Automatic best wireless network selection based on key performance indicators

Introducing cognitive mechanisms at the application layer may lead to the possibility of an automatic selection of the wireless network that can guarantee best perceived experience by the final user. This chapter investigates this approach based on the concept of Quality of Experience (QoE), by introducing the use of application layer parameters, namely Key Performance Indicators (KPIs). KPIs are defined for different traffic types based on experimental data. A model for an ap- plication layer cognitive engine is presented, whose goal is to identify and select, based on KPIs, the best wireless network among available ones. An experimenta- tion for the VoIP case, that foresees the use of the One-way end-to-end delay (OED) and the Mean Opinion Score (MOS) as KPIs is presented. This first implementation of the cognitive engine selects the network that, in that specific instant, offers the best QoE based on real captured data. To our knowledge, this is the first example of a cognitive engine that achieves best QoE in a context of heterogeneous wireless networks.