ACE-Sync: An Adaptive Cloud-Edge Synchronization Framework for Large-Scale Distributed Training
International Journal of Information and Communication Technology Trends · 05.04.2026 · 6 (1)
DOI: 10.66525/ijictt84 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Information and Communication Technology Trends, online erschienen am 05.04.2026. Verfasst von Noah Romano und 3 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
The proliferation of data generated at the network edge has driven the need for decentralized machine learning approaches, specifically large-scale distributed training across the cloud-edge continuum. However, synchronizing deep neural network models between resource-rich cloud servers and resource-constrained edge devices presents significant challenges, primarily due to variable network bandwidth, high latency, and node heterogeneity. This paper introduces an adaptive cloud-edge synchronization framework designed to optimize the communication efficiency and convergence stability of distributed training. The proposed framework dynamically adjusts the synchronization frequency and applies context-aware gradient compression based on real-time network conditions and local training progress. By decoupling the synchronization cycles from strict global epochs, the architecture mitigates the straggler effect common in heterogeneous environments. Comprehensive evaluations demonstrate that the proposed framework significantly reduces communication overhead while maintaining model accuracy comparable to traditional centralized approaches. The adaptive nature of the synchronization protocols ensures robust performance even under severe network degradation. This research provides a scalable solution for deploying complex deep learning models across distributed edge networks, facilitating the next generation of intelligent, privacy-preserving, and bandwidth-efficient applications.
Zitieren
APA
Romano, N., Higgins, E., Keller, N., & Vance, J. (2026). ACE-Sync: An Adaptive Cloud-Edge Synchronization Framework for Large-Scale Distributed Training. International Journal of Information and Communication Technology Trends, 6(1). https://doi.org/10.66525/ijictt84
Vancouver
Romano N, Higgins E, Keller N, Vance J. ACE-Sync: An Adaptive Cloud-Edge Synchronization Framework for Large-Scale Distributed Training. International Journal of Information and Communication Technology Trends. 2026;6(1). doi:10.66525/ijictt84
BibTeX
@article{Romano2026,
author = {Romano, Noah and Higgins, Eleanor and Keller, Noah and Vance, Julian},
title = {ACE-Sync: An Adaptive Cloud-Edge Synchronization Framework for Large-Scale Distributed Training},
journal = {International Journal of Information and Communication Technology Trends},
year = {2026},
volume = {6},
number = {1},
doi = {10.66525/ijictt84}
}
EUVIMED