COLLABORATIVE CLOUD-EDGE COMPUTING: A SURVEY ON ARCHITECTURES, ORCHESTRATION, AND DISTRIBUTED AI PARADIGMS

Abstract

The rapid evolution of digital infrastructure has catalyzed the emergence of novel computing paradigms that integrate cloud computing with edge computing within complex collaborative architectures. This paper presents a comprehensive survey of recent literature (2010–2026) encompassing three principal dimensions: (1) cloud-edge collaborative architectures including hierarchical models, network topologies, and workload distribution mechanisms; (2) orchestration frameworks supporting dynamic resource management, intelligent task scheduling, and latency optimization; and (3) distributed artificial intelligence paradigms including Federated Learning, Split Computing, and on-device inference at the edge layer. Through a systematic analysis of 78 Scopus-indexed publications, we identify evolutionary trends, open challenges, and future research directions. Survey findings indicate that the integration of adaptive policy-based orchestration mechanisms with AI inference capabilities at the edge layer constitutes the most determinant factor in improving energy efficiency and reducing end-to-end latency in cloud-edge systems. These findings are expected to provide guidance for researchers and practitioners in designing next-generation distributed computing systems.