Scalability of Neural Network Models for the Classification and Detection of Threats in Network Traffic

Ayala Rodriguez, Erick (2023) Scalability of Neural Network Models for the Classification and Detection of Threats in Network Traffic. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (986kB) | Preview
Preview	PDF (Configuration manual) Download (823kB) | Preview

Abstract

This study evaluates the impact of parallel data processing implementing multiple GPUs and multiple core CPUs on neural network models for network traffic analysis. The UNSW-NB15 dataset will be used in experiments to test various machine learning models, including Convolutional Neural Networks (CNN), and Sequential Neural Networks (SNN). For reference, was implemented an additional model like Decision Tree Classifier (DTC) to have a comparison with traditional models. TensorFlow was selected as the main framework to work with due to the compatibility with Nvidia GPUs and the CUDA libraries, also the models were implemented using a local machine and then replicated the scenarios in an EC2 instance in AWS to test the scalability of the models incrementing the processing power with multiple GPUs and increasing the number of cores of the CPU. The results show a decrease in the time to train the models which can be helpful to develop new approaches in cybersecurity to detect and classify attacks while utilizing all the capabilities of the systems resources available and achieve lower training times and more accurate detections of network threats. The research aims to contribute to the development of innovative approaches to improve the accuracy and performance of neural network approaches for network traffic analysis.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Gupta, Punit 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:	Tamara Malone
Date Deposited:	10 Aug 2024 13:10
Last Modified:	10 Aug 2024 13:10
URI:	https://norma.ncirl.ie/id/eprint/7042

Actions (login required)

View Item