Edge-Computing-Enabled Engineering Software for Smart Infrastructure: Balancing On-Device Analytics and Cloud Collaboration

Abstract

The fast development of smart infrastructures requires a software engineering approach capable to balance low-latency, edge-based processing with the scalability and collaborative features of cloud ecosystems. The paper introduces the design and development of an edge-computing-supported engineering software framework that combines ondevice analytics, real-time sensor data crunching, multisite cloud synching to enable smart infrastructure lifestyle management. By providing an execution platform at the edge to solve local decision-making (e.g. predictive maintenance, structural health monitoring and energy optimization), it decreases reliance on network and increases responsiveness. At the same time, a cloud scale layer assists in design collaboration, shared models and broad system simulation across distributed engineering groups. The study is using a hybrid architecture based on microservices running inside containers, federated learning for adaptive model retraining and digital twins integration to provide tight feedback loops between physical assets and their digital surrogates. Results show that the proposed edge–cloud collaboration achieves a substantial enhancement in reliability of the system, 60% bandwidth saving and faster real-time analytics in latency-critical missions. The work presented can be seen as a stepping-stone towards discussing cyber-physical integration in the civil and mechanical engineering communities, paving the way for a scalable architectural model of future smart infrastructure software ecosystems.