Search:
Computing and Library Services - delivering an inspiring information environment

Bearicade: A Novel High-Performance Computing User and Security Management System Augmented with Machine Learning Technology

Al-Jody, Taha (2021) Bearicade: A Novel High-Performance Computing User and Security Management System Augmented with Machine Learning Technology. Doctoral thesis, University of Huddersfield.

[img]
Preview
PDF - Accepted Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (16MB) | Preview

Abstract

Despite the rising development and popularity of HPC systems, there have been insufficient advancements towards the security of HPC systems. The substantial computational power, high bandwidth networks, and massive storage capacity provided in the HPC environment are desirable targets for the attackers. The majority of educational institution HPC centres provide their users with simple access methods lacking the modern security needs. Thus, accelerating the systems’ proneness to modern cyber-attacks. The current implementations of HPC access points, such as web portals, offer users direct access to the HPC systems. Consequently, such web portal implementations affect the HPC system with the same security challenges faced by cloud providers and web applications. Although attempts have been made toward securing HPC systems, most of these implementations are outdated, insufficient with the current security standards, or do not integrate well with modern HPC access solutions. To address these security issues, Bearicade, a novel High-Performance Computing (HPC) user and security management system, was designed, developed, implemented and evaluated. Bearicade is a data-driven secure unified framework for managing HPC users and systems security. This framework is an add-on layer to an existing HPC systems software, collecting over 50 different types of information from multiple sources within the HPC systems. It offers Artificial Intelligent security solutions with an added usability and accessibility without adversely affecting the performance and functionality of HPC systems. Throughout this study, the security and usability of Bearicade were validated implementing multiple Machine Learning models. It has been deployed over three years as a production system for students and researchers at the University of Huddersfield QueensGate Grid (QGG) with considerable success, protecting the QGG systems from the summer 2020 attacks that has affected many other HPC systems in research and educational establishments.

Item Type: Thesis (Doctoral)
Subjects: T Technology > T Technology (General)
Schools: School of Computing and Engineering
School of Computing and Engineering > High-Performance Intelligent Computing > High Performance Computing Research Group
Depositing User: Rebecca Hill
Date Deposited: 13 Oct 2021 12:59
Last Modified: 13 Oct 2021 13:00
URI: http://eprints.hud.ac.uk/id/eprint/35579

Downloads

Downloads per month over past year

Repository Staff Only: item control page

View Item View Item

University of Huddersfield, Queensgate, Huddersfield, HD1 3DH Copyright and Disclaimer All rights reserved ©