Amazon DynamoDB

Amazon DynamoDB is a serverless NoSQL database from Amazon Web Services for key-value and document data, billed per request with no servers to size.

DynamoDB stores application data as items in tables and returns it with single-digit millisecond latency for tens of thousands or hundreds of millions of simultaneous users. Amazon Web Services builds and runs it as a fully managed service inside the AWS cloud. A table needs only a primary key, so items in one table can carry different attributes with no schema migration. The service manages partitions itself and splits them as a table grows. Teams have no servers to patch, no upgrades to schedule and no maintenance windows to plan around. Global tables replicate a table to other AWS Regions and accept writes in every replica. Conflicts settle by last writer wins, or a table can run in multi-Region strong consistency mode instead. Transactions group several item writes into one all-or-nothing operation, and table reads choose eventual or strong consistency per request. DynamoDB Streams records every item change in order, and Time to Live deletes items once their own expiry timestamp passes. Point-in-time recovery keeps continuous backups restorable to the second, while vector search answers similarity queries over embeddings stored beside the data. DynamoDB Accelerator (DAX) is a DynamoDB-compatible caching service that runs alongside it. The service is built for developers and platform teams whose applications already run on AWS and must stay fast as demand swings. Financial services teams use it for payments and core banking work that depends on ACID transactions and multi-Region resilience. Gaming, retail and e-commerce, adtech and software companies run internet-scale and mission-critical applications on it. Teams building agentic applications use vector search to give their agents semantic retrieval over their own records. Security teams scope access per table, item and attribute through IAM policies, and can encrypt chosen attributes in the client with the AWS Database Encryption SDK. A common serverless pattern connects DynamoDB Streams to AWS Lambda, so a function runs each time an item changes or expires. In banking it carries core ledgers, fraud detection, messaging workflows and the onboarding of digital users. Analysts export tables to Amazon S3, in full at a point in time or incrementally, and change data capture can feed Amazon Kinesis Data Streams or AWS Glue. It also holds device data for Internet of Things fleets, and it takes over from older databases when a team modernizes an application.