Cluster architecture having a star topology with centralized services

• US 8,190,780 B2
• Filed: 12/30/2003
• Issued: 05/29/2012
• Est. Priority Date: 12/30/2003
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: SAP AG (SAP SE)
• Inventors: Rohland, Hans-Christoph, Kilian, Frank
• Primary Examiner(s): Pwu, Jeffrey
• Assistant Examiner(s): Ali, Farhad
• Application Number:
  US10/750,002
• Publication Number:
  US 20050188021A1
• Time in Patent Office: 3,073 Days
• Field of Search: 709/206, 709/252
• US Class Current: 709/252
• CPC Class Codes: H04L 51/23 Reliability checks, e.g. acknowledgments or fault reporting
• CPC Class Codes: H04L 67/10 in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/1034 Reaction to server failures by a load balancer
• CPC Class Codes: H04L 67/566 Grouping or aggregating service requests, e.g. for unified processing
• CPC Class Codes: H04L 67/567 Integrating service provisioning from a plurality of service providers

Abstract

A cluster of architecture having a star topology with the central services node at its center. Application server instances are organized in a star topology with the central services node at its center. The central services node may include services such as a messaging server for interinstance cluster communications. A locking server may also be provided to provide cluster wide locking to facilitate changes and updates within the cluster. A database may also be shared by all instances in the cluster, thereby reducing the need for data replication. In one embodiment, the message server has no persistent state. In such an embodiment, if the message server fails, it can merely be restarted without any state recovery requirement.

Claims

1. A system comprising:

a message server that is configured to be restarted after a failure without performing state recovery operations; and

a plurality of instances of an application server coupled in a star topology with the message server at a center of the star topology, the message server handling communications between the plurality of instances of the application server, one or more of the plurality of instances to register or reregister instance-specific information with the message server upon a starting or restarting, respectively, of the message server, the instance-specific information identifying one or more services that the one or more of the plurality of instances is configured to provide to each of the plurality of instances.

2. The system of claim 1 wherein each of the plurality of instances comprises:

a dispatcher node; and

a plurality of server nodes.

3. The system of claim 2 wherein each of the plurality of server nodes comprises a java 2 enterprise edition (J2EE) engine.

4. The system of claim 1 further comprising a central lock server to provide cluster wide locks to the plurality of instances.

5. The system of claim 1 wherein the message server comprises:

a first data structure to store a list of connected clients; and

a second data structure to store a list of services provided in the system.

6. The system of claim 1, wherein each of the plurality of instances is started using a first instance-specific bootstrap logic, the first instance-specific bootstrap logic synchronized with a second instance-specific bootstrap logic stored in the database.

7. The system of claim 1, wherein a node within the plurality of instances is started using a first node-specific bootstrap logic, the first node-specific bootstrap logic synchronized with a second node-specific bootstrap logic stored in the database.

8. The system of claim 1, wherein the plurality of instances are unable to communicate with each other during a failure of the message server.

9. A non-transitory computer readable storage media containing executable computer program instructions which when executed cause a digital processing system to perform a method comprising:

starting a central services node to provide a locking service and a messaging service, the messaging service configured to be restarted after a failure without performing state recovery operations, the messaging service handling communications between a plurality of application server instances;

starting the plurality of application server instances;

organizing the application server instances into a cluster having star topology with the central services node at a center of the star topology; and

registering or reregistering instance-specific information with the central services node upon starting or restarting, respectively, of the central services node, the registering or registering initiated by one or more of the plurality of application server instances, the instance-specific information identifying one or more services that the one or more of the plurality of application server instances are configured to provide to each of the plurality of application server instances.

10. The non-transitory computer readable storage media of claim 9, the method further comprising sharing a database among the plurality of application server instances.

12. The non-transitory computer readable storage media of claim 9, the method further comprising starting a message server having no persistent state.

13. The non-transitory computer readable storage media of claim 12, the method further comprising restarting the message server without state recovery responsive to a system failure.

14. The non-transitory computer readable storage media of claim 9, the method further comprising conducting inter instance communication through the messaging service.

11. The non-transitory computer readable storage media of 9, the method wherein starting a plurality of application server instances comprises starting, for each application server instance of the plurality, a dispatcher node and a plurality of server nodes.

15. A system comprising:

means for organizing a plurality of application servers instances into a cluster having a star topology with a central services node at a center of the star topology;

means for sharing a storage resource across the cluster; and

means for performing centralized inter instances communication that is configured to be restarted after a failure without performing state recovery operations, the inter instances communication including registering or reregistering of instance-specific information with the central services node upon a starting or restarting, respectively, of the central services node, the registering or reregistering initiated by one or more of the plurality of application server instances, the instance-specific information identifying one or more services, that the one or more of the plurality of application server instances are configured to provide to each of the plurality of application server instances.

16. The system of claim 15 further comprising means for centrally locking a resource within the cluster.

17. The system of claim 15 wherein the means for performing comprises a message server having no persistent state.

18. The system of claim 15 wherein the means for performing comprises means for recording services provided in the cluster.

19. A method comprising:

starting a central services node to provide a locking service and a messaging service, the messaging service being configured to be restarted after a failure without performing state recovery operations, the messaging service handling communications between a plurality of application server instances;

starting the plurality of application server instances;

organizing the plurality of application server instances into a cluster having a star topology with the central services node at a center of the star topology; and

registering or reregistering instance-specific information with the central services node upon a starting or a restarting, respectively, of the central services node, the instance-specific information identifying one or more services the one or more of the plurality of application server instances are configured to provide to each of the plurality of application server instances.

20. The method of claim 19 further comprising sharing a database among the plurality of application server instances.

21. The method of claim 19 wherein starting a plurality of application server instances comprises starting, for each instance of the plurality, a dispatcher node and a plurality of server nodes.

22. The method of claim 19 wherein the messaging service has no persistent state.

23. The method of claim 22 further comprising restarting the message server without state recovery responsive to a system failure.

24. The method of claim 19 further comprising conducting inter instance communication through the messaging service.

25. The system of claim 19, further comprising notifying each of the plurality of application server instances of the registering or reregistering of the instance-specific information.

26. The method of claim 19, wherein the instance-specific information further includes information about a new service that the one or more of the plurality of instances provide.