AI-Driven Dynamic Resource Allocation in Cloud Computing: Predictive Models and Real-Time Optimization

Abstract

The advancement in cloud computing has brought about the need for resource management and allocating computing, storage, and network resources dynamically to suit the ever-evolving workloads. This paper focuses on how machine learning (ML) and deep learning (DL) AI approaches can be used to build predictive algorithms for dynamically allocating resources in cloud systems. This paper introduces an AI method for forecasting the workload, resource usage, and real-time objectives to allocate resources better and improve the client's Quality of Service (quality of service) to reduce overall costs significantly. Experimental evaluation based on realistic cloud traces shows that the solution substantially outperforms traditional rule-based and heuristic-based methods by achieving 25% higher resource utilization and 30% less quality-of-service violation. Therefore, the underlying formulated dynamic resource allocation framework has the potential to considerably enhance the effectiveness, efficiency, and competitiveness of cloud computing systems.