Topology-Aware GNN-DQN for Optimal Task Offloading in Edge–Fog–Cloud Systems

Evaluating Dynamic and Energy-Efficient Task Offloading Mechanisms in Heterogeneous Fog Computing Systems

The rapid growth of Internet-of-Things (IoT) devices has increased the need for computing support close to end users, particularly for applications that cannot tolerate long processing delays or excessive energy consumption. Fog computing has emerged as a practical extension of the cloud to address these requirements, yet real deployments often involve a mix of devices with different processing abilities, communication characteristics, and power constraints. These differences make it difficult to decide when and where tasks should be offloaded. This study introduces a task-offloading approach that adapts to changing conditions in a heterogeneous fog environment. The method continuously observes factors such as processor utilization, task size, communication delay, and the remaining energy of participating devices. Using this information, the system determines whether a task should run on the originating device, a nearby fog node, or the cloud. The approach aims to limit unnecessary transfers while striking a balance between energy use and execution delay. Simulation experiments conducted in iFogSim indicate that the proposed strategy consistently improves performance over conventional static or energy-unaware schemes. The results show notable reductions in overall energy usage and significant improvements in task-completion success under varying network loads. These findings suggest that integrating real-time monitoring with adaptive decision-making can strengthen the efficiency and responsiveness of fog-based IoT systems.

Lightweight IoT Node Offloading Framework for Real-Time Edge Analytics

The rapid growth of Internet of Things (IoT) deployments has intensified the need for efficient, decentralized computation management at the network edge. This paper presents a lightweight, neighbor-aware one-hop task offloading framework designed for resource-constrained IoT networks. The proposed adaptive scheme combines Exponential Weighted Moving Average (EWMA) load estimation with a queue-depth gate to prevent unnecessary offloading under transient load spikes, and an assignment-pressure mechanism to distribute tasks more evenly across neighboring nodes. We evaluate the framework using a custom-developed discrete-event simulator on a 90-node ringplus-chord topology with heterogeneous hotspot and light nodes, comparing against three baselines: local-only execution, random offloading, and least-loaded neighbor selection. Results show that a load-aware but pressure-unaware least-loaded strategy surprisingly produces the highest load variance (377.25), worse than random offloading (114.84), due to severe task funneling toward persistently fast nodes. The proposed scheme eliminates task drops entirely, achieves an average latency of 148.7 ms, and reduces task-count variance to 44.33 - an $8.5 \times$ improvement over the least-loaded baseline and 4.6× over local-only execution - while requiring only 28.21% of tasks to be offloaded. These results demonstrate that assignment-pressure tracking is essential for fair load distribution in energy-limited IoT deployments.

Internet of Things-Centric Optimized Service Provisioning in Multi-Cloud Environment

A lightweight, QoS-aware service placement algorithm that evaluates latency, bandwidth, and node load in real time is introduced that yields reduced latency and more consistent wait times relative to heuristic and genetic baselines.

Multi-application operator placement in cloud-edge infrastructure for big data stream processing

Task Scheduling in the Fog to Cloud Continuum for IoT Services: A Taxonomy and Structured Synthesis of Distributed Resource Management

The paper argues for a shift from proof of concept scheduling studies toward reproducible, transparent, and deployable fog systems, and identifies several priorities for future work: standardized benchmark workloads, cloud native scheduling that accounts for container lifecycle and microservice dependencies, resilience aware scheduling that treats failures and migration as first class concerns, and carbon aware orchestration that extends beyond energy minimization.

Task Offloading with Temporal Awareness in Heterogeneous Cloud–Edge–Ground Systems

In post-disaster or infrastructure-limited remote-area environments where local terrestrial networking infrastructure is unavailable or severely degraded, heterogeneous cloud–edge–ground (CEG) systems provide a viable platform for computation offloading. Ground devices such as IoT sensors and unmanned ground vehicles (UGVs) generate computation-intensive and latency-sensitive tasks, while fixed-wing unmanned aerial vehicles (UAVs) with edge servers offer intermediate processing and relay capabilities to the cloud. We aim to minimize system energy consumption and task latency by jointly considering offloading ratios, CPU frequencies and transmission power of the ground devices. However, efficiently partitioning tasks across these tiers is challenging due to dynamic wireless network conditions, limited onboard resources, and diverse task demands. We propose a temporal dynamics-aware multi-agent offloading and resource allocation framework for such CEG systems in this work. We introduce KAT-MAPPO, a Kolmogorov–Arnold Transformer-enhanced Multi-Agent Proximal Policy Optimization scheme. By embedding Kolmogorov–Arnold Network layers into the transformer encoder, the framework captures long-range temporal dependencies and improves feature representation, while MAPPO enables decentralized offloading under dynamic conditions. Resource allocation is further decoupled and solved analytically via convex optimization. Using a real-world setup, our simulation results show that KAT-MAPPO significantly outperforms existing baselines, achieving faster convergence, a 43% reduction in average task latency in the training evaluation setting, and up to 53.57% latency reduction for the case study scenarios, while also reducing energy consumption. These results highlight the promise of reinforcement learning considering temporal dynamics for efficient offloading in heterogeneous CEG systems.