How can a functional, self-adaptive Zero Trust framework be designed, implemented and evaluated to enforce security policies in real-time while ensuring transparency based on machine learning predictions within a cloud infrastructure?

Mulay, Maitreyee Vaibhav (2025) How can a functional, self-adaptive Zero Trust framework be designed, implemented and evaluated to enforce security policies in real-time while ensuring transparency based on machine learning predictions within a cloud infrastructure? Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (1MB) | Preview

Abstract

In modern world sophisticated cyber threats are emerging for which traditional perimeter based security systems are proving inadequate to fight such threats. The solution proposed in this research is zero trust security model which follows the rule of ’never trust always verify’, this offers a more robust paradigm. This paper presents the implementation of a self adaptive zero trust framework that utilises machine learning to provide dynamic, real-time policy enforcement in a cloud environment. The dataset CSE-CIC-IDS2018 network traffic is used for model development. An end to end ML operational pipeline is created on AWS (Amazon Web Services). The data pre-processing and segregation into internal and external threat classification is done initially. Machine learning models - Random Forest, SVM (Support Vector Machine) and logistic regression are used in Sagemaker AI notebook. The lambda functions are created to trigger based on simulation which provides scalable and isolated inference endpoints. Cloudwatch is used log the events of the simulation. The results of the model demonstrates high model accuracy. The classifies for external threat models have 89 and 91 percent accuracy while internal threat model achieved is 99 percent. The system successfully indicates the threat detected with recommended action based on which the lambda applies the policy to the user appropriate to the threat level. This research provides a complete blueprint for operating an intelligent Zero Trust system which confirms and feasibility and effectiveness using ML-driven predictions with explainability for adaptive security policy enforcement.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Samarawickrama, Yasantha UNSPECIFIED
Subjects:	T Technology > T Technology (General) > Information Technology > Cloud computing Q Science > QA Mathematics > Computer software > Computer Security T Technology > T Technology (General) > Information Technology > Computer software > Computer Security Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Cloud Computing
Depositing User:	Ciara O'Brien
Date Deposited:	30 Mar 2026 10:50
Last Modified:	30 Mar 2026 10:50
URI:	https://norma.ncirl.ie/id/eprint/9245

Actions (login required)

View Item