database system;

wherein said database is logically part of a destination clustered database system.

24. A volatile or non-volatile medium storing one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 8.

25. A volatile or non-volatile medium storing one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 9.