A COMPARATIVE EVALUATION OF DATA REDUCTION TECHNIQUES FOR IOT EDGE/FOG COMPUTING

The Internet of Things (IoT) is an expanding field with applications in a wide range of domains. As the number of connected IoT devices increases, the resulting volume of sensor data places greater demands on the network and storage capacity. This study investigates how data reduction techniques applied in the edge/fog layer of the IoT system architecture can mitigate these challenges by reducing the volume of data transmitted while preserving relevant information. This study focuses on three data reduction techniques commonly addressed in this field of research: data aggregation, filtering, and prediction. By evaluating solutions based on the core principles of these techniques, the study aims to compare their effectiveness in achieving an optimal balance between transmission reduction and the accuracy with which the reduced data stream represents actual sensor readings from a temperature monitoring IoT device. An experimental evaluation was conducted by implementing each solution on a Raspberry Pi, with the aim of simulating a real-world edge/fog computing setup. The performance of each solution was then analyzed under varying temperature conditions. The results indicate that the most optimal technique depends on the characteristics of the incoming data. The filtering-based solution provided the best balance between transmission reduction and accuracy under stable and gradually changing conditions, whereas the prediction-based solution was more effective under fluctuating conditions. Although the aggregation-based solution consistently reduced transmissions, it was less accurate under rapidly changing conditions. This study provides practitioners with comparative and practical insights for the development of resource-efficient IoT systems.