Tier 1 — Must Know Cold
1. Business vs Technical vs Operational
Business: meaning, definitions, rules, ownership, standards, quality/security context.
Technical: physical structures, mappings, transformations, schemas, jobs, dependencies, technical lineage.
Operational: logs, runtime status, errors, usage, execution time, SLAs, operational evidence.
Deciding clue: meaning/governance vs implementation/movement vs runtime evidence.
2. Five Metadata Management goals
Terminology knowledge; integrate diverse Metadata; quality/consistency/currency/security; standard access; technical standards/exchange.
3. Eight guiding principles
Organizational Commitment → Strategy → Enterprise Perspective → Socialization → Access → Quality → Audit → Improvement.
4. Glossary vs Dictionary vs Directory/Catalog
- Glossary = business concepts/meaning/ownership.
- Dictionary = element/data-set properties and structure.
- Catalog/Directory = discovery/location.
5. Data Model vs Metamodel
Data model = business/data structures.
Metamodel = Metadata entities/attributes/relationships for the Metadata environment.
6. Centralized architecture
Persistent central repository stores harvested copies. Clue: central persistence.
7. Distributed architecture
Common access point, source queries in real time, no persistent central repository.
8. Hybrid architecture
Selected Metadata central; other detail remains source-held and is retrieved as needed.
9. Bi-Directional architecture
Controlled enterprise changes can flow back to originating source repositories.
10. Five core activities
Define Strategy → Understand Requirements → Define Architecture → Create/Maintain Metadata → Query/Report/Analyze Metadata.
Architecture sub-activities: Metamodel; Standards; Manage Stores.
Create/Maintain sub-activities: Integrate; Distribute/Deliver.
11. Strategy before tool
Charter/scope, stakeholders, interviews, current sources, future state, phased plan come before product selection/configuration.
12. Metadata requirements
Content/detail, volatility, workflow/approval, roles, quality, search/access/integration, history/versioning, and security.
Trap: Metadata can reveal sensitive data existence/location/structure even without exposing values.
13. Metadata as a product
Metadata requires planned creation, accountable source processes, standards, stewardship, integration, quality, change control, feedback, and delivery.
14. Integrate vs Deliver
Integrate = harvest/stage/standardize/map/resolve/merge.
Deliver = make the managed result accessible via portals, documents, applications, files, or services.
15. As Designed vs As Implemented
Designed = intended mapping/specification.
Implemented = actual code/jobs/production path.
16. Lineage vs Impact Analysis
Lineage = origin/movement/transformation/path.
Impact = dependencies/consequences of a proposed change.
17. Metadata metrics
Completeness; maturity; Steward representation; usage; glossary activity; Master Data service compliance; documentation quality.
Deciding rule: match the metric to the management question.
Source: pp. 395–423.