AWS Knowledge Hub

English

Tiếng Việt

English

Tiếng Việt

Appearance

Database Service Comparison ​

AWS offers multiple database services, each optimized for different use cases. This guide helps you choose the right database service for your workload.

Database Categories ​

Relational Databases ​

Services:

  • Amazon RDS (MySQL, PostgreSQL, MariaDB, Oracle, SQL Server)
  • Amazon Aurora (MySQL and PostgreSQL compatible)

Characteristics:

  • ACID compliance
  • SQL queries
  • Structured data
  • Vertical and horizontal scaling

NoSQL Databases ​

Services:

  • Amazon DynamoDB
  • Amazon DocumentDB (MongoDB compatible)

Characteristics:

  • Flexible schema
  • High performance
  • Horizontal scaling
  • Eventual consistency (some)

In-Memory Databases ​

Services:

  • Amazon ElastiCache (Redis, Memcached)

Characteristics:

  • Very fast (in-memory)
  • Caching use cases
  • Session storage
  • Real-time analytics

Service Comparison ​

Amazon RDS ​

Best For:

  • Traditional relational databases
  • Applications requiring SQL
  • Existing database migrations
  • Multiple database engine options

Features:

  • Managed service
  • Automated backups
  • Multi-AZ deployments
  • Read replicas (up to 5)
  • Automated patching

Limitations:

  • Vertical scaling (instance size)
  • Limited horizontal scaling
  • Storage limits (up to 64 TB)
  • Manual scaling

Use Cases:

  • Web applications
  • Enterprise applications
  • Content management systems
  • E-commerce platforms

Amazon Aurora ​

Best For:

  • High-performance applications
  • MySQL/PostgreSQL workloads
  • Applications needing high availability
  • Large-scale applications

Features:

  • Up to 5x MySQL, 3x PostgreSQL performance
  • Auto-scaling storage (10 GB to 128 TB)
  • Up to 15 read replicas
  • Multi-Master support
  • Serverless option
  • Global Database

Advantages over RDS:

  • Better performance
  • More read replicas
  • Auto-scaling storage
  • Faster replication
  • Backtrack feature

Use Cases:

  • High-traffic web applications
  • SaaS applications
  • Enterprise databases
  • Global applications

Amazon DynamoDB ​

Best For:

  • Serverless applications
  • High-traffic applications
  • NoSQL workloads
  • Key-value and document data

Features:

  • Fully managed
  • Serverless
  • Automatic scaling
  • Single-digit millisecond latency
  • Global Tables
  • On-demand or provisioned capacity

Data Models:

  • Key-value
  • Document (JSON)

Use Cases:

  • Web applications
  • Gaming applications
  • IoT applications
  • Mobile backends
  • Real-time applications

Amazon DocumentDB ​

Best For:

  • MongoDB workloads
  • Document databases
  • Existing MongoDB applications

Features:

  • MongoDB compatible (3.6, 4.0, 5.0)
  • Managed service
  • High performance (2x MongoDB)
  • Auto-scaling storage (10 GB to 64 TB)
  • Up to 15 read replicas

Use Cases:

  • Content management
  • User profiles
  • Catalogs
  • IoT applications

Amazon ElastiCache ​

Best For:

  • Caching
  • Session storage
  • Real-time analytics
  • High-performance reads

Engines:

  • Redis: Advanced data structures, persistence
  • Memcached: Simple key-value, multi-threaded

Use Cases:

  • Application caching
  • Session stores
  • Real-time leaderboards
  • Gaming applications

Decision Matrix ​

Choose RDS When ​

  • Need traditional SQL database
  • Want multiple engine options
  • Have existing database to migrate
  • Need specific database features
  • Cost optimization for smaller workloads

Choose Aurora When ​

  • Need high performance
  • Want auto-scaling storage
  • Need many read replicas
  • Want Global Database
  • MySQL/PostgreSQL compatible

Choose DynamoDB When ​

  • Building serverless applications
  • Need automatic scaling
  • Want single-digit millisecond latency
  • Have key-value or document data
  • Need global replication

Choose DocumentDB When ​

  • Using MongoDB
  • Need MongoDB compatibility
  • Want managed MongoDB service
  • Need high performance
  • Have document data model

Choose ElastiCache When ​

  • Need caching layer
  • Want in-memory performance
  • Need session storage
  • Want real-time analytics
  • Need high-performance reads

Performance Comparison ​

Throughput ​

  • DynamoDB: Millions of requests per second
  • Aurora: High throughput, scales with replicas
  • RDS: Depends on instance size
  • ElastiCache: Very high (in-memory)

Latency ​

  • ElastiCache: Sub-millisecond (in-memory)
  • DynamoDB: Single-digit milliseconds
  • Aurora: Low latency, < 100ms replication
  • RDS: Depends on instance and workload

Scalability ​

  • DynamoDB: Automatic, unlimited
  • Aurora: Auto-scaling storage, many replicas
  • RDS: Vertical scaling, limited horizontal
  • ElastiCache: Cluster mode for scale

Cost Considerations ​

Pricing Models ​

RDS:

  • Instance-based pricing
  • Storage and IOPS costs
  • Reserved Instances available

Aurora:

  • Instance-based pricing
  • Storage costs (auto-scaling)
  • Serverless option (pay per ACU)

DynamoDB:

  • On-demand: Pay per request
  • Provisioned: Pay for capacity units
  • Storage costs

DocumentDB:

  • Instance-based pricing
  • Storage costs
  • Similar to RDS pricing

ElastiCache:

  • Instance-based pricing
  • Reserved Instances available

Cost Optimization ​

  • Use Reserved Instances for predictable workloads
  • Right-size instances
  • Use appropriate storage types
  • Monitor and optimize usage
  • Delete unused resources

High Availability ​

Multi-AZ Deployments ​

  • RDS: Multi-AZ with automatic failover
  • Aurora: Multi-AZ with 6 copies across 3 AZs
  • DynamoDB: Built-in multi-AZ
  • DocumentDB: Multi-AZ support
  • ElastiCache: Multi-AZ clusters

Backup and Recovery ​

  • RDS: Automated backups, point-in-time recovery
  • Aurora: Continuous backup, backtrack
  • DynamoDB: Point-in-time recovery
  • DocumentDB: Automated backups
  • ElastiCache: Snapshot support (Redis)

Security ​

Encryption ​

  • All services support encryption at rest
  • All services support encryption in transit
  • KMS integration for key management
  • VPC isolation

Access Control ​

  • IAM integration
  • VPC security groups
  • Network isolation
  • Database authentication

Migration Considerations ​

From On-Premises ​

  • Use AWS DMS for migration
  • Minimal downtime options
  • Schema conversion support
  • Continuous replication

Between AWS Services ​

  • Export/import tools
  • DMS for RDS to Aurora
  • Native tools for some services
  • Consider downtime vs. continuous replication

Best Practices ​

  1. Choose Right Service: Match service to use case
  2. Design for Scale: Plan for growth
  3. Implement HA: Multi-AZ deployments
  4. Enable Backups: Automated backup strategy
  5. Monitor Performance: Use CloudWatch
  6. Optimize Costs: Right-size and use Reserved Instances
  7. Secure Data: Encryption and access control
  8. Plan Migrations: Test and validate