A Latency Analysis of Edge vs. Cloud Computing for Internet-of-Mirrors (IoM) Applications

The growing demand for real-time, privacy-preserving health and beauty applications, such as the Internet-of-Mirrors (IoM), necessitates the careful selection of computational architectures. This paper presents a comparative investigation using dental smile analysis as a representative use case to explore the trade-offs between edge and cloud computing in this context. We introduce a framework for evaluating latency and performance across different processing stages in both edge and cloud setups. By utilising a Jetson Xavier edge device for local processing and the Google Cloud platform for cloudbased analysis, we provide quantitative insights into the tradeoffs between local and remote processing in IoM systems. Our study reveals that while edge computing offers significantly lower latency and enhanced privacy, cloud computing provides superior scalability and the ability to handle more complex tasks. These findings highlight the crucial need to balance between real-time performance and computational capabilities in IoM applications. Considering the trade-offs of both approaches, our results revealed that a hybrid edge-cloud architecture, customised for specific use cases, could offer an optimal balance of computational performance and latency requirements. This work provides insights into the design of future IoM systems, with implications for the standardisation and development of hybrid architectures in the health and beauty sector.