SCALABLE SDN-BASED ARCHITECTURE FOR LARGE-SCALE ENTERPRISE NETWORK MANAGEMENT

Abstract

Software-Defined Networking (SDN) has emerged as a transformative paradigm for enterprise network management, offering centralized control and programmability. However, existing SDN architectures face significant scalability challenges when deployed in large-scale enterprise environments with thousands of network devices and dynamic traffic patterns. This study proposes a novel hierarchical SDN architecture incorporating distributed controller clusters, dynamic load balancing, and optimized flow management mechanisms to address these limitations. We implemented and evaluated the proposed architecture using Mininet emulation with Fat-Tree topologies ranging from 16 to 128 hosts. Performance evaluation demonstrates that the proposed architecture achieves a 42% reduction in flow setup latency (5.2 ms vs. 8.5 ms), 18% improvement in throughput (850K vs. 720K flows/sec), and 65% reduction in packet loss compared to traditional single-controller SDN deployments. Statistical analysis using ANOVA confirmed significant performance improvements (p < 0.001) across all metrics. These findings establish that hierarchical SDN architectures with intelligent load balancing provide a viable solution for scalable enterprise network management, enabling organizations to leverage SDN benefits in production environments without compromising performance or reliability.