AI-Driven Cloud Computing: Predictive vs. Conventional Resource Management

Cloud Computing allows one to scale resources up or down on demand, yet the practice-oriented resource management methods currently used are not always capable of responding promptly to the changing needs. The emerging concept of Artificial Intelligence (AI) may offer a solution due to its unique ability to predict future demands. The present paper explores the concept of using AI in Cloud Computing and how it can address the current issues related to resource management, workload, and response time, scalability, security, and cost. A conceptual framework was created and evaluated within the experiment to identify whether an AI-based approach could be better than the status quo. A comparison was made between the traditional threshold-based provisioning and a newly designed predictive model through a simulated workload. On average, the experimental results showed a 32% decrease in response time, an increase in utilization from 64% to 81%, a reduction in scaling reaction delay by over 56%, and a reduction in operational cost by approximately 17%. Altogether, the experiment proved that predictive AI-based resource management can offer significant advantages over conventional methods in Cloud Computing, although there are factors that need to be considered for practical applications.

• Nihar Karra
• Rakesh Kumar Agrawal
• Wasim Mohammed Amin Tambe

Euro Vantage Journal of Artificial Intelligence (EVJAI), Vol. 2, No. 1

Published

Karra, N., Agrawal, R. K., & Tambe, W. M. A. (2025). AI-Driven Cloud Computing: Predictive vs. Conventional Resource Management. Euro Vantage Journal of Artificial Intelligence, 2(1), 76-94. https://doi.org/10.65923/gb1m1788

@article{karra2025aidrivencloud, title={AI-Driven Cloud Computing: Predictive vs. Conventional Resource Management}, author={Karra, Nihar and Agrawal, Rakesh Kumar and Tambe, Wasim Mohammed Amin}, journal={Euro Vantage Journal of Artificial Intelligence}, volume={2}, number={1}, pages={76--94}, year={2025}, doi={10.65923/gb1m1788}}