Autonomic computing system with model transfer

• US 20120203912A1
• Filed: 04/19/2012
• Published: 08/09/2012
• Est. Priority Date: 04/18/2006
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines SA (International Business Machines Corporation)
• Inventors: Chess, David Michael, Das, Rajashi, Hanson, James Edwin, Segal, Alla, Tesauro, Gerald James, Whalley, Ian Nicholas
• Granted Patent: US 8,554,898 B2
• Time in Patent Office: Days
• US Class Current: 709/226
• CPC Class Codes: H04L 41/145 involving simulating, designing, planning or modelling of a network

Abstract

Methods and systems are provided for autonomic control and optimization of computing systems. A plurality of component models for one or more components in an autonomic computing system are maintained in a system level database. These component models are obtained from a source external to the management server including the components associated with the models. Component models are added or removed from the database or updated as need. A system level management server in communication with the database utilizes the component models maintained in the system level database and generic component models as needed to compute an optimum state of the autonomic computing system. The autonomic computing system is managed in accordance with the computed optimum state.

Claims

1. A method for autonomic system management comprising:

maintaining within an autonomic computing system a plurality of component models, each component model associated with one of a plurality of components in the autonomic computing system, each component model simulating the operation of its associated component to predict a state and performance of that component for various operational conditions of the entire computing system, wherein each component model is supplied to the autonomic computing system by and maintained by its associated component independent of a centralized management server and all of the component models are accessible by the management server, wherein maintaining the component models comprises adding a new component model for each new component added to the autonomic computing system by;

identifying new components added to the autonomic computing system; and

obtaining a corresponding component model from each new component, by;

establishing communication between the management server and the new component;

requesting a new component model from the new component;

communicating a list of supported languages for the new component model from the management server to the new component;

reviewing all available new component models;

generating a list of all new component models expressed in languages supported by the management server;

identifying a preferred new component model from the generated list; and

communicating the identified, preferred new component model from the new component to the management server;

accessing the component models maintained in the autonomic computing system using the management server;

using the management server and accessed component models to develop a system-wide model for autonomic computing system performance covering all of the components in the autonomic computing system;

computing an updated operational state for the autonomic computing system using the system-wide model developed using the accessed component models, the updated operational state associated with improved autonomic computing system behavior in at least one measure of behavior over a current operational state of the autonomic computing system; and

managing the autonomic computing system in accordance with the computed updated operational state.

2. The method of claim 1, wherein the step of maintaining the component models further comprises updating existing component models in the autonomic computing system.

3. The method of claim 1, wherein the step of maintaining the component models further comprises:

identifying existing components that have been removed from the autonomic computing system; and

removing existing component models associated with the removed components from the autonomic computing system.

4. The method of claim 1, wherein:

the step of accessing the component models further comprises;

identifying components within the autonomic computing system for which component models are not maintained in the autonomic computing system; and

obtaining a generic model for each identified component for which a component model is not maintained in the autonomic computing system; and

the step of computing the updated operational state for the autonomic computing system further comprises computing the updated operational state for the autonomic computing system using the accessed component models and the obtained generic models.

5. The method of claim 1, wherein the step of managing the autonomic computing system in accordance with the computed optimum state further comprises adjusting operational states of one or more components, setting control variables for one or more components, adjusting resource allocation within the autonomic computing system or combinations thereof.

6. A method for autonomic system management comprising:

maintaining within an autonomic computing system comprising a plurality of components arranged in a decentralized peer-to-peer architecture where each peer comprises one of the components, a plurality of component models, each component model associated with one of the plurality of components in the autonomic computing system and each component model simulating the operation of its associated component to predict a state and performance of that component for various operational conditions of the entire computing system, wherein each component model is supplied by and maintained by its associated component independent of other components;

using each component to forward its component model to one or more peer components within the autonomic computing system using peer-to-peer file sharing protocols;

identifying components in the plurality of components that obtain a sufficient number of components models through the peer-to-peer file sharing to form an aggregate model of the autonomic computing system that is sufficiently detailed to perform system-wide optimization;

selecting one of the identified components comprising to function as a system-wide management server to compute an updated operational state;

computing an updated operational state for the autonomic computing system at the selected identified component using peer-to-peer file sharing protocol forwarded component models, the updated operational state associated with improved autonomic computing system behavior in at least one measure of behavior over a current operational state of the autonomic computing system; and

managing the autonomic computing system in accordance with the computed updated operational state.

7. The method of claim 6, wherein the step of selecting one of the identified components further comprises using a leader election algorithm to select the component.

8. The method of claim 6, wherein the components comprise information technology components, computer, routers, databases, web servers, autonomic storage units, storage area network controllers, performance monitors or database servers.

9. The method of claim 6, wherein each component model is integrated into its associated component as firmware.

10. A method for autonomic system management comprising:

maintaining within an autonomic computing system comprising a plurality of components arranged in a plurality of local zones where each zone comprises two or more of the components, a plurality of component models, each component model associated with one of the plurality of components in the autonomic computing system and each component model simulating the operation of its associated component to predict a state and performance of that component for various operational conditions of the entire computing system, wherein each component model is supplied by and maintained by its associated component independent of other components;

sharing within each local zone component models among the components disposed in that local zone using an epidemic gossip protocol;

using components within each local zone to elect at least one component within that local zone to communicate shared component models from that local zone to other components of the autonomic computing system outside that local zone;

applying the sharing of local zone component models within each local zone and the communication of the shared component models to components outside each local zone recursively to convey a system-wide aggregate model to a top-level component within the autonomic computing system;

computing an updated operational state for the autonomic computing system at the top-level component using the conveyed system-wide aggregate model, the updated operational state associated with improved autonomic computing system behavior in at least one measure of behavior over a current operational state of the autonomic computing system; and

managing the autonomic computing system in accordance with the computed updated operational state.

11. The method of claim 10, wherein each component model is integrated into its associated component as firmware.

12. The method of claim 10, wherein each component comprises a hardware component.

13. The method of claim 10, wherein each component comprises a computer, a router, a database, a web server, an autonomic storage unit, a storage area network controller or a database server.

14. The method of claim 10, wherein the step of maintaining the component models further comprises:

identifying existing components that have been removed from the autonomic computing system; and

removing existing component models associated with the removed components from the autonomic computing system.