IoT Latency and Power consumption : Measuring the performance impact of MQTT and CoAP

Sammanfattning: The purpose of this thesis is to investigate the impact on latency and power consumption of certain usage environments for selected communication protocols that have been designed for resource constrained usage. The research questions in this thesis is based on the findings in Lindén report “A latency comparison of IoT protocols in MES” and seeks to answer the following: ”How does MQTT impact the latency and power consumption on a constrained device?” ”How does CoAP impact the latency and power consumption on a constrained device?” ”How does usage environment influence the latency for MQTT and CoAP?” This thesis only seeks to explore concepts and usage environments related to wireless sensor networks, internet of things and constrained devices. The experiments have been carried out on a ESP WROOM 32 Core board V2 applying MQTT and CoAP as the communication protocols. The overall research method used in this thesis is the experimental research design proposed by Wohlin et al. Experiments have been created to support or disprove hypotheses which are formulated to answer the research questions. The experiments were conducted in test environments, which mimic a real-life wireless sensor network environment. The process is thoroughly recorded to further increase the traceability of this thesis. This decision was made due to a comment Boyle et al. made about the problems with real-life experiments about the wireless communication-based research domain. Where Boyle et al. states that there is “insufficient knowledge” available for the research community. The MQTT related latency experiments showed that QoS level 0 had the lowest latency of all the QoS levels. However, the results also showed that QoS 1 and 2 almost had an identical latency. The CoAP related latency experiments did not indicate any obvious trends. The results from the power consumption related experiments were inconclusive since the data was incomplete. The usage environment related experiments yielded conclusive results. The data showed that there was a small variation in the latency impact across the various usage environments. Furthermore, the data suggest that CoAP and MQTT had lower latencies in a high signal strength environment compared to a lower signal strength environment. However, it is not clear if there were any unknown factors influencing the results.