Before we dive into the different database offering from AWS, let's have quickly a look at the differences in how data is managed and stored (Relational vs Non-Relational) and different use cases - Operational/Transactional (OLTP ) vs Analytical (OLAP))

Relational Non Relational

tables, rows and columns many different data storage models ( key-value pairs, documents or graphs)

rigid schema (SQL ) Flexible schema (NoSql)

rules enforced within the database defined in application code

vertical scalability horizontally

supports joins and complex queries unstructured simple query language

OLTP (Online Transaction Processing) OLAP (Online Analytics Processing

transaction-oriented tasks analysis and decision making oriented tasks

OLTP typically involves inserting, updating, and/or deleting small amounts of data in a database OLAP performs multidimensional analysis of business data and provides the capability for sophisticated data modelling

Short transactions and simple queries long transactions and complex queries

ie production DB ie data-warehouse

OLTP vs OLAP

data in OLAP often comes (is aggregated) from multiple OLTP sources.

OLTP to ETL to OLAP

OLAP data-warehouses are useful to avoid performance hits on the original source, but, for very simple scenarios having a read Replica would do the job.

As with many other services AWS provides us with a very extensive Datastore offering to fit different requirements in terms of use case, scalability, performance and price.

Amazon RDS (OLTP)

DynamoDB (OLTP)

DocumentDB

RedShift (OLAP)

Elasticache

EMR (hadoop) (OLAP)

Database on EC2 ( other database not listed above can be directly installed on EC2)