Databases on AWS

RDS vs DynamoDB vs Aurora vs Redshift — matching database services to relational, NoSQL, caching, and analytics needs.

10 min read

AWS offers purpose-built databases. The exam gives you a workload description and expects you to name the database. Two axes matter most: relational vs non-relational, and transactional vs analytical.

Relational (SQL)

Amazon RDS

Managed relational databases (MySQL, PostgreSQL, MariaDB, Oracle, SQL Server): AWS handles patching, backups, and Multi-AZ failover.

Amazon Aurora

AWS's cloud-native MySQL/PostgreSQL-compatible engine: up to 5x MySQL performance, 6 copies of data across 3 AZs. Aurora Serverless scales automatically.

Amazon Redshift

Data warehouse for analytics (OLAP): run complex SQL over petabytes of historical data for BI and reporting.

Tip

RDS Multi-AZ = availability (standby replica for failover). Read replicas = performance (scale out reads). This pairing appears in many questions.

Non-relational (NoSQL) and specialized

Amazon DynamoDB

Serverless key-value/document NoSQL with single-digit-millisecond latency at any scale. No servers to manage; global tables for multi-Region.

Amazon ElastiCache

Managed in-memory caching (Redis/Memcached) for microsecond reads that take load off databases.

Amazon MemoryDB

Redis-compatible, durable in-memory database.

Amazon DocumentDB

Managed MongoDB-compatible document database.

Amazon Neptune

Graph database for highly connected data: social networks, recommendations, fraud rings.

Amazon QLDB

Immutable, cryptographically verifiable ledger database for systems of record.

Amazon Keyspaces / Timestream

Managed Apache Cassandra-compatible, and time-series databases respectively.

Exam tip

Keyword → database: "relational / SQL / transactions" → RDS/Aurora; "key-value, millisecond latency, serverless" → DynamoDB; "data warehouse / BI / analytics on historical data" → Redshift; "in-memory cache" → ElastiCache; "graph / relationships" → Neptune; "MongoDB" → DocumentDB; "immutable ledger" → QLDB.

Think of it like this

RDS is a filing cabinet with strict folders and cross-references (relational). DynamoDB is a wall of labeled cubbies — grab by label, lightning fast, endlessly expandable (key-value). Redshift is the records archive where analysts run big historical studies. ElastiCache is the sticky notes on your monitor for the things you look up constantly.

Knowledge check
Question 1 of 4

Which AWS service provides a managed, serverless key-value NoSQL database with single-digit millisecond performance?