Scalability study of backhaul capacity sensitive network selection scheme in LTE-WiFi HetNet

Wireless heterogeneous network (HetNet) with small cells presents a new backhauling challenge that differs from those of experienced by conventional macro-cells. In practice, the choice of backhaul technology for these small cells whether fibre, xDSL, point-to-point and point-to-multipoint wireless, or multi-hop/mesh networks, is often governed by availability and cost and not by required capacity. Therefore, the resulting backhaul capacity of the small cells in HetNet is likely to be non-uniform because of the mixture of backhaul technologies adopted. In such an environment, a question then arises whether a network selection strategy that considers the small cells’ backhaul capacity will improve the end users’ usage experience. In this paper, a novel dynamic backhaul capacity sensitive (DyBaCS) network selection scheme is proposed and compared with two commonly used network selection schemes, namely, WiFi First and physical data rate in an LTE-WiFi HetNet environment. The proposed scheme is evaluated in terms of average connection or user throughput^† and fairness among users. The effects of varying WiFi backhaul capacity (uniform and non-uniform distribution), WiFi-LTE coverage ratio, user density and WiFi access points (APs) density within the HetNet form the focus of this paper. Results show that the DyBaCS scheme generally provides superior fairness and user throughput performance across the range of backhaul capacity considered. Besides, DyBaCS is able to scale much better than WiFi First and physical data rate across different user and WiFi densities.