CDS Architecture & System Design 2 — Questions and Answers
Question 1: Which architectural pattern separates read and write operations into distinct models to optimize scalability?
- Event Sourcing
- CQRS (Command Query Responsibility Segregation) (Correct answer)
- Microservices
- Data Vault
Correct answer: CQRS (Command Query Responsibility Segregation)
CQRS splits the data model into separate command (write) and query (read) sides, allowing each to be scaled and optimized independently.
Question 2: In a federated data architecture, what is the primary responsibility of a data domain owner?
- Managing the central data warehouse ETL jobs
- Owning, publishing, and maintaining data products for their domain (Correct answer)
- Enforcing enterprise-wide data quality SLAs
- Controlling access to the master data management system
Correct answer: Owning, publishing, and maintaining data products for their domain
In a federated (data mesh) architecture, domain owners are responsible for treating their data as a product and maintaining its quality and availability.
Question 3: What does a 'data contract' define in a modern data architecture?
- The legal agreement between a vendor and the data team
- The SLA between a data producer and its consumers regarding schema, quality, and delivery (Correct answer)
- The encryption standard applied to data at rest
- The cost allocation model for cloud data storage
Correct answer: The SLA between a data producer and its consumers regarding schema, quality, and delivery
A data contract is a formal agreement between producer and consumer specifying schema, semantics, quality expectations, and delivery guarantees.
Question 4: Which type of data store is BEST suited for managing highly connected entity relationships, such as social networks or fraud detection graphs?
- Columnar data warehouse
- Relational OLTP database
- Graph database (Correct answer)
- Time-series database
Correct answer: Graph database
Graph databases natively store nodes and edges, making traversal of complex relationships far more efficient than joins in relational systems.
Question 5: A 'schema-on-read' approach, common in data lakes, means:
- The schema is enforced when data is written to storage
- The schema is applied and interpreted when data is queried (Correct answer)
- All data must conform to a single global schema
- Schema validation is performed by the ETL pipeline before ingestion
Correct answer: The schema is applied and interpreted when data is queried
Schema-on-read defers schema enforcement to query time, allowing raw data to be stored flexibly and interpreted differently by each consumer.
Question 6: What is the primary purpose of a data virtualization layer in an enterprise architecture?
- To physically consolidate all data into a single repository
- To provide a unified query interface over distributed data sources without moving data (Correct answer)
- To compress and archive historical data to reduce storage costs
- To enforce row-level security across all data warehouses
Correct answer: To provide a unified query interface over distributed data sources without moving data
Data virtualization presents a logical unified view of data across disparate sources, enabling queries without requiring physical data movement or replication.
Question 7: In the Lambda architecture, what is the role of the 'speed layer'?
- To reprocess the entire historical dataset for accuracy
- To serve pre-computed batch views with low latency
- To process recent streaming data and provide low-latency approximate results (Correct answer)
- To manage schema evolution across batch and serving layers
Correct answer: To process recent streaming data and provide low-latency approximate results
The speed layer processes real-time streaming data to fill the latency gap of the batch layer, providing recent but potentially approximate results.
CDS Architecture & System Design 2
Which architectural pattern separates read and write operations into distinct models to optimize scalability?