From Network Segmentation to AI-Driven Zero Trust: A Systematic Survey of Micro-segmentation Technologies

The rapid adoption of cloud-native architectures has exposed critical limitations in traditional network security approaches, particularly as organizations struggle to protect dynamic workloads across hybrid environments. While micro-segmentation promises granular security controls, practitioners face a diverse range of implementation choices—from basic VLAN configurations to sophisticated AI-driven policy engines. This paper presents a systematic analysis of micro-segmentation technologies through a comprehensive review of recent literature from major academic and industry venues. We trace the evolution from static network-based segmentation through identity-centric approaches to emerging AI-powered Zero Trust implementations. Our analysis reveals that modern kernel-level solutions have fundamentally altered the security-performance equation, enabling fine-grained protection without the prohibitive overhead of earlier approaches. We identify distinct implementation patterns, including cloud-provider native services, ranging from straightforward flow-based segmentation to complex service mesh architectures. Through comparative analysis, we find that containernative approaches now dominate production environments, with several open-source projects demonstrating maturity for mission-critical deployments. This survey provides practitioners with actionable guidance for selecting appropriate micro-segmentation strategies based on their security requirements, infrastructure constraints, and operational maturity.