Cloud Technology and Services
Managed Databases and Database Migration
CoreDistinguish relational, NoSQL, and in-memory database services, then separate data movement from schema conversion during migration.
Aligned to the current CLF-C02 exam guide, verified August 15, 2026.
Why this matters
Database questions reward recognition of the data model and management need, not engine tuning or query syntax.
Must Know
- Amazon RDS is a managed relational database service for supported engines. Amazon Aurora is an AWS-designed relational engine compatible with MySQL and PostgreSQL.
- A database on Amazon EC2 gives the customer operating-system and database administration control. RDS and Aurora automate more infrastructure tasks such as supported backups, patching, and failover.
- Amazon DynamoDB is a fully managed, serverless NoSQL key-value and document database.
- Amazon ElastiCache is a managed in-memory data store and cache used to reduce latency and load on a primary database.
- AWS Database Migration Service (AWS DMS) moves and can continuously replicate supported data. AWS Schema Conversion Tool (AWS SCT) assesses and converts schemas and code objects for supported heterogeneous migrations.
Compare and Distinguish
- EC2-hosted database vs RDS: EC2 offers OS-level control and more administration; RDS manages more of the database infrastructure.
- RDS/Aurora vs DynamoDB: RDS and Aurora are relational; DynamoDB is NoSQL key-value/document.
- DynamoDB vs ElastiCache: DynamoDB is a durable NoSQL system of record; ElastiCache commonly accelerates access to data held elsewhere.
- DMS vs SCT: DMS moves data; SCT converts supported schema objects when the target engine differs.
Scenario examples
- Scenario: An application needs SQL relationships, but the company wants AWS to handle routine database infrastructure tasks. Think: Amazon RDS or Amazon Aurora.
- Scenario: A large application needs a serverless key-value database. Think: Amazon DynamoDB.
- Scenario: Frequently read product data should be served from memory to reduce load on the primary database. Think: Amazon ElastiCache.
- Scenario: A migration changes database engines and needs both schema conversion and ongoing data replication. Think: AWS SCT for schema conversion and AWS DMS for data movement.
Exam traps
- Managed database does not mean the customer has no responsibility for data, identities, schemas, or application queries.
- ElastiCache is not automatically the durable database of record.
- DMS does not replace schema conversion in every heterogeneous migration, and SCT does not move the production data.
- NoSQL is not simply “SQL with less management”; it uses a different data model and access approach.
Key takeaways
- Managed relational: RDS or Aurora. Serverless NoSQL: DynamoDB. In-memory cache: ElastiCache.
- Self-managed database with OS control: EC2.
- Move data: DMS. Convert schema: SCT.
How it works
- RDS and Aurora provision managed database infrastructure while the customer creates schemas, controls access, and uses the data.
- DynamoDB stores items by keys in tables and automatically manages the underlying database infrastructure.
- ElastiCache keeps suitable data in memory so applications can avoid repeated slower reads.
- During migration, SCT can prepare a target schema and DMS can load and replicate data.
When to use it
- Use RDS or Aurora for managed relational workloads.
- Use DynamoDB for a managed NoSQL key-value or document workload.
- Use ElastiCache when an in-memory cache should reduce response time or database load.
- Use DMS for data migration and SCT when supported cross-engine schema conversion is required.
Security and governance implications
- Managed services reduce infrastructure administration, but customers still control identities, network access, encryption choices, and data classification.
- Migration permissions should be scoped to the source and target resources required.
Troubleshooting signals
- If a service choice looks ambiguous, identify whether the requirement is relational structure, durable NoSQL storage, or temporary acceleration.
- For a migration mismatch, separate source-to-target connectivity and replication from schema compatibility.
More detail
- Relational databases organize structured data with defined relationships. NoSQL databases support other models and access patterns that can scale without relational joins.
- Other in-scope databases have more specialized recognition roles: Amazon DocumentDB supports MongoDB-compatible document workloads, and Amazon Neptune is a graph database.
- For supported migrations, DMS can keep source and target synchronized to reduce downtime. SCT helps evaluate and translate compatible schema objects before a different target engine is used.
Ready for the quiz?
- Which requirement makes DynamoDB more suitable than RDS?
- What administration remains with a customer using a managed database?
- Why might a heterogeneous migration use both DMS and SCT?
Related objectives
- D3.4