Edge–Cloud Continuum: A Conceptual Perspective on Distributed Cloud Computing

Abstract

The rapid growth of connected devices, artificial intelligence applications, and data-driven services has led to an extraordinary increase in distributed data generation. Traditionally, machine learning models are trained using centralized datasets collected from multiple devices and stored in cloud servers. Although this approach provides strong computational capabilities, it raises major concerns related to privacy, data security, and communication overhead. In many real-world environments such as healthcare systems, smart cities, and industrial IoT networks, transmitting raw data to centralized locations may be impractical or even prohibited due to privacy regulations.
Federated learning has arisen as a promising solution to these challenges by enabling collaborative model training while keeping data on local devices. Instead of sharing raw datasets, participating devices train models locally and send only model updates to a central coordinator for aggregation. This approach significantly reduces the exposure of sensitive information and allows organizations to utilize distributed data resources more efficiently.
The combination of federated learning with edge computing and cloud infrastructures has created new opportunities for distributed intelligence across the edge–cloud continuum. In such environments, computation is performed across multiple layers including devices, edge nodes, and centralized cloud systems. This paper presents a conceptual examination of federated learning within distributed edge–cloud environments. The study reviews existing research, analyzes current challenges, and proposes a conceptual framework for implementing federated learning in distributed computing systems. The findings suggest that federated learning can play a key role in enabling scalable, privacy-aware, and efficient artificial intelligence applications in future computing ecosystems.