Amazon DynamoDB

Amazon DynamoDB is a serverless, fully managed, distributed NoSQL database designed for single-digit millisecond performance at any scale. It requires no database servers to provision or patch, and tables can scale automatically, including to zero in on-demand mode. Data access uses primary keys, with global and local secondary indexes for queries using alternate keys. Global tables support multi-Region, multi-active operation; AWS says they are designed for up to 99.999% availability. DynamoDB also supports native vector search for similarity searches and agent memory use cases. Data is encrypted in transit and at rest by default, and IAM policies can limit access down to specific items and attributes. Point-in-time recovery provides continuous backups with per-second granularity and restoration within a configured period of 1 to 35 days. The free tier uses provisioned capacity and the Standard table class; its request volume is described as about 200 million requests per month depending on item size. On-demand usage is billed per request consumed, while provisioned capacity is billed even if the application uses less.

DynamoDB combines managed NoSQL storage with automatic scaling, global tables, and fine-grained access controls. The choice between on-demand and provisioned capacity matters because their billing bases differ.

DynamoDB suits teams building cloud applications that need managed NoSQL storage, automatic scaling, or multi-Region operation. Its listed use cases include mission-critical, internet-scale, agentic, modernization, and Internet of Things applications.

Overview

Amazon DynamoDB is a managed NoSQL database for teams building cloud applications that need to scale without operating database servers. It suits workloads from internet-scale services to IoT and agentic applications, but the choice between request-based and capacity-based billing deserves close attention.

Its data model centers on primary keys, with global and local secondary indexes for queries through alternate keys. Data persists on disk; each value can be up to 0.4 MB. HTTP(S), JSON and PartiQL are supported, along with time-to-live (TTL). These are useful building blocks, though the value-size ceiling makes DynamoDB a poor fit for storing larger individual values.

Key features

Managed scaling and access

AWS manages the servers, so there is no provisioning or patching to handle. Tables can scale automatically, and on-demand mode can scale to zero. That makes on-demand attractive when demand is hard to predict; provisioned capacity is more deliberate, but charges apply to the capacity specified even when the application uses less.

Primary keys suit direct access patterns, while secondary indexes offer alternate query paths. Native vector search supports similarity-search and agent-memory use cases, extending the service beyond conventional key-value access.

Global operation and protection

Global tables support multi-Region, multi-active operation and are designed by AWS for up to 99.999% availability. That is a strong fit for applications that need to operate across Regions, but it is an AWS design target rather than a guarantee stated here.

Data is encrypted in transit and at rest by default. IAM policies can restrict access down to specific items and attributes, and AWS lists KMS encryption, IAM controls and VPC endpoints with PrivateLink among its integrations. Point-in-time recovery provides continuous, per-second backups and restoration within a configured window of 1 to 35 days. AWS also describes its security controls and compliance standards as addressing requirements of governments, global banks and other high-sensitivity organizations.

Pricing

DynamoDB uses a freemium model, with two free-tier listings and paid capacity choices whose prices are custom pricing.

A free trial is also offered. Free capacity is limited by the listed quotas, and neither paid capacity option has a published price here; estimate costs against request volume or specified capacity before choosing.

Platforms

DynamoDB is a cloud deployment available across Android, iOS, Linux, macOS, Windows and web, with API access. Its broad client reach does not change the central requirement: the database itself is deployed in the cloud.

Who it's for

DynamoDB is a strong candidate for teams building mission-critical, internet-scale, modernization, IoT or agentic applications that benefit from managed scaling, multi-Region operation and granular IAM controls. It is less suitable when workloads need values larger than 0.4 MB or when predictable spending is difficult to achieve under request-based billing or provisioned-capacity commitments.

Pros and cons

Pros

• Managed servers and automatic scaling, including scaling to zero in on-demand mode, reduce infrastructure work.
• Multi-active global tables and fine-grained IAM controls support applications with cross-Region and sensitive-data requirements.
• Per-second point-in-time recovery with a configurable 1-to-35-day window adds a defined recovery option.
• Native vector search supports similarity-search and agent-memory workloads alongside indexed key access.

Cons

• Provisioned capacity is billed whether or not all specified capacity is consumed, so poor estimates can waste spend.
• On-demand billing tracks each request, which can make costs harder to anticipate when traffic varies.
• The 0.4 MB maximum value size rules out larger individual values.
• The free tier is provisioned and uses the Standard table class, so it does not represent every capacity choice.

Alternatives

RocksDB is worth considering when a free, open-source C++ library with GPLv2 or Apache 2.0 dual licensing better fits the project.

Google Cloud Bigtable is an alternative to assess for teams considering its Standard and Premium Bigtable editions or Data Boost.

Upstash Redis may suit readers looking for a freemium Redis option with a free tier that includes 256 MB of data, 10 GB monthly bandwidth and 500K monthly commands.

Aerospike Database is another freemium database option.

Riak KV is a free, Apache 2-licensed option with a key-value data module, built-in masterless replication and HTTP API and Protocol Buffers support.

VMware Workstation Pro is also listed as a free alternative.

etcd is a free, Apache 2.0-licensed self-hosted option.

Infinispan offers a free Apache 2.0-licensed community edition as an embedded library or standalone server.

Browse Key-Value Databases, Multimodel Databases or Database Software for more options.

Verdict

Choose DynamoDB for cloud applications that need managed scaling, global tables and tightly scoped access controls without running database servers. Its strongest case is operational reach; look elsewhere if larger values are essential or if capacity-based billing does not suit your workload economics.