FULL PDF
Abstract
With the surge in cloud-based applications demanding real-time responsiveness, conventional static orchestration models fall short in coping with the heterogeneity and dynamism of modern infrastructures. This paper introduces an adaptive machine learning (ML)-based orchestration framework tailored for heterogeneous cloud environments, enhancing performance optimization through dynamic resource allocation and predictive analysis. Leveraging reinforcement learning and deep learning strategies, the system achieves real-time workload adaptability, reducing latency and optimizing energy efficiency. Experimental insights and literature synthesis reveal the potential of ML to revolutionize cloud orchestration across diverse environments, from edge to fog to centralized cloud systems.
View more >>