Reliability on future software-defined communication systems

• Autor(a) principal: Rehman, Asad Ur
• Data de Publicação: 2021
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10773/33216
• topic: Software-defined networking; Network functions virtualization; Cloud computing; Software-defined and virtualized networks; Fault-tolerance; Reliable 5G networks; Hybrid networks; Service assurance; Auto-scaling

Software-defined and virtualized networks are attracting significant attention from both the academic research community and the industry, due to their envisaged constructive impacts on the development of future telecommunication systems (fifth-generation (5G) networks and beyond). Communication service providers will also need to engage in digital transformation to adapt to the next generation of telecommunication services based on the novel software-driven technologies associated with 5G networks. Recently, several studies have explored various aspects of these technologies, such as software-defined networking (SDN), network function virtualization (NFV), cloud computing, and mobile/multi-access edge computing (MEC). However, major standardization developments, regarding both the architecture itself and support for networking environments, are still ongoing. Hence, the development of network applications for software-defined and virtualized networks has not yet accelerated. The advent of such networks presents new challenges and opens new paths for the development of novel strategies, architectures, and standards to enable high reliability, fault tolerance, and service assurance in 5G networks and beyond. Along this line, this thesis explores SDN, NFV, and cloud computing technologies. The objectives are to better understand software-driven and virtualized networks, address challenges in their architectural implementation, explore fault-tolerance and service assurance support, delineate obstacles to commercial deployment at a conceptual level and identify future research directions for software-defined and virtualized network developments. In this regard, this work proposes an out-of-band, self-healing, centralized, hybrid, software-defined, networking control design to reliably conceptualize the management and automation of future dynamic networks. Furthermore, the thesis evaluates the performance of virtualization technologies in reliable 5G scenarios and presents a recovery mechanism that can achieve near-zero service downtime. In addition, it describes the implementation of service assurance components and subsequent auto-scaling testing, using open-source management and orchestration, to dynamically extend or reduce the computing resources allocated to network applications in a softwaredefined and virtualized networking environment, at run time, or as required. Overall, the thesis focuses on proposing, evaluating, testing, and validating solutions in terms of reliability, fault-tolerance, and service assurance using SDN, NFV, and cloud computing-based implementations to determine the feasibility and reliability of software-defined and virtualized networks.