A Federated Learning Framework for Testing in Distributed Cloud Environments

Abstract

The rapid expansion of the Internet of Things (IoT) poses crucial challenges for the deployment of IoT applications in IT infrastructures, in particular the latency issues inherent in cloud-based platforms, despite their cost and productivity advantages. This highlights the need to rethink compliance testing frameworks for cloud environments, with a focus on improving coordination and observability during distributed data processing. To address these challenges, this study proposes a federated learning (FL) framework, in which fog nodes independently train local models and share only parameter updates - not raw data - with a central server. These updates are aggregated into a global model, which is then redistributed to IoT users for refinement using local datasets. This iterative process improves the accuracy of the global model while preserving data confidentiality at device level. Focusing on IoT edge computing, where computational tasks are decentralized to optimize resource efficiency, we present a new FL-driven test architecture designed to streamline coordination and fault detection in distributed cloud systems. The methodology is rigorously evaluated to demonstrate its effectiveness in balancing workload distribution and improving system reliability.