Access Network Selection based on Available Bandwidth Estimation for Heterogeneous Networks

Abstract With the new 5G trend, the amount of heterogeneous networks (HetNets) has increased exponentially. Cooperation between the most commonly used HetNets such as: WiFi and Long Term Evolution (LTE) is required to ensure a balanced traffic load which benefits the end user with maximum throughput and high quality of service. A balanced traffic problem with WiFi- LTE heterogeneous wireless networks is selecting which network to use, LTE or WiFi, if both are available. In other words, a node would have to send/receive traffic using either LTE or WiFi. We address this issue and propose both a WiFi available bandwidth estimation technique called ”Available Bandwidth with Error Probability (ABEP)” and a traffic steering strategy to improve the performance of WiFi-LTE wireless networks. Our approach provides a better estimation of available bandwidth in WiFi with respect to relevant bandwidth estimation techniques. The bandwidth utilization in a WiFi Network is obtained at an initial stage of the traffic steering strategy. The steering of traffic is then triggered depending on events and the network conditions. Different from other approaches, our proposed traffic steering algorithm considers both a minimum bandwidth and a minimum SNR threshold to connect to WiFi, which are necessary to have an efficient transmission. One of the benefits of our approach regards reducing the use of the LTE network and its corresponding monetary charges.