Best Practices and Advanced Features

This final lesson covers DynamoDB's advanced features — Time to Live (TTL), DynamoDB Accelerator (DAX), Global Tables, and PartiQL — along with a comprehensive set of best practices for production workloads.

---

Time to Live (TTL)

TTL automatically deletes expired items from a table at no cost (no WCU consumed):

How TTL Works

1. You designate an attribute as the TTL attribute (e.g., ExpiresAt)
2. The attribute value must be a Unix epoch timestamp (Number type)
3. DynamoDB periodically scans for items where the TTL value is in the past
4. Expired items are deleted (typically within 48 hours of expiration)
5. Deletions are recorded in DynamoDB Streams (as REMOVE events with userIdentity.type = "Service")

Enabling TTL

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

Writing an Item with TTL

# Set expiry to 24 hours from now (Unix timestamp)
aws dynamodb put-item \
  --table-name Sessions \
  --item '{
    "SessionId": {"S": "sess-abc123"},
    "UserId": {"S": "user-001"},
    "ExpiresAt": {"N": "1709654400"}
  }'

Common TTL Use Cases

Use Case	TTL Attribute Value
Session management	Session creation time + session duration
Cache entries	Current time + cache TTL
Temporary tokens	Token issue time + validity period
Regulatory data retention	Data creation time + retention period
Event logs	Event time + retention window

Tip: Expired items may still appear in reads for up to 48 hours after expiration. Filter them out in your application by checking the TTL attribute.

---

DynamoDB Accelerator (DAX)

DAX is a fully managed, in-memory cache purpose-built for DynamoDB:

How DAX Works

Application → DAX Cluster → DynamoDB Table
                │
        In-Memory Cache
        (microsecond reads)

DAX is a write-through cache — reads check the cache first, and writes update both the cache and the table.

Key Features

Feature	Description
Latency	Microsecond response times (vs millisecond for DynamoDB)
Compatibility	Drop-in replacement — same API as DynamoDB
Managed	Fully managed — no cache invalidation logic needed
Write-through	Writes go to DynamoDB and update the cache
Cluster	Multi-node cluster with read replicas
Encryption	Supports encryption at rest and in transit

DAX Cluster Setup

aws dax create-cluster \
  --cluster-name my-dax-cluster \
  --node-type dax.r5.large \
  --replication-factor 3 \
  --iam-role-arn arn:aws:iam::123456789012:role/DAXRole \
  --subnet-group-name my-subnet-group

When to Use DAX

Good Fit	Poor Fit
Read-heavy workloads	Write-heavy workloads
Repeated reads of the same items	Mostly unique reads
Latency-sensitive applications	Applications tolerant of millisecond latency
Eventually consistent reads	Strongly consistent reads (DAX does not cache these)

DAX vs ElastiCache

Feature	DAX	ElastiCache (Redis/Memcached)
Purpose	DynamoDB-specific cache	General-purpose cache
API	DynamoDB-compatible	Redis/Memcached protocol
Cache invalidation	Automatic (write-through)	Manual
Data model	Items and queries	Key-value, data structures
Consistency	Eventually consistent	Application-managed

---

Global Tables

Global Tables provide multi-Region, multi-active replication for DynamoDB:

How Global Tables Work

                    Writes
                      │
        ┌─────────────┼─────────────┐
        ▼             ▼             ▼
   eu-west-1     us-east-1     ap-southeast-1
   ┌──────────┐  ┌──────────┐  ┌──────────┐
   │ Replica  │  │ Replica  │  │ Replica  │
   │ Table    │←→│ Table    │←→│ Table    │
   └──────────┘  └──────────┘  └──────────┘
        Active       Active       Active

• All replicas accept reads and writes (multi-active)
• Changes are replicated to all Regions, typically within one second
• Conflict resolution uses last writer wins based on timestamps

Creating a Global Table

# First, create the table in the primary Region
aws dynamodb create-table \
  --table-name GlobalUsers \
  --billing-mode PAY_PER_REQUEST \
  --attribute-definitions AttributeName=UserId,AttributeType=S \
  --key-schema AttributeName=UserId,KeyType=HASH \
  --stream-specification StreamEnabled=true,StreamViewType=NEW_AND_OLD_IMAGES

# Then add replicas
aws dynamodb update-table \
  --table-name GlobalUsers \
  --replica-updates '[
    {"Create": {"RegionName": "us-east-1"}},
    {"Create": {"RegionName": "ap-southeast-1"}}
  ]'

Global Table Features