CDT Architecture & System Design 2 — Questions and Answers
Question 1: Which document management architecture pattern separates content storage from metadata indexing to optimize search performance?
- Monolithic repository
- Federated content store
- Decoupled content and index architecture (Correct answer)
- Flat file hierarchy
Correct answer: Decoupled content and index architecture
Decoupled content and index architecture stores binary content separately from searchable metadata, allowing each layer to be optimized independently.
Question 2: In a distributed document management system, what is the primary purpose of a content-addressable storage (CAS) layer?
- To enforce user access controls on documents
- To deduplicate files by storing them by their cryptographic hash (Correct answer)
- To route documents to the nearest geographic node
- To convert documents between file formats automatically
Correct answer: To deduplicate files by storing them by their cryptographic hash
CAS stores files indexed by a hash of their content, so identical files are stored only once, enabling automatic deduplication.
Question 3: A document system requires that all edits be reversible and auditable indefinitely. Which storage pattern best supports this requirement?
- Delta compression with periodic snapshots
- Append-only immutable versioning (Correct answer)
- In-place update with transaction logs
- Cyclic overwrite buffer
Correct answer: Append-only immutable versioning
Append-only immutable versioning never modifies existing records, ensuring every state is preserved and auditable.
Question 4: Which component in an enterprise content management (ECM) architecture is responsible for enforcing document retention schedules automatically?
- Rendition engine
- Records management module (Correct answer)
- OCR pipeline
- Full-text search index
Correct answer: Records management module
The records management module applies retention policies, triggering disposition actions (hold, archive, or delete) based on schedule rules.
Question 5: When designing a system for high-volume document ingestion, which queuing strategy prevents back-pressure from overwhelming downstream processors?
- Synchronous REST calls with retry logic
- Priority queue with dead-letter channel (Correct answer)
- Round-robin database insertion
- Direct socket streaming to all consumers
Correct answer: Priority queue with dead-letter channel
A priority queue with a dead-letter channel buffers incoming documents, routes failures for inspection, and absorbs ingestion bursts without dropping data.
Question 6: In document system design, what does 'eventual consistency' imply for a replicated content store?
- All nodes always reflect the latest write in real time
- Reads may return stale data until replication completes (Correct answer)
- Documents are locked during cross-node synchronization
- Only one node is authoritative and others are read-only
Correct answer: Reads may return stale data until replication completes
Eventual consistency allows replicas to diverge temporarily; given no new updates, all nodes will converge to the same state over time.
Question 7: Which architectural approach allows a document management system to support multiple client application types (web, mobile, desktop) without duplicating business logic?
- Fat client with embedded database
- API-first headless architecture (Correct answer)
- Monolithic server-side rendering
- Thick client with local document cache only
Correct answer: API-first headless architecture
An API-first headless architecture exposes business logic through standardized APIs, letting any client consume the same services without logic duplication.
Which document management architecture pattern separates content storage from metadata indexing to optimize search performance?