A Real Network Performance Analysis Testbed for Encrypted MQTT in DMS
Antonio Francesco Gentile, Emilio Greco, Domenico Luca Carnì · 2024
The widespread adoption of IoT devices has led to the development of Distributed Measurement Systems (DMS). However, cyber attacks aimed at destroying critical infrastructure and retrieving sensitive data are rising. We proposed a security-oriented VLAN testbed deployed with opensource firmware and constrained hardware to address this issue. The approach uses local MQTT brokers, TLS tunnels for local sensor data, and an SSL tunnel to transmit encrypted data to a cloud-based central broker. The proposal evaluates critical metrics such as Total Ratio, Total Runtime, Average Runtime, Message time, Average Bandwidth, and Total Bandwidth to predict the minimum network throughput for the selected QoS and security. From a measurement science perspective, lower productivity may result in phenomena being observed less frequently, potentially leading to misinterpretations. This paper does not introduce a new method for safeguarding measurement data, but rather evaluates the network performance of commonly used methods applicable to captive and commercial hardware. This scenario is typical for IoT-based DMS. The proposal considers a security-focused VLAN approach, along with targeted solutions for hardware constraints. Given the nature of the worst case, this allows for broad application of the results obtained. Our study marks the initial phase of a larger and more extensive research effort. Specifically, we used commercial hardware such as the Raspberry Pi 4. Our goal is to extend the scope of the research and delve into more complex and detailed questions in our next work. The primary objective is to identify algorithms that ensure optimal data transmission and encryption ratios and explore algorithms that ensure maximum compatibility with existing infrastructures supporting MQTT technology and will facilitate secure connections for geographically dispersed DMS IoT networks, particularly in challenging environments.