Adaptive Resource Orchestration for Fog-Edge Computing in Software-Defined Networks Using Machine Learning

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

This article is not in any list yet, why not save it to one of your lists.
Log in to save this article

Abstract

The convergence of fog and edge computing with software-defined networking (SDN) enables adaptive resource management for latency-sensitive applications in distributed environments. This paper introduces an adaptive machine learning-driven resource orchestration framework that dynamically optimizes task offloading and resource allocation across fog and edge nodes using distributed optimization techniques. By integrating Lyapunov optimization with machine learning approaches, including reinforcement learning and predictive modeling, and harnessing SDN's centralized control capabilities, our adaptive method significantly reduces latency and energy consumption while maintaining quality of service (QoS). The framework employs a mixed integer non-linear programming (MINLP) model, enhanced with heuristic-based relaxation for improved scalability, to manage resources adaptively in fog-enhanced edge networks. Extensive simulations on fog-edge-cloud topologies demonstrate that our adaptive approach achieves superior performance with lower latency and higher resource efficiency compared to traditional centralized methods, validated through comprehensive performance evaluations. The results highlight the critical role of machine learning in enabling adaptive responses to dynamic workloads and the importance of fog computing in extending computational capabilities closer to end-users in evolving network environments.

Article activity feed