Though highly energy-intensive components of modern industrial infrastructure, cloud data centers have become essential as digital services have grown exponentially. Environmental sustainability requires the change from conventional, device-level power management to autonomous, system-level orchestration as worldwide digital needs match those of whole towns. This seminar offers an exhaustive analysis of Artificial Intelligence (AI) an‎d Machine Learning (ML) techniques aimed at attaining green cloud networking. The study shows how predictive modeling an‎d smart workload consolidation can greatly span the energy proportionality gap by examining modern architectures including Deep Reinforcement Learning (DRL), Transformers, an‎d Generative Adversarial Networks (GANs). Among important technical solutions highlighted are Software-Defined Networking (SDN) for energy-efficient traffic engineering an‎d the use of Digital Twins to construct virtual san‎dboxes for risk-free policy instruction. The seminar also investigates the move toward carbon-conscious computing, which entails temporal an‎d spatial migration of workloads to match actual renewable energy supply. Despite the promise of these AI-driven frameworks, critical challenges such migration overhead, hardware heterogeneity, an‎d the scalability of AI models in the 6G era still exist. This research combines the present literature to provide a strategic path for next-generation, self-optimizing infrastructures that match high-performance computer throughput with worldwide net-zero emission goals.

2/18/2026 12:00:00 AM

89791

1404/11/29

A Review of AI-Based Workload Consolidation to Reduce Energy Usage in Data Centers

Green Cloud Computing, , AI-Driven Resource Orchestration, , Carbon-Aware Computing , Deep Reinforcement Learning, , Energy-Efficient Traffic Engineering.