A rollback is the process of reverting a system's state to a previously known, stable version. In the context of network-as-code, where infrastructure is defined and versioned like software, a bug introduced in a new release can be quickly remediated by deploying the last known good version of the code. This action directly and immediately restores functionality to the production site, which is the primary goal after a failed deployment. The other options are preventative measures that occur before a release, not restorative actions taken after a failure has occurred in production.

A. Staging: This is a pre-production environment used for testing before a release. It is a preventative safeguard, not a method to restore a live production site.

B. Code review: This is a quality assurance process to detect flaws before code is deployed. It is a preventative measure, not a recovery mechanism for a live system.

C. Code testing: This is the process of executing code to find bugs before a release. It is a preventative control, not a method for restoring a failed production environment.

1. Official Vendor Documentation: AWS CloudFormation User Guide

"Stack updates

" section on "Monitoring the progress of a stack update." The documentation states

"By default

if the stack update fails

CloudFormation rolls back all changes to the stack's previous working state." This exemplifies rollback as a standard recovery mechanism in an Infrastructure as Code (IaC) service.

2. Peer-reviewed Academic Publication: Rahman

A.

& Williams

L. (2016). "On the relationships between build failures

continuous integration

and fault-proneness: An empirical study of open source projects." 2016 IEEE International Conference on Software Quality

Reliability and Security (QRS). While focused on CI

the paper's context on build management highlights that a key tenet of CI/CD is the ability to quickly revert changes

i.e.

rollback

if a build introduces faults (p. 160

Section III-B). DOI: 10.1109/QRS.2016.29

3. University Courseware: Ousterhout

J. (2018). Lecture notes for CS 140: Operating Systems. Stanford University. Lecture 18

"Crash Recovery

" discusses the fundamental principle of using a stable

previous state (like a checkpoint or log) to recover from failures

which is the core concept of a rollback in deployment scenarios (pp. 5-7).