Method of predicting availability of a system

• US 7,756,803 B2
• Filed: 03/03/2005
• Issued: 07/13/2010
• Est. Priority Date: 03/03/2005
• Status: Active Grant

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Narayan, Ranjani, Natanasabapathy, Gautham, Varadarajan, Keshavan
• Primary Examiner(s): HOLMES, MICHAEL B
• Application Number:
  US11/885,475
• Publication Number:
  US 20080168314A1
• Time in Patent Office: 1,958 Days
• Field of Search: 706/45
• US Class Current: 706/45
• CPC Class Codes: G06F 11/008 Reliability or availability analysis
• CPC Class Codes: G06F 11/076 by exceeding a count or rate limit, e.g. word- or bit count limit

Abstract

The computer availability is predicted by determining a hazard fail rate for the computer system that is based on the hazard fail rate of the individual computer components and the computer system structure s6, as well as the mean repair rate for the system. The computer hazard rates of individual computer system components are based on the error monitoring of those computer components during the computer system operation in real time.

Claims

1. A method of predicting availability of a system, the system comprising a plurality of components, the method comprising:

determining a hazard rate for the system;

determining a mean repair rate for the system; and

calculating the system availability from the system hazard rate and the system mean repair rate.

2. A method according to claim 1, wherein the step of determining a hazard rate for the system comprises:

determining a system structure, the structure comprising interconnections between the components;

determining a hazard rate for each of the components; and

determining the system hazard rate from the component hazard rates.

3. A method according to claim 2, wherein the step of determining the system hazard rate from the component hazard rates comprises:

combining the component hazard rates for each of the components based on the determined structure.

4. A method according to claim 3, wherein the step of determining a reliability for each of the components from the component hazard rate comprises using the equation:

R=e<sup>−

λ

t</sup>

where

R is the component reliability;

λ

is the component hazard rate; and

t is the time after the first error for which reliability is being assessed.

5. A method according to claim 2, wherein the step of determining a hazard rate for each of the components includes monitoring each of the components during operation of the system to determine an error history.

6. A method according to claim 5, wherein the error history comprises the number of errors occurring over a predetermined time period.

7. A method according to claim 5, further comprising:

comparing the error history with an error threshold; and

determining the hazard rate as the probability of crossing the error threshold given the error history.

8. A method according to claim 1, wherein the step of determining a mean repair rate for the system comprises:

determining a mean repair rate for each of the components;

calculating individual availabilities of each of the components based on the mean repair rate;

calculating a steady state system availability based on the individual availabilities; and

calculating the mean repair rate for the system from the steady state system availability.

9. A method according to claim 8, wherein the step of calculating a steady state system availability comprises calculating a weighted function, in which the weights for each individual component represent an importance of the component relative to other components.

10. A method according to claim 8, comprising calculating an individual availability for a component in accordance with the equation:

$A=\frac{\mathrm{\mu <br><br>}}{\mathrm{\lambda <br><br>}+\mathrm{\mu <br><br>}}$

where

A is the availability of an individual component;

μ

is the mean repair rate of the component; and

λ

is the hazard rate of the component.

11. A method according to claim 8, comprising calculating the mean repair rate of the system in accordance with the equation:

$\stackrel{\_}{\mathrm{\mu <br><br>}}=\frac{\stackrel{\_}{A}<br><br>\stackrel{\_}{\mathrm{\lambda <br><br>}}}{1-\stackrel{\_}{A}}$

where

Ā

is the steady state availability of the system;

∥

is the mean repair rate for the system; and

λ

is the hazard rate for the system.

12. A method according to claim 1, comprising calculating the system availability in accordance with the equation:

$\stackrel{\_}{A}<br><br>\left(t\right)=\frac{\stackrel{\_}{\mathrm{\mu <br><br>}}}{\stackrel{\_}{\mathrm{\lambda <br><br>}}+\stackrel{\_}{\mathrm{\mu <br><br>}}}+\frac{\stackrel{\_}{\mathrm{\lambda <br><br>}}}{\stackrel{\_}{\mathrm{\lambda <br><br>}}+\stackrel{\_}{\mathrm{\mu <br><br>}}}<br><br>e^{-\left(\stackrel{\_}{\mathrm{\lambda <br><br>}}+\stackrel{\_}{\mathrm{\mu <br><br>}}\right)<br><br>t}$

where

Ā

(t) is the instantaneous availability of the system;

λ

is the system hazard rate; and

μ

is the mean repair rate of the system.

13. A method of maintaining a system at a predetermined service level, comprising:

predicting system availability according to claim 1; and

acting to maintain the predetermined service level in response to the predicted availability.

14. A method according to claim 13, wherein the step of acting to maintain the predetermined service level comprises replacing a component while the system is operational, the component being a component that is predicted to fail.

15. A method according to claim 13, wherein the step of acting to maintain the predetermined service level comprises enabling migration of a process to other hardware, when predicted availability indicates a degradation in process performance.

16. Apparatus for maintaining a system at a predetermined service level, comprising:

means for monitoring the system for errors;

means for predicting availability of the system based on the errors; and

means for acting in response to the predicted availability to maintain the system at the predetermined service level.