Analyzing software change impact based on machine learning

Patent metadata

Field	Value
Publication number	US-11200043-B2
Application number	US-201816049083-A
Country	US
Kind code	B2
Filing date	Jul 30, 2018
Priority date	Jul 30, 2018
Publication date	Dec 14, 2021
Grant date	Dec 14, 2021

Abstract

Analyzing software change impact is provided. Feedback is received regarding a predicted impact of a software change to a set of components of a service. It is determined whether the predicted impact of the software change is correct based on the feedback. In response to determining that the predicted impact of the software change is correct based on the feedback, the software change is applied to the set of components of the service increasing performance of the service and a server computer hosting the service.

First claim

What is claimed is:

1. A computer-implemented method for analyzing software change impact, the computer-implemented method comprising:
receiving, by a computer, an input to apply a software change on a set of components of a service provided to registered clients by a cloud computing platform;
generating, by the computer using a software change impact evaluation model, a predicted impact of the software change on the set of components of the service;
receiving, by the computer, feedback regarding the predicted impact of the software change to the set of components of the service;
determining, by the computer, whether the predicted impact of the software change is correct based on the feedback;
responsive to the computer determining that the predicted impact of the software change is correct based on the feedback, applying, by the computer, the software change to the set of components of the service increasing performance of the service and a server computer hosting the service;
recording, by the computer, impact data corresponding to the software change to the set of components of the service, wherein the impact data comprises at least one of (i) component dependencies that represent a set of one or more dependencies that a component of the service directly affected by the software change has with other components of the service, (ii) a component importance that represents a level of importance of the component of the service directly affected by the software change, (iii) a component relationship with other cloud services that represents one or more relationships that the component of the service directly affected by the software change has with one or more of the other cloud services, (iv) a component failure rate that represents a rate of failure associated with the component of the service directly affected by the software change, and (v) a component impact on applications in the service that represents an impact of the component of the service directly affected by the software change on the applications in the service;
optimizing, by the computer, using machine learning, the software change impact evaluation model based on the recorded impact data corresponding to the software change on the set of components of the service;
accessing, by the computer, a software change history data structure to identify references to same or similar software changes as the software change to be applied to the set of components of the service;
determining, by the computer, whether a reference to a same or similar software change exists in the software change history data structure;
responsive to the computer determining that a reference to a same or similar software change does not exist in the software change history data structure, retrieving, by the computer, a value and associated weight for each impact factor in a plurality of impact factors corresponding to the software change on the set of components of the service from a software change impact table;
generating, by the computer, the predicted impact corresponding to the software change using values and associated weights for the plurality of impact factors;
generating, by the computer, a software change date, a software change identifier, and a service identifier corresponding to the software change; and
recording, by the computer, the predicted impact, the software change date, the software change identifier, and the service identifier corresponding to the software change in the software change history data structure.

2. The computer-implemented method of claim 1 further comprising:
responsive to the computer determining that the predicted impact of the software change is correct based on the feedback, generating, by the computer, a real impact result report corresponding to the software change that includes impact level, priority level, and other services impacted by the software change applied to the service; and
providing real impact feedback corresponding to the software change to the software change impact evaluation model for refinement and optimization of the software change impact evaluation model.

3. The computer-implemented method of claim 1 further comprising:
responsive to the computer determining that a reference to a same or similar software change does exist in the software change history data structure, selecting, by the computer, a closest matching software change to the software change to be applied to the set of components of the service;
retrieving, by the computer, historical impact data corresponding to the closest matching software change from the software change history data structure; and
generating, by the computer, using an impact evaluation model, the predicted impact with severity, priority, and combined service impact based on the historical impact data corresponding to the closest matching software change.

4. The computer-implemented method of claim 1 , wherein the plurality of impact factors includes component dependencies, component importance, component relationship with other services provided by the cloud computing platform, component failure rate, and component impact on applications in the service.

5. The computer-implemented method of claim 1 further comprising:
generating, by the computer, a related services topological structure corresponding to the service based on retrieved related services data;
identifying, by the computer, each related service that will be affected by the software change to be applied to the set of components of the service; and
generating, by the computer, a degree of importance score for each related service that will be affected by the software change to be applied to the set of components of the service.

6. The computer-implemented method of claim 5 further comprising:
retrieving, by the computer, a point value and associated weight for each install recommendation factor in a plurality of install recommendation factors corresponding to the software change from a software change recommendation table; and
generating, by the computer, a software change recommendation value for the software change using point values and associated weights for the plurality of recommendation factors and the degree of importance score for each related service that will be affected by the software change to be applied to the set of components of the service.

7. The computer-implemented method of claim 6 , wherein the plurality of install recommendation factors include compliance, emergency, importance, relevance, and impact level.

8. The computer-implemented method of claim 2 further comprising:
generating, by the computer responsive to the computer determining that the predicted impact of the software change is correct based on the feedback, a set of impact dashboards corresponding to the software change on the set of components of the service for system administrator review and feedback, wherein the service is a cloud service provided by the cloud computing platform.

9. The computer-implemented method of claim 1 further comprising:
comparing, by the computer, the predicted impact result with a real impact result of the software change to determine whether the predicted impact result is accurate.

10. The computer-implemented method of claim 5 further comprising:
retrieving, by the computer, change data corresponding to the software change from a change management system; and
retrieving, by the computer, related services data corresponding to the service from a service configuration management database, wherein the related services data retrieved from the service configuration management database identify relationships between different cloud services of the cloud computing platform.

Assignees

• IBM

Inventors

• Shi Hui Qing
• Wang Yi Bin
• Yuan Yuan
• Zhou Ya Pei
• Wang Wei

Classifications

• G06N20/00

  Machine learning · CPC title

• G06F8/65 Primary

  Updates (security arrangements therefor G06F21/57) · CPC title

• G06N5/04

  Inference or reasoning models · CPC title

• G06F16/219

  Managing data history or versioning (querying versioned data G06F16/2474; querying temporal data G06F16/2477) · CPC title