CIA Metadata Strategy and Design 3 — Questions and Answers
Question 1: When designing metadata for a data lake, which challenge is MOST unique compared to a traditional data warehouse?
- Defining business glossary terms
- Capturing metadata for schema-on-read and diverse unstructured formats (Correct answer)
- Assigning data stewards to datasets
- Creating data dictionaries
Correct answer: Capturing metadata for schema-on-read and diverse unstructured formats
Data lakes store raw, unstructured, and semi-structured data where schemas are applied at read time, making metadata capture far more complex than warehouse environments.
Question 2: Which ISO standard specifically addresses the framework for metadata registries and data element concepts?
- ISO 8000
- ISO 11179 (Correct answer)
- ISO 15836
- ISO 25012
Correct answer: ISO 11179
ISO/IEC 11179 defines the framework for metadata registries, specifying how data elements, their concepts, and representations should be registered and managed.
Question 3: An organization is implementing metadata tagging for sensitive data. Which type of metadata classification supports automated data protection policies?
- Descriptive metadata
- Data sensitivity or classification metadata (Correct answer)
- Structural metadata
- Provenance metadata
Correct answer: Data sensitivity or classification metadata
Data sensitivity classification metadata (e.g., PII, Confidential) enables automated policy enforcement by data governance and security tools.
Question 4: What is 'passive metadata' in the context of enterprise information architecture?
- Metadata that has been archived and is no longer active
- Metadata that is stored but not actively used to drive system behavior or decisions (Correct answer)
- Metadata without owner assignments
- Metadata that describes inactive data assets
Correct answer: Metadata that is stored but not actively used to drive system behavior or decisions
Passive metadata exists in repositories but is not operationalized to influence system processing, access control, or automated decisions.
Question 5: In a metadata maturity model, what capability characterizes the 'managed' level?
- Ad hoc metadata creation with no standards
- Metadata standards defined but inconsistently applied
- Metadata processes are measured, monitored, and consistently applied enterprise-wide (Correct answer)
- Metadata is optimized and continuously improved with predictive capabilities
Correct answer: Metadata processes are measured, monitored, and consistently applied enterprise-wide
At the managed level, metadata processes are quantitatively controlled with consistent application of standards and active monitoring of metadata quality.
Question 6: Which technique is used to automatically discover and classify metadata based on patterns within the data itself?
- Manual data profiling
- Data fingerprinting and pattern-based classification (Correct answer)
- User-submitted tagging only
- Schema import from source systems
Correct answer: Data fingerprinting and pattern-based classification
Data fingerprinting and pattern-based classification use algorithms to scan data content and automatically assign metadata tags based on recognized patterns (e.g., SSN format, email structure).
Question 7: A business glossary term has multiple competing definitions from different business units. What is the RECOMMENDED resolution approach in metadata governance?
- Accept all definitions and create separate entries for each
- Delete all definitions and start fresh
- Facilitate a business-driven consensus process to agree on a single authoritative definition (Correct answer)
- Let the IT department choose the technical definition
Correct answer: Facilitate a business-driven consensus process to agree on a single authoritative definition
Governance best practice calls for a facilitated business consensus process so the authoritative definition reflects agreed-upon business meaning across all stakeholders.
CIA Metadata Strategy and Design 3
When designing metadata for a data lake, which challenge is MOST unique compared to a traditional data warehouse?