DAS EquipmentMonitoring Platform

The Problem

One company's distributed antenna system spanned over 12 buildings with equipment from five different vendors — Zinwave, ADRF, JMA Teko, Eaton, and AVTECH — each with its own management interface, proprietary protocols, and alarm formats.

Technicians had to log into each device individually or drive to sites to check status. Faults were discovered only after users reported coverage issues, sometimes days after the underlying failure. There was no historical trending, no centralized visibility, and no way to identify degrading performance before it became an outage.

The Solution

We built a cloud-hosted monitoring platform that connects to all sites via encrypted WireGuard VPN tunnels. The system polls every device for health metrics via SNMP and receives real-time fault notifications via SNMP traps, consolidating everything into a single dashboard with automated daily email digests.

Architecture

What we built

• 17 custom monitoring templates covering 8 device types from 5 vendors, each built from scratch by analyzing vendor MIB files
• Custom SNMP trap pipeline that handles NAT translation, attributing every trap to the correct physical device regardless of network topology
• Automated daily digest emails with severity-coded event summaries and direct links to device management interfaces
• Encrypted site connectivity via WireGuard VPN tunnels to cellular routers at each building
• Separate management and monitoring VPNs — the monitoring tunnel carries only SNMP polling and trap traffic, while a separate management VPN provides secure access to device web interfaces for configuration and troubleshooting.

The Outcome

• Single dashboard replaced 30+ individual device logins
• Faults detected within seconds instead of days
• Eliminated diagnostic site visits — technicians know exactly which building, rack, and module has the fault before dispatching
• Adding a new building requires only a VPN tunnel and template assignment
• Runs on a single cloud instance with no per-device licensing fees

The Systems Engineering Challenge

The hard part wasn't setting up a monitoring server. It was understanding five vendors' SNMP implementations, alarm semantics, and data formats well enough to build a unified view. ADRF devices NAT through site routers, so traps arrive from the wrong IP. Zinwave stores RF power values as scaled integers. Each vendor defines "fault" differently.

Solving this required systems engineering — understanding the domain deeply enough to design abstractions that work across all of it. That's the work AI can't do alone.