Electronic Thesis and Dissertation Repository

Thesis Format

Integrated Article


Doctor of Philosophy


Electrical and Computer Engineering


Shami, Abdallah


In today’s digital age, the increasing demand for networks, driven by the proliferation of connected devices, data-intensive applications, and transformative technologies, necessitates robust and efficient network infrastructure. This thesis addresses the challenges posed by virtualization in 5G networking and focuses on enhancing next-generation Radio Access Networks (RANs), particularly Open-RAN (O-RAN). The objective is to transform virtualized networks into highly reliable, secure, and latency-aware systems. To achieve this, the thesis proposes novel strategies for virtual function placement, traffic steering, and virtual function security within O-RAN. These solutions utilize optimization techniques such as binary integer programming, mixed integer binary programming, column generation, and machine learning algorithms, including supervised learning and deep reinforcement learning. By implementing these contributions, network service providers can deploy O-RAN with enhanced reliability, speed, and security, specifically tailored for Ultra-Reliable and Low Latency Communications use cases. The optimized RAN virtualization achieved through this research unlocks a new era in network architecture that can confidently support URLLC applications, including Autonomous Vehicles, Industrial Automation and Robotics, Public Safety and Emergency Services, and Smart Grids.

Summary for Lay Audience

In today’s world, as more devices and applications connect to the internet and as technology is injected into our daily lives, robust and faster network infrastructure is needed. This thesis focuses on addressing the challenges faced by 5G systems and their underlying infrastructure, particularly Open Radio Access Networks (O-RANs). The main goal is to make these networks more reliable, secure, and fast. To achieve this, the thesis has developed novel strategies, optimization algorithms, and machine-learning models to enhance the security and performance of O-RAN. Putting those contributions into action allows network service providers to offer more reliable, secure, and fast services. With optimized O-RAN, networks can support different applications such as Autonomous Vehicles, Industrial Automation and Robotics, Public Safety and Emergency Services, and Smart Grids. This thesis paves the way for a new era where our networks are smarter and more secure for these critical applications.