On the evaluation of application level delays in public LoRaWAN networks

The Internet of Things paradigm aims to connect a very large number of smart objects located all around the world for increasing the knowledge about the environment. Low-power wireless communications are compulsory to reduce the smart object invasiveness and ensure reasonable duration of autonomous node. For this reason, several Low-Power Wide-Area Network (LPWAN) technologies have been proposed in the recent past, because of their large range and low consumption. Most notably, the LoRaWAN solution emerged as the most accepted LPWAN and several research works appeared, discussing its advantages and limitations. However, very few, if any, discuss timing performance in a real scenario, considering not only the wireless frontend but also the wired backend. In this work, the authors implement a real use case based on a public LoRaWAN infrastructure and propose metrics to formally evaluate the delays at the application level (i.e. considering the whole infrastructure as a black box). The result is an average delay in the range 400-700 ms, and the proof that the delay is directly dependent on both the LoRaWAN backend and the connection speed of information endpoints.