Authors: Narendra Reddy Burramukku
Abstract: The rapid evolution of cloud computing and software-defined networking has led to increasingly complex, distributed, and dynamic network infrastructures. Managing performance, reliability, and security in such environments requires intelligent, automated, and adaptive network management solutions. Artificial Intelligence (AI) and Machine Learning (ML) have emerged as transformative technologies for enhancing network management by enabling predictive analytics, anomaly detection, traffic optimization, and autonomous decision-making. This paper presents a comprehensive survey of AI- and ML-driven techniques for modern network management, with a focus on cloud, software-defined, and virtualized network environments. The study systematically reviews supervised, unsupervised, and reinforcement learning approaches applied to network monitoring, fault detection, traffic engineering, and security management. It further analyzes architectural frameworks integrating AI/ML with Software-Defined Networking (SDN), Network Function Virtualization (NFV), and cloud-native platforms. Key challenges such as data quality, model scalability, explainability, real-time decision-making, and security risks are critically discussed. Additionally, the paper highlights emerging research directions including self-healing networks, intent-based networking, federated learning, and AI-driven zero-touch network operations. By consolidating algorithms, architectures, and application domains, this survey provides a structured reference for researchers and practitioners aiming to design intelligent, scalable, and resilient network management systems.
