5G and Beyond: Smart Devices as Part of the Network Fabric

by B.aldomero Coll-Perales, Javier Gozalvez, Juan Luis Maestre

Institute of Electrical and Electronics Engineers (IEEE)

5G networks mainly rely on infrastructure-centric cellular solutions to address data traffic and service demands. Continuously scaling infrastructure- centric cellular networks is not exempt from challenges, and beyond 5G networks should consider the active coexistence and coordination of infrastructure-centric and device-centric wireless networks. Device-centric wireless networks will build from D2D and MCNs. Device-centric wireless networks can push the limits of edge computing and networking to smart devices (including smartphones, vehicles, machines and robots), and exploit their mobile computing, storage and connectivity capabilities. These capabilities can be more efficiently utilized using demand-driven opportunistic networking that establishes the connections between devices and nodes not just based on their presence, but also on their capacity to support the requested demand and services. This article presents results from experimental field tests that demonstrate the cellular spectral efficiency gains that can be achieved from the combined use of device-centric wireless communications and demand-driven opportunistic networking. The field trials demonstrate that these technologies can improve the cellular spectral efficiency of conventional cellular communications by up to a factor of 4.7 and 12 in outdoor pedestrian and vehicular scenarios, respectively, under the evaluated scenarios and conditions. These gains have been obtained using 4G and IEEE 802.11 technologies. However, the potential of device-centric wireless networks is not constrained to any particular radio interface. The results presented in this article provide empirical evidences that further motivate progressing toward a new paradigm where edge networking capabilities are moved to smart mobile devices that become part of the network fabric, and can opportunistically and locally integrate network management functions to ensure that sufficient resources are placed where the demand arises.