QoE-Based Flow Admission Control in Small Cell Networks

An important requirement on 5G mobile systems is to accommodate massive numbers of wireless devices and users. Heterogeneous networks are expected to play a crucial role in meeting this requirement. In this vein, small cells are expected to become an integral part of these heterogeneous networks. However, their success would not last longer unless they offer services at a quality similar to that currently ensured by the macro cellular networks. Mitigating congestion of the backhaul links to small cell networks is a crucial factor. With this regard, this paper proposes an admission control that makes decisions to redirect IP flows, fully or partially, to the macro or small cell networks, or to reject the incoming flows. The decision mechanism is based on predictions of users' Quality of Experience (QoE). It is modeled as a Markov decision process (MDP), whereby the aim is to derive the optimal policy (i.e. reject or accept flows in the macro or the small cell) that maximizes users' QoE. Through computer simulations, we evaluate the performance of the proposed admission control and compare it against a random policy decision. We also numerically illustrate its optimal policies in different scenarios under different traffic load conditions.