Self-Optimizing cloud substrate networks: An AI-driven approach to dynamic infrastructure optimization
Year: 2025
Type: article
Abstract: Self-Optimizing Cloud Substrate Networks represent a paradigm shift in cloud infrastructure management, combining graph theory foundations with artificial intelligence to create dynamic, adaptive systems. This article explores a comprehensive framework for such networks, detailing the mathematical representation of substrate networks as attribute-rich graphs and introducing sophisticated mechanisms for dynamic resource mapping. By incorporating application-specific optimization tailored to diverse workload requirements and leveraging predictive resource allocation through machine learning, these systems proactively address potential performance bottlenecks before they emerge. Experimental results demonstrate significant improvements over traditional network management approaches in key metrics including latency management, resource utilization, adaptation to changing conditions, and failure recovery. The implementation balances the benefits of specialized optimization with the practicality of generalized approaches, while identifying promising future research directions to enhance scalability, explainability, and cross-domain optimization capabilities.
Author: Mohan Ranga Rao Dontineni
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4410464264