In the first fiscal of this project, we have focused on the development of the context-aware IoT system. We have made achievements as follows.
1) Context-ware IoT mobile router: We have designed and created IoT mobile router, which consists of Raspberry board, Sensor HAT board and LTE modem board. We attached the context-info such as sensor name and sensor device id to the trailer of data packet. The packet gateway, built with our deeply programmable FLARE node, will get the context-info of the sensor data and will perform traffic engineering accordingly. The prototyped system has been demonstrated in Japan IT week at Tokyo Big Sight during 2019/5/8-5/10.
2) Context-aware Private LTE. We have prototyped a context-aware UE slicing architecture utilizing in-network deep learning. In our implementation, we tag the downlink packets with the identified context info utilizing deep learning at the packet gateway (PGW) and transmit them to RAN slice with different spectrum resource blocks (RBs) and scheduling algorithms. The context info can be defined be as application name, location info or others as user defined. The work has been published in 3 papers and one paper got IEICE ICM English Session Encouragement Award.
3) High-speed PPPoE Router for home IoT network. We build a compact 10Gbps softwarePPPoE router on a commercial mini-PC for home IoT network, which can achieve much higher throughput than a commercial PPPoE router tested in a production environment. The related work has been presented at a conference, and the paper has been recommended for IEICE journal paper.