HyperTrust-Fog: Hypergraph-Based Trust-Aware-Federated Orchestration with Energy Adaptive Scheduling for Hierarchical Cloud Fog Edge Systems
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.Abstract
The HyperTrust-Fog system is proposed as an orchestration approach for hierarchical cloud, fog, and edge infrastructures that support latency-sensitive and energy-constrained IoT services. It begins from the observation that many existing federated learning (FL) or graph-based orchestration methods rely on pairwise interaction models and largely static trust assumptions. Such systems are inadequate for fog environments where collaboration is naturally many-to-many, device reliability drifts over time, and adversarial clients can distort collective optimization. To address these issues, HyperTrust-Fog introduces a dynamic hypergraph representation in which hyperedges encode higher-order cooperative groups across fog and edge nodes. A hypergraph neural encoder extracts time-varying embeddings that capture group-based context beyond simple links. These embeddings feed into a Trust-aware Federated Aggregation (TAFA) mechanism that adaptively reweights local updates, attenuating contributions from unreliable or Byzantine participants. In parallel, an energy-adaptive scheduler is also integrated to leverage the exact learned representations. It helps to coordinate task placement so that latency and power consumption are jointly reduced rather than treated as separate objectives. Empirical evaluation on EdgeCloudSim and IoT-EdgeBench, augmented with a synthetic trust drift and attack process, considers deployments of 1,000 nodes with 200 fog and 800 edge devices, and further tests scalability up to 1,200 nodes. Relative to FedAvg, trust-based federated baselines, and graph neural orchestrators, HyperTrust-Fog attains 97.6% orchestration accuracy with a trust classification AUC of 0.981. These reliability gains correspond to system-level improvements, including a 32.4% average latency reduction and a 28.7% increase in energy efficiency, while maintaining accuracy above 95% at larger scales. Overall, HyperTrust-Fog provides a scalable, robust, trust-driven, and energy-efficient orchestration solution for cloud-fog edge systems.