Self-aware and self-healing computing system

• US 9,971,652 B2
• Filed: 05/08/2015
• Issued: 05/15/2018
• Est. Priority Date: 08/30/2005
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Gangadhar, Deepak K.
• Primary Examiner(s): Truong, Loan L. T.
• Application Number:
  US14/707,133
• Publication Number:
  US 20150242281A1
• Time in Patent Office: 1,103 Days
• CPC Class Codes: G06F 11/008 Reliability or availability analysis
• CPC Class Codes: G06F 11/1423 by reconfiguration of paths
• CPC Class Codes: G06F 11/3447 Performance evaluation by modeling
• CPC Class Codes: G06F 11/3452 Performance evaluation by statistical analysis
• CPC Class Codes: G06F 11/3495 for systems
• CPC Class Codes: G06F 2201/875 Monitoring of systems including the internet

Abstract

A method and a computing system for performing the method. At least two microstates of at least two components of a computing system are organized into at least two macrostates of the computing system. Each microstate represents a state that a component of the computing system is able to individually enter. Each macrostate represents a state that the computing system is able to enter as a whole. The macrostates are organized into attractors. Each attractor is a stable state in which the computing system is stable. An attractor separation map is constructed. The attractor separation map indicates how the attractors are separated from one another by at least two hamming distances. Each hamming distance is a number of bits that differ between two attractors.

Claims

1. A method, comprising:

organizing a plurality of microstates of a plurality of components of a computing system into a plurality of macrostates of the computing system, wherein each microstate represents a state that a component of the computing system is able to individually enter, and wherein each macrostate represents a state that the computing system is able to enter as a whole,

wherein the macrostates are organized into attractors, each attractor being a stable state in which the computing system is stable, and wherein the method further comprises constructing an attractor separation map indicating how the attractors are separated from one another by a plurality of hamming distances, each hamming distance being a number of bits that differ between two attractors,

wherein the macrostates of the computing system are included within a plurality of meta-dynamic states of the computing system subject to each meta-dynamic state being a set of macrostates of the macrostates of the computing system, and

wherein the method further comprises;

detecting perturbations of the computing system, wherein a perturbation of the computing system will result in movement thereof to a new meta-dynamic state;

determining that the new meta-dynamic state is undesirable;

determining a path to cause the computing system to move back to a desirable meta-dynamic state; and

causing the computing system to move on the path to the desirable meta-dynamic state.

2. The method of claim 1, wherein the macrostates are organized into both the attractors and basins of attraction, and wherein each basin of attraction is a transient state in which the computing system is unstable, but which leads to one of the attractors.

3. The method of claim 1, wherein said determining the path further comprises:

construing a self-heal graph from the microstate/meta-dynamic state graph, such that the new meta-dynamic state that is undesirable is at a center of the self-heal graph, and the desirable meta-dynamic state is at a periphery of the self-heal graph, wherein the self-heal graph comprises a plurality of paths from the new meta-dynamic state that is undesirable to the desirable meta-dynamic state, each path having an associated cost that is incurred for the computing system to move along the path from the new meta-dynamic state that is undesirable to the desirable meta-meta-dynamic state.

4. The method of claim 1, wherein said determining the path comprises:

constructing a microstate/meta-dynamic state graph that depicts which of the meta-dynamic states the computing system enters when one or more of the microstates are changed according to one or more parameters, wherein construction of the microstate/meta-dynamic state graph is completed when the desirable meta-dynamic state has been reached from the new meta-dynamic state that is undesirable.

5. The method of claim 4, wherein the one or more parameters comprise:

a maximum number of components that the computing system is able to force a microstate change on at the same time;

a cost associated with each microstate change that the computing system is able to force;

a time limit for how long movement of the computing system to the desirable meta-dynamic state from the new meta-dynamic state that is undesirable is able to maximally take; and

a desired meta-dynamic state that is the desirable meta-dynamic state to which the computing system is to move.

6. The method of claim 1, wherein the meta-dynamic states have a dynamic system invariance property, such that the computing system is capable of being in different macrostates when in a meta-dynamic state, but remains dynamically invariant regardless of which macrostate the computing system is in.

7. The method of claim 1, said method further comprising:

organizing the meta-dynamic states into a plurality of desirable meta-dynamic states, a plurality of undesirable meta-dynamic states, and a plurality of transient meta-dynamic states, wherein each desirable and undesirable meta-dynamic state is a stable meta-dynamic state of the computing system, and wherein each transient meta-dynamic state is an unstable meta-dynamic state of the computing system, such that the computing system automatically proceeds to one of the desirable and undesirable meta-dynamic states from the transient meta-dynamic state over time.

8. A computer-readable storage medium comprising a computer program stored thereon, said computer program configured be executed by a computing system to perform a method, said method comprising:

organizing a plurality of microstates of a plurality of components of a computing system into a plurality of macrostates of the computing system, wherein each microstate represents a state that a component of the computing system is able to individually enter, and wherein each macrostate represents a state that the computing system is able to enter as a whole,

wherein the macrostates are organized into attractors, each attractor being a stable state in which the computing system is stable, and wherein the method further comprises constructing an attractor separation map indicating how the attractors are separated from one another by a plurality of hamming distances, each hamming distance being a number of bits that differ between two attractors,

wherein the macrostates of the computing system are included within a plurality of meta-dynamic states of the computing system subject to each meta-dynamic state being a set of macrostates of the macrostates of the computing system, and

wherein the method further comprises;

detecting perturbations of the computing system, wherein a perturbation of the computing system will result in movement thereof to a new meta-dynamic state;

determining that the new meta-dynamic state is undesirable;

determining a path to cause the computing system to move back to a desirable meta-dynamic state; and

causing the computing system to move on the path to the desirable meta-dynamic state.

9. The storage medium of claim 8, wherein the macrostates are organized into both the attractors and basins of attraction, and wherein each basin of attraction is a transient state in which the computing system is unstable, but which leads to one of the attractors.

10. The storage medium of claim 8, wherein said determining the path further comprises:

construing a self-heal graph from the microstate/meta-dynamic state graph, such that the new meta-dynamic state that is undesirable is at a center of the self-heal graph, and the desirable meta-dynamic state is at a periphery of the self-heal graph, wherein the self-heal graph comprises a plurality of paths from the new meta-dynamic state that is undesirable to the desirable meta-dynamic state, each path having an associated cost that is incurred for the computing system to move along the path from the new meta-dynamic state that is undesirable to the desirable meta-meta-dynamic state.

11. The storage medium of claim 8, wherein said determining the path comprises:

constructing a microstate/meta-dynamic state graph that depicts which of the meta-dynamic states the computing system enters when one or more of the microstates are changed according to one or more parameters, wherein construction of the microstate/meta-dynamic state graph is completed when the desirable meta-dynamic state has been reached from the new meta-dynamic state that is undesirable.

12. A computing system comprising a computer-readable storage medium, said storage medium comprising a computer program stored thereon, said computer program configured be executed by the computing system to perform a method, said method comprising:

organizing a plurality of microstates of a plurality of components of a computing system into a plurality of macrostates of the computing system, wherein each microstate represents a state that a component of the computing system is able to individually enter, and wherein each macrostate represents a state that the computing system is able to enter as a whole,

wherein the macrostates are organized into attractors, each attractor being a stable state in which the computing system is stable, and wherein the method further comprises constructing an attractor separation map indicating how the attractors are separated from one another by a plurality of hamming distances, each hamming distance being a number of bits that differ between two attractors,

wherein the macrostates of the computing system are included within a plurality of meta-dynamic states of the computing system subject to each meta-dynamic state being a set of macrostates of the macrostates of the computing system, and

wherein the method further comprises;

detecting perturbations of the computing system, wherein a perturbation of the computing system will result in movement thereof to a new meta-dynamic state;

determining that the new meta-dynamic state is undesirable;

determining a path to cause the computing system to move back to a desirable meta-dynamic state; and

causing the computing system to move on the path to the desirable meta-dynamic state.

13. The computing system of claim 12, wherein the macrostates are organized into both the attractors and basins of attraction, and wherein each basin of attraction is a transient state in which the computing system is unstable, but which leads to one of the attractors.

14. The computing system of claim 12, wherein said determining the path further comprises:

construing a self-heal graph from the microstate/meta-dynamic state graph, such that the new meta-dynamic state that is undesirable is at a center of the self-heal graph, and the desirable meta-dynamic state is at a periphery of the self-heal graph, wherein the self-heal graph comprises a plurality of paths from the new meta-dynamic state that is undesirable to the desirable meta-dynamic state, each path having an associated cost that is incurred for the computing system to move along the path from the new meta-dynamic state that is undesirable to the desirable meta-meta-dynamic state.