Carbon and Energy-Aware Cloud Resource Management and Task Scheduling Framework

The exponential growth of cloud computing has intensified energy consumption and carbon emissions, posing a critical challenge to environmental sustainability.To address this, the present study proposes a novel CarbonLiteFormer-GSBO framework for energy-and carbon-aware cloud resource management and task scheduling.The proposed CarbonLiteFormer, a lightweight adaptive Transformer architecture, integrates an Energy-Carbon Dual Attention (ECDA) mechanism that dynamically learns correlations between resource utilization, carbon footprint, and workload complexity.Unlike conventional deep or static schedulers, CarbonLiteFormer leverages hybrid spatial-temporal embeddings to predict task demands while maintaining minimal model complexity.The optimization phase employs the Genetic Secretary Bird Optimization (GSBO) algorithm, a synergistic metaheuristic that combines the global search capability of Genetic Algorithms with the adaptive exploration of the Secretary Bird Optimizer to achieve cost-effective and energyefficient scheduling decisions.The framework was evaluated using two benchmark datasets, the CloudSim plus Task Scheduling Dataset and the PlanetLab Energy-Aware Workload Trace Dataset, under multiple QoS parameters such as makespan, throughput, energy efficiency, and carbon emission index.Experimental results demonstrate that CarbonLiteFormer-GSBO outperforms existing state-of-the-art models, achieving 23.8% lower energy consumption, 18.6% carbon reduction, and 16.4% improvement in scheduling efficiency.The proposed approach provides a sustainable, scalable, and adaptive solution for next-generation green cloud computing environments.