Transforming Large-Scale Virtualized Networks: Advancements in Latency Reduction, Availability Enhancemenit, and Security Fortification

• Main Author: Tamim, Ibrahim
• Format: Dissertation
• Language: English
• Published: ProQuest Dissertations & Theses 01-01-2023
• Subjects: Artificial intelligence; Capital expenditures; Computer science; Information Technology; Integer programming; Internet and social media studies; Internet of Things; Linear programming; Machine learning; Neural networks; Traffic congestion; Transportation; Web studies
• ISBN: 9798381024579

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 (URLLC) 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.