AWS DynamoDB

Amazon DynamoDB is a fully managed NoSQL database service that provides fast and predictable performance with seamless scalability.

Key Benefits

DynamoDB offers single-digit millisecond latency at any scale, with built-in security, backup, and in-memory caching.

Key Concepts

Tables

The primary resource in DynamoDB. Unlike relational databases, tables don't have a fixed schema.

Items

Individual records in a table. Each item is a collection of attributes.

Attributes

Data elements within items. Similar to columns in relational databases.

Data Access Patterns

Primary Key

Every table requires a primary key:

Type	Description
Partition Key	Single attribute that uniquely identifies an item
Composite Key	Partition key + sort key for more complex queries

Choose your partition key carefully! A good partition key distributes data evenly and supports your query patterns.

Data Types

Type	Description
S	String
N	Number
B	Binary
BOOL	Boolean
NULL	Null

Type	Description
M	Map (nested attributes, like JSON objects)
L	List (ordered collection, like arrays)

Type	Description
SS	String Set (unique strings)
NS	Number Set (unique numbers)
BS	Binary Set (unique binary values)

Table Design

Example: E-commerce Orders

create-table.json

{
  "TableName": "Orders",
  "KeySchema": [
    {"AttributeName": "customerId", "KeyType": "HASH"},
    {"AttributeName": "orderId", "KeyType": "RANGE"}
  ],
  "AttributeDefinitions": [
    {"AttributeName": "customerId", "AttributeType": "S"},
    {"AttributeName": "orderId", "AttributeType": "S"}
  ],
  "BillingMode": "PAY_PER_REQUEST"
}

Item Example

order-item.json

{
  "customerId": {"S": "CUST-001"},
  "orderId": {"S": "ORD-2024-001"},
  "orderDate": {"S": "2024-01-15"},
  "total": {"N": "99.99"},
  "items": {"L": [
    {"M": {"productId": {"S": "PROD-1"}, "quantity": {"N": "2"}}}
  ]},
  "status": {"S": "SHIPPED"}
}

Secondary Indexes

Global Secondary Index (GSI)

Query data using different partition and sort keys.

gsi-definition.json

{
  "IndexName": "StatusIndex",
  "KeySchema": [
    {"AttributeName": "status", "KeyType": "HASH"},
    {"AttributeName": "orderDate", "KeyType": "RANGE"}
  ],
  "Projection": {"ProjectionType": "ALL"}
}

GSIs have their own provisioned capacity and can be created/deleted at any time.

Local Secondary Index (LSI)

Same partition key but different sort key. Must be created with the table.

LSIs share throughput with the base table and cannot be created after table creation.

Capacity Modes

Pay per request pricing
Automatic scaling for spiky workloads
No capacity planning needed

Best for:

New applications with unknown workloads
Unpredictable traffic patterns
Preferring pay-per-use pricing

Specify read/write capacity units (RCU/WCU)
More cost-effective for predictable workloads
Auto-scaling available

Set Provisioned Capacity

aws dynamodb update-table \
  --table-name Orders \
  --provisioned-throughput ReadCapacityUnits=100,WriteCapacityUnits=50

Read Consistency

Type	Description
Eventually Consistent	Default, lower latency, may return stale data
Strongly Consistent	Always returns latest data, higher latency

Query vs Scan

Requires partition key
Efficient and fast
Uses indexes

Always use Query when possible!

Reads every item in table
Expensive and slow
Consumes lots of RCUs

Warning

Avoid Scan operations in production. Design your data model to use Query operations instead.

Streams

Capture item-level changes for real-time processing:

Enable Streams

aws dynamodb update-table \
  --table-name Orders \
  --stream-specification StreamEnabled=true,StreamViewType=NEW_AND_OLD_IMAGES

Stream View Types:

KEYS_ONLY - Only the key attributes
NEW_IMAGE - The item after modification
OLD_IMAGE - The item before modification
NEW_AND_OLD_IMAGES - Both versions

Use DynamoDB Streams with Lambda to build real-time applications and event-driven architectures.

TTL (Time to Live)

Automatically delete expired items:

Enable TTL

aws dynamodb update-time-to-live \
  --table-name Orders \
  --time-to-live-specification Enabled=true,AttributeName=expirationTime

TTL deletions don't consume WCUs and are eventually consistent (may take up to 48 hours).

Best Practices

Data Modeling

Start with access patterns, not data structure
Use composite keys for related data
Denormalize when appropriate
Use GSIs for alternative access patterns

Performance

Distribute workload evenly across partitions
Use Query instead of Scan
Enable DAX for read-heavy workloads
Use batch operations for multiple items

Cost Optimization

Use on-demand for variable workloads
Enable TTL to delete old data
Project only needed attributes
Use reserved capacity for predictable workloads

DAX (DynamoDB Accelerator)

In-memory cache for DynamoDB:

Microsecond latency for reads
No application changes needed (API compatible)
Write-through caching
Ideal for read-heavy workloads

Next Steps

DynamoDB CLI Reference

Complete CLI commands for DynamoDB

RDS

Relational Database Service

DynamoDB CLI Reference

RDS

On this page