Autonomic control of a distributed computing system using rule-based sensor definitions

• US 20060173857A1
• Filed: 07/07/2005
• Published: 08/03/2006
• Est. Priority Date: 01/31/2005
• Status: Active Grant

• Current Assignee: Computer Associates Think Inc. (Broadcom, Inc.)
• Original Assignee: Cassatt (Broadcom, Inc.)
• Inventors: Jackson, Jerry R.
• Granted Patent: US 7,454,427 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 2209/505 Clust
• CPC Class Codes: G06F 9/5072 Grid computing

Abstract

A distributed computing system conforms to a multi-level, hierarchical organizational model. One or more control nodes provide for the efficient and automated allocation and management of computing functions and resources within the distributed computing system in accordance with the organization model. A user, such as a system administrator, interacts with the control nodes to logically define the hierarchical organization of distributed computing system. The control node includes an automation subsystem having one or more rule engines that provide autonomic control of the application nodes in accordance with a set of one or more rules. For example, the rules may define sensors or methods of configuring the distributed computing system. More specifically, the sensor rules define the conditions under which the automation subsystem may update a model of the distributed computing system based on data received from the distributed computing system.

Claims

1. A distributed computing system comprising:

a plurality of application nodes interconnected via a communications network; and

a control node that includes an automation subsystem having one or more rule engines that provide autonomic control of the application nodes in accordance with a set of one or more rules,

wherein the rules define one or more sensors that when processed by the rules engine update a model of the application nodes based on status data received from the application nodes.

2. The distributed computing system of claim 1, wherein the rule engines are forward-chaining rule engines.

3. The distributed computing system of claim 1, wherein the automation subsystem includes a rule compiler that compiles the rules into one or more discrimination networks.

4. The distributed computing system of claim 1, wherein the control node further comprises:

a monitoring subsystem that collects the status data from the application nodes and communicates the status data to the automation subsystem, wherein the status data represents an actual state for the application nodes; and

a business logic tier that provides expected state data to the automation subsystem, wherein the expected state data represents an expected state for the application nodes,

wherein the rule engines analyze the status data from the monitoring subsystem and apply the set of rules to produce action requests to the business logic tier to control the application nodes to reduce any difference between the actual state and the expected state.

5. The distributed computing system of claim 4, wherein the monitoring subsystem comprises a set of modular information collectors that collect the status data from the application nodes.

6. The distributed computing system of claim 5, wherein each of the information collectors conforms to a different communication protocol.

7. The distributed computing system of claim 5, wherein the application nodes are organized into tiers, and wherein each of the information collectors collects the status data from the application nodes of different tiers.

8. The distributed computing system of claim 5, wherein the automation unit registers collection requests with the monitoring subsystem, and the monitoring subsystem includes a monitoring engine that invokes the collectors in response to the collection requests.

9. The distributed computing system of claim 4, wherein the automation unit further comprises a sensor subsystem that receives the state data from the monitoring subsystem and alters the model of the application nodes in accordance with the state data and in accordance with the rules defining the sensors.

10. The distributed computing system of claim 9, wherein the sensor subsystem communicates the modified model to the rule engines in accordance with a schedule.

11. The distributed computing system of claim 4, an effector subsystem that outputs action requests to the business logic tier in response to application of the rules by the rule engines.

12. The distributed computing system of claim 4, wherein the business logic tier includes a set of web services to control the application nodes.

13. The distributed computing system of claim 4, wherein the business logic tier presents interfaces by which a user defines one or more goals for the expected state of the application nodes.

14. The distributed computing system of claim 13, wherein the business logic tier registers the goals with the automation unit for achievement by the rule engines in accordance with the rule set.

15. The distributed computing system of claim 1, wherein the rules are units of code that conform with a rule language.

16. The distributed computing system of claim 13, wherein the automation unit automatically formulates dependencies between the goals.

17. The distributed computing system of claim 13, wherein the business logic tier further comprises a task manager to invoke tasks in response to requests from the automation unit, and wherein the business logic tier includes an event subsystem that tracks the progress of each of the tasks.

18. The distributed computing system of claim 13, wherein the goals include at least one of software image management goals, node allocation goals, goals for harvesting nodes, tier capacity goals, asset requirement goals and goals for gathering the status data.

19. The distributed computing system of claim 1, wherein the control node presents interfaces by which a user defines the sensors.

20. The distributed computing system of claim 4, further comprising:

a database storing data that defines a model for a hierarchical organization of the distributed computing system, wherein the model specifies a fabric having a one or more domains, and wherein each domain has at least one tier that includes at least one of the application nodes,

wherein the rule engines automatically produce action requests to configure the application nodes in accordance with a hierarchical organization defined by the model.

21. The distributed computing system of claim 20, wherein each tier represents a different computing function.

22. The distributed computing system of claim 1, wherein each of the rules engines comprise:

a rule compiler to generate a discrimination network based on a set of match conditions supplied by a set of rules;

a set of sensors to update a model of the actual state of the distributed computing system, wherein the set of rules defines each sensor in the set of sensors and wherein each sensor only updates the model when the discrimination network indicates that a set of match conditions associated with the sensor have been satisfied;

a memory to store the model of the actual state; and

an execution engine to perform an action associated with a rule in the set of rules, wherein the discrimination network indicates that the exists a set of match conditions associated with the rule where all match conditions in the set of match conditions are satisfied.

23. A method comprising:

receiving input that defines an expected state for a distributed computing system having a plurality of application nodes interconnected via a communications network;

receiving status data from the distributed computing system regarding an actual state of the nodes;

processing sensor-defining rules with a set of rule engines to determine operations for updating a model of the actual state of the nodes, based on received status data;

processing rules with the set of rule engines to determine operations for reducing any difference between the expected state and the actual state as interpreted by the sensor-defining rules; and

applying the operations to the distributed computing system to control the application nodes in accordance with the rules.

24. The method of claim 23, wherein processing rules comprises processing the rules with a plurality of forward-chaining rule engines that compile the rules into one or more discrimination networks.

25. The method of claim 23, further comprising collecting the status data from the application nodes with a plurality of modular information collectors.

26. The method of claim 23, further comprising:

registering collection requests with a monitoring subsystem; and

executing a monitoring engine that invokes the collectors in response to the collection requests.

27. The method of claim 23, further comprising:

aggregating the state data in accordance with a hierarchical model of deployment of the application nodes; and

processing the aggregated state data with the rule engines.

28. The method of claim 23, further comprising:

presenting an interface by which a user defines the expected state as one or more goals; and

registering the goals with the rule engines.

29. The method of claim 28, further comprising:

storing data that defines a model for a hierarchical organization of the distributed computing system, wherein the model specifies a fabric having a one or more domains, and wherein each domain has at least one tier that includes at least one of the application nodes,

automatically producing the operations with the rule engines to configure the application nodes in accordance with a hierarchical organization defined by the model.

30. The method of claim 23, further comprising:

presenting an interface by which a user provides the sensor-defining rules; and

registering the sensor-defining rules with the rule engines.

31. The method of claim 23, wherein processing sensor-defining rules comprises:

providing the sensor-defining rules to a compiler, wherein each sensor-defining rule comprises a match condition and an action;

receiving a discrimination network from the compiler based on the match condition of each sensor-defining rule;

referencing the discrimination network an a model of the current state of the distributed computing system to determine a set of active sensors;

updating the model of the current state of the distributed computing system with the action specified by a sensor in the set of active sensors when the status data matches a pattern associated with the sensor in the set of active sensors.

32. A computer-readable medium comprising instructions that cause a programmable processor to:

process sensor-defining rules with a set of forward-chaining rule engines to determine the conditions under which to listen for a pattern in a stream of data from components of a distributed computing system and how to modify a model of an actual state of the distributed computing system a what action to perform when the pattern occurs;

process rules with the set of forward-chaining rule engines to determine operations for reducing differences between the model of the actual state of the distributed computing system and an expected state based on the action performed as a result of a sensor-defining rule; and

apply the operations to the distributed computing system to provide autonomic control of the application nodes in accordance with the rules.