Amazon DynamoDB Developer Guide

What DynamoDB is

AWS describes DynamoDB as a serverless, fully managed NoSQL database with single-digit millisecond performance. It supports key-value and document data models. There are no servers, versions or maintenance windows to manage: you create a table and send reads and writes. It has no JOIN operator, so you design your tables around the exact questions your app will ask.

What the guide covers

• Core components: tables, items, attributes, the partition key and the optional sort key that together form the primary key.
• Secondary indexes: global and local secondary indexes for querying by something other than the primary key.
• Capacity modes: on-demand (pay per request, scales down to zero) and provisioned (you set read and write capacity).
• Consistency and transactions: strongly consistent reads when you ask for them, and ACID transactions across items and tables.
• Streams: item-level change records that can trigger AWS Lambda functions.
• Best practices: NoSQL design, single-table design and modelling relationships without joins.
• Tools: the console, the AWS CLI, NoSQL Workbench and a downloadable local version for development.

Is it worth learning?

If you build on AWS, especially serverless apps with Lambda, DynamoDB is the default database many teams reach for, and the modelling style is worth understanding. It also teaches the key-value way of thinking: you get speed and simplicity by deciding your access patterns first. The trade-off is flexibility: new kinds of query often mean new indexes or a changed design, so a relational database is usually easier while requirements are still moving.

Prerequisites and time

Basic programming and an AWS account. A weekend to model a small app (users, posts, comments), write the access patterns down and implement them.

Free or paid?

The guide is free. As of October 2026 the guide says DynamoDB is in the AWS always free tier with 25 GB of storage and 25 read and 25 write capacity units in provisioned mode; check the pricing page before relying on it, and set a billing alarm.

Common mistake

Designing tables like a relational schema, one table per entity, then discovering you cannot join them. List the queries first.

Best for: backend developers building on AWS, particularly serverless apps.