Databases on AWS

Why this matters

Almost every application needs to store data somewhere, and the database is where that data lives. On the CLF-C02 exam, you do not need to design schemas or write queries. You need to recognize the AWS database services, know whether a service is relational or non-relational, and match a described need to the right one.

This lesson sets up the rest of the topic. It covers what a managed database service is, how it differs from running a database yourself, the split between relational and non-relational data, and the idea of purpose-built databases. The next lessons go deeper into the specific services.

Managed service versus self-managing

You can run a database on AWS in two broad ways.

The first is to launch an Amazon EC2 instance, install the database software yourself, and manage everything: the operating system, patches, backups, scaling, and recovery. This gives you full control, but you carry all the operational work, and small mistakes during a manual patch or upgrade can cause downtime.

The second is to use a managed database service such as Amazon RDS. AWS provisions the database, applies patches, takes automated backups, detects failures, and handles recovery. You focus on your data and your application instead of the plumbing underneath. AWS recommends a managed service as the default for most workloads.

The table below shows who handles what across three options.

Notice the pattern: as you move to a managed service, AWS takes on more of the routine work. One thing never moves. Tuning your own queries stays with you in every model, because it depends on your data and how your application uses it.

Relational and non-relational databases

Databases fall into two high-level groups, and the exam expects you to tell them apart.

A relational database (often called SQL) stores data in tables made of rows and columns, with a defined schema. You can link, or join, tables to answer questions that span related data, such as matching customers to their orders. Relational databases are a strong fit for structured data and transactions where accuracy matters, like finance, orders, and inventory.

A non-relational database (often called NoSQL) uses other data models that trade some of the strict structure for scale, speed, and flexibility. Instead of fixed tables, it might store key-value pairs, JSON-like documents, graphs, or time-stamped points. These models shine when you need very high throughput, low latency, or a flexible schema that changes often.

Neither group is "better." They solve different problems, which is exactly why AWS offers both.

Purpose-built databases

AWS does not push one database for every job. Instead it offers a range of purpose-built databases, each optimized for a specific data model and workload. The idea is simple: choose the right tool for each task. A single application can use several at once, such as a relational database for billing and a key-value database for a shopping cart.

You do not need deep detail on every service for the exam. You do need to recognize the name, know which category it belongs to, and match a scenario to it. The next three lessons cover the most tested services in each group, starting with relational databases.

Exam tips

• A managed service (like Amazon RDS) means AWS handles patching, backups, failover, and hardware. Running a database on EC2 means you handle all of that yourself.
• Query tuning is always your job, even with a managed service.
• Relational (SQL) = tables, rows, columns, joins, fixed schema. Non-relational (NoSQL) = key-value, document, graph, and other flexible models built for scale.
• Purpose-built means AWS offers a database for each kind of workload, and you pick the right one rather than forcing everything into a single database.
• Learn to map a category to its service: relational to RDS and Aurora, key-value to DynamoDB, in-memory to ElastiCache, graph to Neptune, document to DocumentDB.