Assessing health of a subsystem or service within a networked system

• US 7,426,502 B2
• Filed: 06/14/2001
• Issued: 09/16/2008
• Est. Priority Date: 06/14/2001
• Status: Active Grant

• Current Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Singhal, Sharad, Karp, Alan H., Bronstein, Alexandre, Das, Joydip
• Primary Examiner(s): Decady; Albert
• Assistant Examiner(s): Chang; Sunray
• Application Number:
  US09/882,581
• Publication Number:
  US 20030018494A1
• Time in Patent Office: 2,651 Days
• Field of Search: 706/47, 706/16, 706/52, 707/104.1, 701/29, 705/2, 709/203
• US Class Current: 706/47
• CPC Class Codes: G06Q 10/10 Office automation; Time management
• CPC Class Codes: Y10S 707/99945 Object-oriented database structure processing
• CPC Class Codes: Y10S 707/99948 Application of database or data structure, e.g. distributed, multimedia, or image

Abstract

A health assessor for assessing health of a target element within a multi-element system includes multiple sensors, each being operatively coupled to the target element to produce measures of the target element. The health assessor also includes measure collectors, each of which collects a measure from one of the sensor. In addition, the health assessor includes evaluators. Each evaluator evaluates at least a subset of all the measures collected by the measure collectors in accordance with (1) a predefined evaluation definition for the respective evaluator and (2) at least a subset of all historical measures to provide an assessment. A probabilistic reasoning network is coupled to the evaluators to receive the assessment from each of the evaluators and to combine all the assessments in accordance with a pre-configured reasoning definition so as to provide an overall health assessment of the target element. A health assessment system including the health assessor is also described.

Claims

1. A health assessor for assessing health of a target element within a multi-element system, comprising:

a plurality of sensors, each being operatively coupled to the target element to produce a measure of the target element;

a plurality of measure collectors, each collecting a measure from one of the sensors, wherein each of the collectors also stores the measure as historical measure;

a plurality of evaluators, each evaluating at least a subset of all the measures collected by the measure collectors in accordance with a pre-configured evaluation definition for the respective evaluator to provide an assessment;

a probabilistic reasoning network coupled to the evaluators to receive the assessment from each of the evaluators and to analyze the assessments in accordance with a pre-configured reasoning definition so as to provide an overall probabilistic health assessment of the target element,

the probabilistic reasoning network to analyze the assessments from the evaluators according to information indicating reliabilities of the respective evaluators.

2. The health assessor of claim 1, wherein the evaluation definition of an evaluator determines which of the measures collected by all of the measure collectors are to be received by the respective evaluator.

3. The health assessor of claim 1, wherein each of the evaluators further comprises

an evaluation definition store that stores the pre-configured evaluation definition of the evaluator;

an evaluation module coupled to the evaluation definition store to provide the assessment by statistically comparing the subset of the measures and the historical measures based on the pre-configured evaluation definition.

4. The health assessor of claim 3, wherein the operation of an evaluator can be changed by replacing the pre-configured evaluation definition for that evaluator with a new evaluation definition.

5. The health assessor of claim 1, wherein the reasoning network is a Bayesian network probabilistic reasoning network.

6. The health assessor of claim 1, wherein the overall probabilistic health assessment of the target element is a probability that the target element is healthy.

7. The health assessor of claim 1, wherein the operation of the probabilistic reasoning network can be changed by replacing the pre-configured reasoning definition with a new reasoning definition.

8. The health assessor of claim 1, wherein the evaluation definition of a first one of the evaluators specifies that the first evaluator is to receive measures from a first group of the measure collectors, and the evaluation definition of a second one of the evaluators specifies that the second evaluator is to receive measures from a second, distinct group of the measure collectors.

9. A health assessment system for assessing health of a target element within a multi-element system, comprising:

a health assessment engine that receives measures of the target element, the health assessment engine having a probabilistic reasoning network to provide health assessment of the target element based on the received measures and historical measures of the target element, wherein the historical measures have already been stored in the health assessment engine;

a result formatting module that formats the health assessment into a report;

a web interface that transmits the formatted report to a remote access system via the Internet.

10. The health assessment system of claim 9, wherein the health assessment engine further comprises

a plurality of sensors, each being operatively coupled to the target element to produce a measure of the target element;

a plurality of measure collectors, each collecting a measure from one of the sensors, wherein each of the collectors also stores the measure as historical measure;

a plurality of evaluators, each evaluating at least a subset of all the measures collected by the measure collectors in accordance with a (1) pre-configured evaluation definition for the respective evaluator and (2) at least a subset of all historical measures of the target element that have already been stored in the collector to provide an assessment;

wherein the probabilistic reasoning network is coupled to the evaluators to receive the assessment from each of the evaluators and to analyze all the assessments in accordance with a pre-configured reasoning definition so as to provide an overall health assessment of the target element.

11. The health assessment system of claim 10, wherein each of the evaluators further comprises

an evaluation definition store that stores the pre-configured evaluation definition of the evaluator;

an evaluation module coupled to the evaluation definition store to provide the assessment by statistically comparing the subset of the measures and the historical measures based on the pre-configured evaluation definition.

12. The health assessment system of claim 11, wherein the operation of an evaluator can be changed by replacing the pre-configured evaluation definition for that evaluator with a new evaluation definition.

13. The health assessment system of claim 9, wherein the probabilistic reasoning network is a Bayesian network probabilistic reasoning network.

14. The health assessment system of claim 13, wherein the probabilistic reasoning network further comprises a reasoning definition store that stores a pre-configured reasoning definition, wherein the pre-configured reasoning definition defines how the probabilistic reasoning network should analyze all the assessments received from the evaluators to produce the overall health assessment of the target element.

15. The health assessment system of claim 14, wherein the operation of the probabilistic reasoning network can be changed by replacing the pre-configured reasoning definition with a new reasoning definition.

16. The health assessment system of claim 9, wherein the remote access system is the target element.

17. The health assessment system of claim 9, wherein the remote access system is a remote access terminal or an application.

18. The health assessment system of claim 9, wherein the health assessment engine further comprises sensors to provide the measures, measure evaluators to provide assessments based on the measures from the sensors and the historical measures,

wherein the probabilistic reasoning network generates the health assessment based on the assessments provided by the measure evaluators and based on information indicating trustworthiness of respective measure evaluators.

19. The health assessment system of claim 9, wherein the result formatting module formats the health assessment into an XML report.