Implementation of Zero Trust Architecture in Online Examination Systems to Prevent Cheating

Abstract

The increasing use of online exam systems (e-exams) in the post-pandemic cyberlearning ecosystem presents critical challenges to academic integrity, particularly related to vulnerabilities to increasingly sophisticated technical fraud such as the use of virtual machines, remote desktops, and even artificial intelligence-based identity manipulation. Traditional perimeter-based security models—verifying only the initial login—have proven inadequate in detecting dynamic threats that occur during exam sessions. This research proposes the implementation of Zero Trust Architecture (ZTA) with the core principle of "Never Trust, Always Verify" as a comprehensive solution for securing online exams. The methodology used is Research and Development (R&D), which includes designing a security architecture based on the five main pillars of Zero Trust: identity, device, network, application, and data. The developed system implements a Continuous Authentication mechanism through periodic facial biometric validation and a Device Health Check feature to ensure the integrity of the devices used by participants. Through a micro-segmentation approach, user access is strictly limited to authorized exam resources. The research results are expected to demonstrate that implementing a Zero Trust model can improve the effectiveness of real-time fraud detection and block unauthorized access without significantly compromising system performance. This research provides an important contribution to educational institutions in building a digital examination infrastructure that is independent, transparent, and able to minimize dependence on manual proctor supervision.