Design and Development of a Smart Home Device Simulator

Testing IoT devices poses significant challenges, especially due to the need for real hardware to accurately simulate device interactions in real-world environments like smart homes. Unlike traditional software, which can often be tested in virtual settings, IoT systems rely on physical setups to account for critical factors such as network behavior, interoperability, and real-time responses. This dependency on actual devices increases costs and limits scalability, making it difficult to conduct large-scale, comprehensive tests across complex IoT networks. This thesis addresses this problem by designing and developing a simulator specifically for Commercial-Off-The-Shelf (COTS) smart devices, focusing on the second generation of Shelly devices. The simulator replicates core functionalities and interactions between these devices, as well as their integration within both the Smartotum ecosystem and Home Assistant platform. The primary goal is to create a flexible, cost-effective environment where smart home solutions can be tested thoroughly without needing physical devices. Using Object-Oriented Programming (OOP) in Python, the simulator models various Shelly Gen. 2 device types and uses MQTT and REST and RPC communication protocols to closely mimic the behaviors and interactions found in real-world smart home setups. The simulator is designed to work with Smartotum, a decentralized smart home ecosystem that prioritizes privacy, interoperability, and reliability by removing reliance on cloud-based services. Instead, Smartotum uses distributed hash tables to manage device data, which helps to mitigate the vendor lock-in issues typical in many smart home systems, allowing different brands and devices to function together seamlessly. The modular design of the simulator supports tests involving both individual devices and complex, multi-device scenarios, where multiple simulated devices operate simultaneously, mimicking a cohesive and realistic smart home environment. Performance tests were performed using the simulator to evaluate the ability of the Smartotum ecosystem to manage high-traffic scenarios and automated device actions. Stress tests were performed to assess how CPU and RAM usage scaled under increased loads of simulated device interactions, providing information on its performance capabilities. By gradually adding simulated devices and increasing data update frequency, the simulator revealed Smartotum’s performance under peak loads, identifying areas where resource optimization may be required. These tests highlight the effectiveness of the Smartotum ecosystem in supporting high-demand smart home interactions and underscore the simulator’s utility as a versatile testing platform. The results showcase the system’s potential for scalable, cross-platform smart home development, offering a valuable tool for testing complex automation, evaluating system performance, and facilitating future enhancements in smart home technology.