Amazon DynamoDB

What is Amazon DynamoDB?

Amazon DynamoDB is AWS’s managed NoSQL database. Unlike relational databases, DynamoDB does not organise data into tables with fixed relationships, but rather as flexible key-value and document structures. This makes it particularly fast and easily scalable – at the cost that complex queries spanning multiple tables cannot be expressed in the same way as in SQL.

DynamoDB is designed to deliver consistently low response times in the single-digit millisecond range, regardless of the volume of data. Whether a table contains a thousand or a billion entries makes no difference to the access speed.

When is a NoSQL database suitable?

DynamoDB really comes into its own in certain scenarios:

• High, fluctuating load: Applications with unpredictable peak loads, such as during campaigns or events.
• Simple access patterns: Data is predominantly read and written using a clear key.
• Very large data volumes: Tables that exceed the limits of a single relational database.
• Serverless architectures: Combined with AWS Lambda, this creates a backend entirely without managed servers.

For applications with complex, frequently changing queries across many linked entities, however, a relational database such as Amazon RDS is usually the better choice.

Capacity and scaling

DynamoDB offers two billing modes. In On-Demand mode, the database scales fully automatically and you pay per actual request – ideal for unknown or highly fluctuating loads. In Provisioned mode, you set an expected capacity, which is more cost-effective for consistent, predictable loads. You can switch between modes at any time.

Availability and data protection

DynamoDB automatically replicates data across multiple Availability Zones, making it inherently highly available. Data is encrypted at rest. Operated in an EU region, DynamoDB can be used in compliance with the GDPR. With Global Tables, data can also be replicated across multiple regions.

DynamoDB for SMEs

DynamoDB is not a general replacement for relational databases, but a specialised tool. For session storage, user profiles, shopping baskets, IoT data or event logs, the database is often the simpler and more cost-effective solution. It is crucial to think through the access patterns carefully in advance – because in DynamoDB, the data model is consistently determined by the type of queries.