Why DynamoDB

Hub › AWS › Advanced › Why DynamoDB

Goal

After this page you will understand why a Lambda function returning hardcoded JSON is not enough for a real application, and why DynamoDB is a natural next step.

Prerequisites

• AWS Intermediate tier — you have a Lambda function behind API Gateway serving JSON from memory.

The problem with in-memory data

The intermediate tier Lambda returns a hardcoded array:

const items = [
  { id: 1, name: "Notebook" },
  { id: 2, name: "Pen" },
  { id: 3, name: "Sticky notes" },
];

Every invocation returns the same data. Nothing persists between calls, and no other client or Lambda can write to it. This is fine for a demo, but any real application needs:

• Durability — data survives Lambda cold starts and redeploys.
• Shared state — multiple Lambda invocations read and write the same data.
• Queryability — filter, sort, and paginate through records.

The three database options on AWS

Service	Best for	Managed?
DynamoDB	Key-value + document, single-digit-millisecond, serverless	Yes
RDS (Aurora)	Relational, joins, complex queries	Yes (provisioned or serverless)
MemoryDB / ElastiCache	Sub-millisecond cache, session store	Yes

DynamoDB wins for this stack because:

• Fully serverless — no servers to provision, scales to zero.
• Pay-per-request — no charge when idle.
• Single-digit-millisecond latency at any scale.
• Tight AWS integration — the AWS SDK DynamoDB client is available in every Lambda runtime.

What this tier adds

Seven new pieces compared to the intermediate tier:

• DynamoDB table — persistent item storage.
• Cognito User Pool — user sign-up, sign-in, JWT tokens.
• Lambda authorizer — protects the API with Cognito JWTs.
• CDK — Infrastructure as Code (TypeScript) for repeatable deployments.
• SAM — local Lambda development and testing.
• CodePipeline — CI/CD from GitHub to deploy.
• CloudWatch alarms — monitor errors and latency.
• Production hardening — WAF, rate limiting, multi-region.

You build each piece in order across the next nine pages.

Next → Create a DynamoDB table