<p>Internet of Things (IoT) is a network of physical devices embedded with sensors, software, and network connectivity, allowing them to acquire data, share it across a network, and derive insights or decisions from the data. Real-time IoT applications require minimal time between data acquisition and information extraction, which is achieved using an edge computing architecture. In addition, most sensor nodes are battery-operated devices requiring power optimization. This paper presents CoMIEC a dynamically configurable MQTT-based framework that achieves a trade-off between sensor node power consumption of various program cycles and latency, the key configuration parameter being the rate at which message packets are communicated by the sensor nodes and consequently processed at the gateway or the cloud, as well as the size of the message packet. The proposed framework uses MQTT topics for dynamic reconfiguration, evaluated on a physical IoT testbed for machine-condition and structural-health monitoring applications. The results for various configurations showed a energy variation of 11.01% and a latency variation of 46.07% in the cloud scenario and 35.35% in the edge computing scenario. Hence, the proposed CoMIEC framework achieves a latency–power trade-off by varying the message packet publishing rate according to battery power level and latency, compared with preset thresholds.</p>