DynamoDB is Amazon's managed NoSQL database for workloads that need predictable key-value access at very large scale. During Prime Day 2023, DynamoDB handled more than 126 million requests per second at peak, which gives a useful sense of the scale it is built to support.

DynamoDB achieves this by making a clear trade-off: it gives up relational query flexibility in exchange for predictable performance on well-modeled access patterns.

The starting question is not "What entities exist?" but "How will the application read and write this data?" That shift is the heart of DynamoDB design.

Choose DynamoDB for the wrong workload, and the model will feel awkward quickly. Model your keys poorly, and a small number of hot partition keys can throttle important paths even when the table looks underused overall.

For interviews, the main skill is not listing DynamoDB features. It is showing that the workload has known access patterns, explaining the key design, and calling out the limits around hot keys, GSIs, transactions, TTL, and DAX.