AI-Augmented Graph Neural Networks for Dynamic Resource Allocation in Cloud-Based Big Data Systems

Abstract

The exponential growth of cloud-based big data systems necessitates innovative approaches to optimize resource allocation dynamically. This study introduces an AI-augmented framework leveraging Graph Neural Networks (GNNs) for efficient and adaptive resource management in cloud environments. By modeling resource allocation as a dynamic graph problem, the proposed framework captures intricate relationships between resources, workloads, and system nodes. GNNs are utilized to predict resource demands and recommend optimal allocation strategies in real time, significantly improving system efficiency and performance. Extensive experiments were conducted using synthetic and real-world datasets from cloud-based systems. The results demonstrate that the proposed framework outperforms traditional methods in resource utilization, workload balancing, and response time reduction. Key contributions include the development of a dynamic graph representation for cloud resources, integration of AI techniques for predictive analytics, and a scalable architecture suitable for heterogeneous cloud environments. The study concludes that the AI-augmented GNN framework offers a robust and scalable solution to the challenges of resource allocation in dynamic cloud systems, paving the way for future advancements in intelligent cloud management.