Electronic Thesis and Dissertation Repository


Doctor of Philosophy


Electrical and Computer Engineering


Dr. Xianbin Wang


Due to the explosive growth of mobile data traffic and the shortage of spectral resources, 5G networks are envisioned to have a densified heterogeneous network (HetNet) architecture, combining multiple radio access technologies (multi-RATs) into a single holistic network. The co-existing of multi-tier architectures bring new challenges, especially on resource management and security provisioning, due to the lack of common interface and consistent policy across HetNets. In this thesis, we aim to address the technical challenges of data traffic management, coordinated spectrum sharing and security provisioning in 5G HetNets through the introduction of a programmable management platform based on Software-defined networking (SDN).

To address the spectrum shortage problem in cellular networks, cellular data traffic is efficiently offloaded to the Wi-Fi network, and the quality of service of user applications is guaranteed with the proposed delay tolerance based partial data offloading algorithm. A two-layered information collection is also applied to best load balancing decision-making. Numerical results show that the proposed schemes exploit an SDN controller's global view of the HetNets and take optimized resource allocation decisions. To support growing vehicle-generated data traffic in 5G-vehicle ad hoc networks (VANET), SDN-enabled adaptive vehicle clustering algorithm is proposed based on the real-time road traffic condition collected from HetNet infrastructure. Traffic offloading is achieved within each cluster and dynamic beamformed transmission is also applied to improve trunk link communication quality.

To further achieve a coordinated spectrum sharing across HetNets, an SDN enabled orchestrated spectrum sharing scheme that integrates participating HetNets into an amalgamated network through a common configuration interface and real-time information exchange is proposed. In order to effectively protect incumbent users, a real-time 3D interference map is developed to guide the spectrum access based on the SDN global view. MATLAB simulations confirm that average interference at incumbents is reduced as well as the average number of denied access.

Moreover, to tackle the contradiction between more stringent latency requirement of 5G and the potential delay induced by frequent authentications in 5G small cells and HetNets, an SDN-enabled fast authentication scheme is proposed in this thesis to simplify authentication handover, through sharing of user-dependent secure context information (SCI) among related access points. The proposed SCI is a weighted combination of user-specific attributes, which provides unique fingerprint of the specific device without additional hardware and computation cost. Numerical results show that the proposed non-cryptographic authentication scheme achieves comparable security with traditional cryptographic algorithms, while reduces authentication complexity and latency especially when network load is high.