BMC Discovery — You Can't Manage What You Can't See 🗺️

A question that sinks careers. Your boss asks: "if we shut down server DB-CORE this weekend, what breaks?" If your answer is "...let me ask around and get back to you," you have already lost. In a 10,000-server estate, no human knows by heart. BMC Discovery does — automatically.

What it is: BMC Discovery (historically ADDM — Atrium Discovery and Dependency Mapping) is a tool that scans your entire IT estate — servers, applications, network gear, cloud instances — and builds an always-current map of what you have and how it is connected. That map lives in a CMDB.

The CMDB — the single source of truth

A CMDB (Configuration Management Database) is the inventory of every CI (Configuration Item) — each server, database, app, service — and the relationships between them. ITIL's processes from the last lesson all lean on it: you cannot assess the impact of a change if you do not know what depends on what.

🧠 Mental model: Discovery is the Google Maps of your data centre. It does not just list the buildings (servers) — it draws the roads between them (dependencies), and keeps the map updated as things change.

Why automatic discovery beats a spreadsheet

Teams used to track infrastructure in a hand-maintained spreadsheet. It was wrong the day after it was written — someone spins up a server, nobody updates the sheet. Discovery scans continuously, so the map reflects reality, not someone's memory. A stale CMDB is worse than none — because people trust it.

The killer feature: impact analysis

Because Discovery knows the dependency graph, it can answer the career-saving question instantly: "if this fails, what is affected?" Trace the arrows pointing at a failing item and you have your blast radius.

🔧 Where you meet it

ServiceNow, BMC Helix, Device42 and others all do this "discovery + CMDB" job. When an incident ticket auto-fills "affected service: Trading Platform," that link came from a discovery-built CMDB.

Your Task

Here is a tiny slice of a CMDB: a map of each system and what it runs on. Run an impact analysis — if one item fails, find everything that directly depends on it.