Context-aware, user-driven, network-controlled RAT selection for 5G networks

摘要

It is expected that in the very near future, cellular networks will have to deal with a massive data traffic increase, as well as a vast number of devices. Optimal placement of the end devices to the most suitable access network is expected to provide the best Quality of Service (QoS) experience to the users but also the maximum utilization of the scarce wireless resources by the operators. Several on-going proposals attempt to overcome the existing barriers by enabling the use of Wi-Fis and femto-cells to cater for part of the load generated by the end devices. The evolution of the Access Network Discovery and Selection Function (ANDSF) for the core part of the cellular network, as well as the Hotspot 2.0 approach, are currently being subject to thorough discussions and studies and are expected to facilitate a seamless 3GPP-WiFi interworking. During the past years, several Radio Access Technology (RAT) selection schemes have been proposed. However, these schemes do not take into consideration the opportunities offered by these new standardized approaches. Our paper acts in a manifold way: Firstly, it proposes COmpAsS, a Context-Aware RAT Selection mechanism, the main part of which operates on the User Equipment (UE)-side, minimizing signaling overhead over the air interface and computation load on the base stations. Secondly, we discuss in detail the architectural perspective; i.e., the extensions needed in the network interfaces that enable the exchange of the required context information among the respective network entities and in accordance with the 3GPP trends in relation to the context-aggregating entities. Furthermore, we quantify the signaling overhead of the proposed mechanism by linking it to the current 3GPP specifications and performing a comprehensive per-parameter analysis. Finally, we evaluate the novel scheme via extensive simulations in a complex and realistic 5G use case, illustrating the clear advantages of our approach in terms of key QoS metrics, i.e. the user-experienced throughput and delay, both in the uplink and the downlink. (C) 2016 Elsevier B.V. All rights reserved.