A software-defined networking (SDN) architecture is capable of integrating all radio frequency and optical wireless small cell networks (e.g. fifth generation (5G), long-term evolution (LTE) femtocell, wireless fidelity (WiFi), light fidelity (LiFi)) in one network domain. This paper considers a SDN-enabled heterogeneous network (HetNet) comprised of LiFi, LTE femtocell and WiFi access points (APs). The HetNet control plane maintains the state of the network topology and wireless resources, which can support the development of intelligent service provisioning and efficient data communications in x generation (xG) wireless networks. The SDN applications use the network state to provide services in the data plane. However, when the state of network and wireless resources constantly changes, the SDN applications cannot provide reliable and guaranteed services to the wireless user equipments. This paper develops a queuing theoretic framework, which provides a performance evaluation for the SDN-enabled HetNet and applications convergence. A traffic engineering (TE) scheme is developed to support dynamic agnostic downlink flows routing to APs and differentiated granular services across the HetNet. Network and user centric policies are developed to make applications aware of network resource availability on the northbound and southbound interfaces of a SDN controller. Numerical models are introduced to study the impact of the computation and communication resources of northbound and southbound interfaces on the SDN-enabled HetNet scalability and the quality-of-service (QoS) guarantee of applications. Also, simulation scenarios are conducted to evaluate the performance of the TE scheme in provisioning effective and reliable services for subscribers.
|Number of pages||20|
|Publication status||Published - 6 Apr 2020|
- SDN contoller
- traffic engineering
- heterogeneous wireless networks
- LTE software agents