Comparative Analysis and dataset generation for evaluating IoT Protocols: MQTT-SN , CoAP

Abstract

The rapid growth of the Internet of Things (IoT) has intensified the need for efficient communication protocols in constrained environments. This study evaluates two key IoT protocols, MQTT and CoAP, through systematic simulations in COOJA/Contiki OS, analyzing their energy efficiency, latency, throughput, and reliability under varied network conditions. We generate a machine learning-ready dataset capturing protocol performance across different payload sizes, transmission intervals, and network scales. This dataset can be used to enables predictive analysis of protocol behavior and supports data-driven IoT system design. Results show MQTT achieves better energy efficiency and scalability for large networks, while CoAP excels in low-latency scenarios. Beyond performance comparison, this work provides: (1) a methodological framework for protocol evaluation, and (2) a structured dataset for future IoT research, bridging protocol analysis with machine learning applications. The combined results and dataset offer practical insights for protocol selection in smart infrastructure and industrial IoT deployments, where constrained resources demand optimized communication solutions.