Method and computer system for queue processing

• US 20060129660A1
• Filed: 11/10/2005
• Published: 06/15/2006
• Est. Priority Date: 11/12/2004
• Status: Abandoned Application

• Current Assignee: SAP AG (SAP SE)
• Original Assignee: SAP AG (SAP SE)
• Inventors: Mueller, Wolfgang G., Willy, Hartmut
• Application Number:
  US11/272,047
• Publication Number:
  US 20060129660A1
• Time in Patent Office: Days
• US Class Current: 709/220
• CPC Class Codes: G06F 9/546 Message passing systems or structures, e.g. queues

Abstract

The scheduler receives a first message from one of the plurality of computer systems and registers the first message in a first queue. Then, the scheduler receives a second message from one of the plurality of computer systems. The second message has a logical dependency on the first message. The second message is registered in a second queue and in the first queue, wherein the queue entry of the second message in the first queue occurs after the queue entry of the first message in the first queue. The scheduler identifies the logical dependency of the second message on the first message by the order of the corresponding queue entries in the first queue and finally releases the first message for processing before releasing the second message according to the logical dependency.

Claims

1. A computer program product configured to process messages of at least one of a plurality of computer systems in a further computer system of the plurality of computer systems, the computer program product having a plurality of instructions that when loaded into a memory of the further computer system cause at least one processor of the further computer system to execute the following steps:

receiving a first message from one of the plurality of computer systems;

registering the first message in a first queue;

receiving a second message from one of the plurality of computer systems, the second message having a logical dependency on the first message;

registering the second message in a second queue and in the first queue, wherein the queue entry of the second message in the first queue occurs after the queue entry of the first message in the first queue;

identifying the logical dependency of the second message on the first message by the order of the corresponding queue entries in the first queue; and

releasing the first message for processing before releasing the second message according to the logical dependency.

2. The computer program product of claim 1, configured to cause the at least one processor to execute the further steps:

receiving a third message from one of the plurality of computer systems after having received the first and second messages, the third message being logically independent from the first and second messages;

registering the third message in a third queue; and

releasing the third message for processing before the second message.

3. The computer program product of claim 2, configured to cause a further processor of the further computer system to release the third message to be processed in parallel with the first message.

4. The computer program product of claim 1, wherein the registering steps provide a unique timestamp for each received message and the queue entries are registered strictly monotonous over time according to the timestamp of the corresponding message.

5. The computer program product of claim 1, wherein the identifying step uses a topological sort algorithm to identify the logical dependency between the first and second messages within the first queue.

6. The computer program product of claim 5, wherein the topological sort algorithm is applied to further queues to which a message has registered to identify further logical dependencies.

7. The computer program product of claim 1, wherein the logical dependency is defined by an object that is affected when the first message is processed and is affected when the second message is processed.

8. The computer program product of claim 1, comprising further instructions that cause the at least one processor to execute the steps of:

after a restart of the computer program, receiving a plurality of persisted messages from persistence means that corresponds to the plurality of messages registered before the restart; and

merging the plurality of the persisted messages with messages that are already registered in corresponding queues after the restart by registering the persisted messages in the corresponding queues so that all logical dependencies are taken into account.

17. A computer system configured to process messages of at least one of a plurality of computer systems, comprising:

a memory storing a computer program product according to claim 1;

and

at least one processor configured to execute the computer program.

9. A method for processing messages of at least one of a plurality of computer systems in a further computer system of the plurality of computer systems, comprising the steps:

receiving a first message from one of the plurality of computer systems;

registering the first message in a first queue;

receiving a second message from one of the plurality of computer systems, the second message having a logical dependency on the first message;

registering the second message in a second queue and in the first queue, wherein the queue entry of the second message in the first queue occurs after the queue entry of the first message in the first queue;

identifying the logical dependency of the second message on the first message by the order of the corresponding queue entries in the first queue; and

releasing the first message for processing before releasing the second message according to the logical dependency.

10. The method of claim 9 comprising the further steps:

receiving a third message from one of the plurality of computer systems after having received the first and second messages, the third message being logically independent from the first and second messages;

registering the third message in a third queue; and

releasing the third message for processing before the second message.

11. The method of claim 10, wherein the third message is released before the second message to be processed in parallel with the first message.

12. The method of claim 9, wherein the registering steps provide a unique timestamp for each received message and the queue entries are registered strictly monotonous over time according to the timestamp of the corresponding message.

13. The method of claim 9, wherein the identifying step uses a topological sort algorithm to identify the logical dependency between the first and second messages within the first queue.

14. The method of claim 13, wherein the topological sort algorithm is applied to further queues to which a message has registered to identify further logical dependencies.

15. The method of claim 9, wherein the logical dependency is defined by an object that is affected when the first message is processed and is affected when the second message is processed.

16. The method of claim 9, comprising the further steps of:

restarting a computer program that already has performed the receiving message and registering message steps for a plurality of messages;

receiving from persistence means a plurality of persisted messages that corresponds to the plurality of messages that were registered before the restarting step; and

merging the plurality of the persisted messages with messages that are already registered in corresponding queues after the restarting step by registering the persisted messages in the corresponding queues so that all logical dependencies are taken into account.