
Sustainable cloud scheduling through hybrid neuromorphic-quantum digital twins
Abstract
Efficient and sustainable scheduling techniques with the potential to balance energy efficiency, carbon emissions, performance, and scalability are necessary for the swift development of cloud computing. In an effort to establish sustainable cloud resource management, we introduce SpikeSynth, a novel hybrid neuromorphic-quantum digital twin framework. Spiking neural networks (SNNs), quantum optimization (QAOA on QUBO formulations), and self-healing digital twins are used in this system. An energy-efficient neuromorphic layer that uses Leaky Integrate-and-Fire (LIF) SNNs for real-time workload pattern recognition, a quantum computing layer that solves NP-hard scheduling problems, and a carbon-aware digital twin layer that predicts failures with 92% accuracy with adaptive solar-powered edge orchestration are its main contributions. These layers are integrated into a unified framework by the proposed architecture. In nine quality of service criteria, SpikeSynth has been proven to be superior to state-of-the-art baselines (LSTM-DT, Quantum-Only, CNN-LSTM, QLSTM, and CNN-SLSTM). This was demonstrated through exhaustive trace-driven assessments conducted on Google Cluster Traces v3 (2019) and Azure Public Dataset (2025). These evaluations scale from 100,000 to 2 million workloads across 1,000 to 10,000 VMs. SpikeSynth achieves the lowest mean energy consumption (0.110 ± 0.068 normalized kWh/M ops), operational cost (0.107 ± 0.085), latency (62.97 ± 48.4 s), and carbon footprint (1.81 ± 1.05 gCO₂/op), while delivering the highest solar power efficiency (53.41 ± 0.95%), throughput (1142 ± 366 tasks/s), accuracy (92.50 ± 0.28%), and scalability. Ablation tests use 38.2% less energy, have a 55.0% lower carbon footprint, have 56.7% less latency, and accelerate recovery by 80.6% compared to LSTM. By economically addressing the Energy-Accuracy-Latency Trilemma, fluctuating green energy sources, and hybrid systems, SpikeSynth enables 6G-ready low-carbon, high-performance cloud infrastructure. Smart utilities, the industrial internet of things, and other critical sustainability sectors are integral components of this framework.
© 2026 N. R. Sabat, R. R. Sahoo, S. Radhakrishnan, published by Faculty of Science, University of Peradeniya, Sri Lanka
This work is licensed under the Creative Commons Attribution 4.0 License.