Ensuring End-to-End Dependability Requirements in Cloud-based Bluetooth Low Energy Applications

Bluetooth Low Energy (BLE) is increasingly used for time-critical IoT applications, where BLE-based smart objects need to exchange data with a remote server within stringent end-to-end latency and reliability bounds. While existing research has investigated how to timely send packets between pairs of BLE devices, it is still unclear how a BLE device can sustain time-critical end-to-end communication
with a remote server, for example, hosted in the cloud. In this paper, we tackle this problem and show how BLE devices can autonomously measure and cope with end-to-end network delays and loss along the path to the remote server. To this end, we first devise an analytical model of the communication between a BLE end-node and the cloud. We then leverage this model to dynamically adapt the communication parameters of the BLE device and sustain the desired end-to-end dependability requirements while minimizing the energy expenditure. Specifically, we design and implement two adaptation strategies on the popular nRF52 platform, and experimentally show that they both allow to sustain a given end-to-end reliability and a given end-to-end latency for data transmissions from/to the BLE node, while limiting the node’s power consumption.