Conceptual Architecture of Automated Inventory and Topological Modeling in Chaotic Multi-Vendor Networks
Abstract
Modern enterprise networks are complex heterogeneous systems comprising equipment from various vendors (HPE, Huawei, SNR, Fortinet, Ubiquiti). This paper proposes and validates a conceptual architecture for automated device inventory and topological modeling in chaotic multi-vendor LANs. The network topology is modeled using Graph Theory (G=(V,E,W)), SNMP/LLDP data is automatically normalized, and identification is optimized using Go microservices. Real-world testing across a ~1000 IP enterprise network demonstrated that the system fully inventoried 67 management devices and 1600+ endpoints in 10 minutes and 19 seconds. A two-phase scanning approach (ICMP Discovery + Deep Scan) yielded a 4.3x overall speedup over sequential methods (44x speedup in initial discovery). Furthermore, automated trunk-port filtering of virtual machines resulted in a 91.4% precision rate in detecting unauthorized Shadow IT devices. Experimental validation confirmed significant structural advantages over traditional NMS platforms (Zabbix), serving as a robust foundation for future Software-Defined Networking (SDN) and AI/ML-based monitoring models.