Estimating server-change risk by corroborating historic failure rates, predictive analytics, and user projections

Patent metadata

Field	Value
Publication number	US-10084645-B2
Application number	US-201514954134-A
Country	US
Kind code	B2
Filing date	Nov 30, 2015
Priority date	Nov 30, 2015
Publication date	Sep 25, 2018
Grant date	Sep 25, 2018

Abstract

A method and associated systems for predicting a degree of risk associated with a planned change to a computer server or other electronic component. A computerized change-management system receives Probability and Impact inputs derived from user-derived input, from which it determines a Baseline risk of change failure. The system processes mined data to determine an historic change-failure rate as a function of a type of change, and computes a predictive incident probability based on a predictive analytics engine's forecast of whether a particular type of server will be problematic. The system then computes a final Change Risk by adjusting the Baseline risk as a function of the historic change-failure rate, the predictive incident-probability, and a Baseline-specific weighting factor. If the resulting Change Risk is judged to be elevated, the system initiates collateral actions and notifications intended to reduce the probability and impact of a change failure.

First claim

What is claimed is:

1. A computerized change-management system of a network-management platform comprising a processor, a memory coupled to the processor, and a computer-readable hardware storage device coupled to the processor, the storage device containing program code configured to be run by the processor via the memory to implement a method for estimating server-change risk by corroborating historic failure rates, predictive analytics, and user projections, the method comprising:
the system receiving a set of probability input values that each represent a probability that a planned server change will fail, where the planned server change comprises a reconfiguration of a virtual server of a virtual network;
the system further receiving a set of impact inputs that describe an impact of a failure of the planned server change;
the system computing a baseline risk of change failure as functions of the set of probability inputs and of the set of impact inputs;
the system identifying a historic change-failure rate as a function of mined data that characterizes past attempted server changes;
the system calculating a predictive incident probability as a function of a predictive analytics engine's prediction of a likelihood that the virtual server is of a type that is known to be problematic during server changes; and
the system deriving a change-risk value by adjusting the baseline risk as a function of the historic change-failure rate and of the predictive incident probability; and
the system initiating collateral activities as a function of the change-risk value,
wherein the collateral activities are initiated in order to reduce an impact of a failure of the planned server change, and
wherein the collateral activities consist of a combination of two or more activities selected from the group consisting of:
updating a previously installed operating system of the virtual server to a latest version,
blocking a mission-critical application from accessing the virtual server during the planned server change,
allocating additional virtual RAM to the virtual server,
allocating additional virtual disk storage to the virtual server, and
initiating an enhanced error-collection or system-logging procedure during a window of time surrounding a scheduled time of the planned server change.

2. The system of claim 1 , wherein the collateral activities are initiated in order to reduce a probability of a failure of the planned server change.

3. The system of claim 1 , wherein the collateral activities are initiated in order to notify an interested party about a risk-related characteristic of the planned server change.

4. The system of claim 1 , wherein the set of probability inputs and the set of impact inputs are derived from users' answers to survey questions related to the planned server change.

5. The system of claim 1 , wherein the identifying a historic change-failure rate further comprises:
the system organizing the mined data into a set of change classes that are each associated with a type of server change; and
the system associating each change class of the set of change classes, as a function of the mined data, with a probability of change failure.

6. The system of claim 1 , wherein the predictive analytics engine's prediction of the likelihood is a function of:
the predictive analytics engine's organization of information related to previously documented server changes into categories based on characteristics of each server that was previously changed, and
the predictive analytics engine's characterization of each organized category as being either problematic or non-problematic.

7. The system of claim 6 , further comprising:
the system determining that the server to undergo the planned server change does not fit into any of the organized categories;
the system synthesizing a new category, based on the predictive analytics engine's organization of information related to previously documented server changes, wherein the new category is associated with characteristics of the server to undergo the planned server change; and
the system determining how likely the planned server change is to be problematic as a function of the new category and as a further function of the predictive analytics engine's organization of previously documented server changes into categories.

8. The system of claim 1 , wherein the deriving a change-risk value further comprises: the system determining a baseline-specific weighting factor; and
the system further adjusting the baseline risk as a function of the baseline-specific weighting factor.

9. A method for estimating server-change risk by corroborating historic failure rates, predictive analytics, and user projections, the method comprising:
receiving, by a change-management system of a network-management platform, a set of probability input values that each represent a probability that a planned server change will fail, where the planned server change comprises a reconfiguration of a virtual server of a virtual network;
further receiving, by the change-management system, a set of impact inputs that describe an impact of a failure of the planned server change;
computing, by the change-management system, a baseline risk of change failure as functions of the set of probability inputs and of the set of impact inputs;
identifying, by the change-management system, a historic change-failure rate as a function of mined data that characterizes past attempted server changes;
calculating, by the change-management system, a predictive incident probability as a function of a predictive analytics engine's prediction of a likelihood that the virtual server is of a type that is known to be problematic during server changes
deriving, by the change-management system, a change-risk value by adjusting the baseline risk as a function of the historic change-failure rate and of the predictive incident probability; and
initiating collateral activities, by the change-management system, as a function of the change-risk value,
wherein the collateral activities are initiated in order to reduce an impact of a failure of the planned server change, and
wherein the collateral activities consist of a combination of two or more activities selected from the group consisting of:
updating a previously installed operating system of the virtual server to a latest version,
blocking a mission-critical application from accessing the virtual server during the planned server change,
allocating additional virtual RAM to the virtual server,
allocating additional virtual disk storage to the virtual server, and
initiating an enhanced error-collection or system-logging procedure during a window of time surrounding a scheduled time of the planned server change.

10. The method of claim 9 ,
wherein the collateral activities are further initiated in order to reduce a probability of a failure of the planned server change and to notify an interested party about a risk-related characteristic of the planned server change.

11. The method of claim 9 , wherein the set of probability inputs and the set of impact inputs are derived from users' answers to survey questions related to the planned server change.

12. The method of claim 9 , wherein the identifying a historic change-failure rate further comprises:
organizing, by the change-management system, the mined data into a set of change classes that are each associated with a type of server change; and
associating, by the change-management system, each change class of the set of change classes, as a function of the mined data, with a probability of change failure.

13. The method of claim 9 , whe

Assignees

• IBM

Inventors

• Bouz Firas
• Jasionowski Pawel

Classifications

• H04L41/069

  using logs of notifications; Post-processing of notifications · CPC title

• H04L41/147

  for predicting network behaviour · CPC title

• H04L41/142 Primary

  using statistical or mathematical methods · CPC title

• H04L41/149

  for prediction of maintenance · CPC title

• H04L41/082 Primary

  the condition being updates or upgrades of network functionality · CPC title