Dynamodb

Introduction

• Fully managed, highly available with replication across AZs
• NoSQL Database
• Scales to massive workloads, highly distributed database
• Fully integrates with IAM
• Enable event driven programming with DynamoDB streams

Basics

• DynamoDB is made of tables
• Each table will have primary key (must be decided at creation time)
• Each table can have an infinite number of rows (rows are items)
  • Each row can have a max of 400KB of data

Primary keys

• how to choose a primary key?
• Two types of primary keys
  • partition keys
    • Partition key must be unique for each item
    • Partition key must be ‘diverse’ so that data is distributed
  • partition key + sort key
    • data is grouped by partition key
    • each combination must be unique per item

Read/Write Capacity Modes

• Provisioned Mode
  • Provision all capacity in advance by providing values for RCU and WCU
  • Pay up front
  • Throughput can temporarily be exceeded using a burst capacity
  • If you exhaust the burst capacity, you can retry using exponential backoff
• On-demand Mode
  • Reads/writes scale on-demand
  • Pay for what you use
• You can switch between the two modes once every 24 hours

Local Secondary Index

• Alternate sort key for your table
• Must be defined at table creation time
• Up to 5 local secondary indexes per table
• The sort key consists of one scaler attribute (string, number, or binary)

Global Secondary Index

• Alternative primary key from the base table

PartiQL

• Use SQL-like syntax to query DynamoDB tables

Optimistic Locking

• DynamoDB has a feature called “conditional writes” to ensure an item has not changed before writing to it
• Each item has an attribute that acts as a version number
• Useful for when you have multiple writers attempters to write to an item

DynamoDB Accelerator (DAX)

• Fully managed, highly available, seamless in-memory cache for DynamoDB
• Microseconds latency for cached reads and queries
• Does not require that you change any application code
• Solves the “hot key” problem. If you read a specific key (item) too many times, you may get throttled.
• 5 minutes TTL for cache (default)
• Up to 10 DAX nodes per cluster
• Multi-AZ (3 nodes minimum recommended for production)
• Secure (Encryption in transit and at rest)

DynamoDB Streams

• Streams are an ordered item-level modifications (create/update/delete) in a table
• Stream records can be:
  • Sent to Kinesis Data Streams
  • Read by AWS Lambda
  • Read by Kinesis Client Library Apps
• Data retention for up to 24 hours
• Use cases:
  • react to changes in real-time
  • Analytics
  • Insert into derivative tables
  • Insert into OpenSearch service