Skip to content
← All casesCMDB / ITSMEnterprise

CMDB from Scratch

Configuration management database for IT landscape of a 140K+ railcar operator

Problem

What doesn't work

No unified configuration item (CI) database for the IT landscape of a major operator. Information about servers, applications, networks, and dependencies was stored in scattered spreadsheets. During incidents, it was impossible to quickly determine impact and dependent systems.

Solution

Architectural approach

Designed CMDB from scratch: CI data model, relationships and dependencies, integration with monitoring and incident management. Automatic CI discovery and updates. Dependency visualization for impact analysis.

Challenges

What made it hard

Server data was scattered across dozens of Excel spreadsheets owned by different admins — consolidating into a unified data model required personally visiting each owner. Auto-discovery only works in managed networks — part of the infrastructure was 'dark' with no monitoring agents. Keeping CMDB current is the hardest part: without a change management process, the database goes stale within weeks.

Role

My role & contribution

CTO / Technical Director

Designed the CMDB data model from scratch: CI classes (servers, applications, networks, DBs), relationships and dependencies. Defined auto-discovery and monitoring integration requirements. Linked change management processes to CMDB.

Demo

How it looks

Implementation

How it works

CI data model design → class and attribute definition → monitoring system integration → automatic discovery → data verification. Change management processes linked to CMDB.

Architecture Decision

Why this way

Custom CMDB instead of off-the-shelf product

Alternative

Deploy enterprise CMDB solution (ServiceNow, BMC)

Why it didn't fit

Off-the-shelf solutions: high license costs, excessive functionality, complex customization for railway operator specifics. Custom: exact match to PGK's IT landscape data model.

Result

CMDB precisely reflecting the operator's IT infrastructure. No enterprise ITSM license costs

Metrics

Results

01
CMDB designed and implemented from scratch
02
All CI classes: servers, applications, networks, DBs
03
Automatic discovery and updates
04
Impact analysis via dependencies
05
Incident management integration
Business Impact

Impact on business

Reduced incident MTTR through instant impact analysis. Full IT landscape transparency for change planning. Foundation for ITSM processes: change management, problem management, capacity planning.

Methods

Algorithms & patterns

CMDB Data ModelingAuto-DiscoveryImpact AnalysisITIL Processes
Stack

Technologies

  • CMDB
  • ITSM
  • Auto-Discovery
  • Monitoring Integration

Ready to discuss?

If you need an architect who builds autonomous AI systems — reach out.

Serbia-based · CET/CEST timezone · EU-aligned working hours · International contracts experience