CMDB vs CSDM — Scope and Purpose Explained

Although often used interchangeably, CMDB and CSDM serve fundamentally different purposes. Confusing the two is one of the main reasons 75% of the CMDB initiatives fail.
What Is the CMDB?
The Configuration Management Database (CMDB) is an operational data store.
CMDB scope
The CMDB contains:
-
Deployed configuration items (CIs)
-
Runtime instances and their technical relationships
-
Discovered and federated infrastructure and platform components
-
Service instances and their dependencies
-
Operational attributes required for ITSM, ITOM, SecOps, and risk use cases
CMDB characteristics
-
Focused on runtime reality
-
Populated primarily through automated discovery and integrations
-
Continuously changing
-
Optimised for operations, impact analysis, incident, change, and risk
-
Not authoritative for design-time or portfolio decisions
The CMDB answers: “What is running, where, and what depends on it—right now?”
What Is CSDM?
The Common Service Data Model (CSDM) is a logical reference model.
CSDM scope
CSDM defines:
-
Which object types exist (applications, services, offerings, instances, products)
-
How those objects relate to each other
-
Where ownership, accountability, and commitments belong
-
Which artefacts are portfolio-managed versus runtime-discovered
CSDM characteristics
-
Stable and opinionated
-
Service-centric by design
-
Independent of tools and discovery methods
-
Focused on governance, consistency, and shared understanding
-
Applicable beyond the CMDB (APM, ITSM, IRM, TBM, EA)
CSDM answers: “What should exist, how it fits together, and who is accountable?”
The Key Difference in One Sentence
CSDM defines the structure and rules; the CMDB contains the data that conforms to those rules.
Scope Comparison
| Aspect | CMDB | CSDM |
|---|---|---|
| Nature | Operational data store | Logical reference model |
| Purpose | Operational visibility & impact | Structural consistency & governance |
| Focus | Runtime reality | Service & portfolio structure |
| Change frequency | High (continuous) | Low (intentional) |
| Data source | Discovery, integrations | Architecture & design principles |
| Owns data | No (consumes authoritative sources) | No (defines where ownership belongs) |
| Audience | Ops, ITSM, SecOps | Architects, governance, platform teams |
What Belongs Where (Rule of Thumb)
Belongs in the CMDB
-
Service Instances
-
Deployed servers, containers, cloud resources
-
Runtime dependencies
-
Discovered software and infrastructure
-
Dynamic CI groups
Defined by CSDM (but instantiated in CMDB)
-
Business and Technical Applications
-
Business and Technology Services
-
Service Offerings
-
Service Instance patterns
-
Ownership and accountability model
Does not belong in the CMDB
-
Design-time artefacts
-
CI/CD pipelines
-
Source code and build metadata
-
Ephemeral runtime events
-
DevSecOps tool internals
Why This Distinction Matters
When CMDB and CSDM are conflated:
-
CMDBs become bloated and unmaintainable
-
Discovery data loses context
-
Governance becomes manual and political
-
DevOps teams resist integration
-
Audit evidence becomes inconsistent
When they are correctly separated:
-
Discovery becomes safe and scalable
-
Service ownership is clear
-
Impact analysis works
-
Risk and compliance evidence aligns naturally
-
DevSecOps and CMDB coexist without friction
The Modern Operating Model
-
CSDM provides the stable, service-centric blueprint
-
CMDB provides the dynamic, runtime data
-
Discovery and federation keep the two aligned
-
Portfolio items are managed; runtime items are observed



