SDN and MEC can accelerate the beam training in mmWave outdoor communications

Applying mmWave technology to mobile communications has recently received increasing attention. One of the major technical challenges is to handle the frequent handover caused by mobility and blockage. Aiming to address this problem, we abstract the network layer to the physical layer via a cross-layer method and perform handovers in a centralized way based on software defined networking. In addition, in order to maintain enough user throughputs and reduce latency during handover procedure, we also propose a fast semi-offline beam training scheme, which takes advantage of mobile edge computing to pre-store the power delay angular profile data of the mmWave channel. The proposed scheme can ensure seamless connections when users are moving into a non-line-of-sight area, thereby providing users with better quality of service for promising applications in the fifth-generation networks, like augmented reality. We use data of field measurement to initially analyze the network throughput and latency for outdoor mmWave networks, which shows that our proposed system is reliable, scalable, and implementable.