Amazon DocumentDB

Core Architecture

Amazon DocumentDB features a cloud-native architecture that decouples storage and compute, allowing each to scale independently. A DocumentDB cluster consists of a cluster volume and DB instances.

• Cluster Volume (Storage Layer): This is the database storage layer. It's a single, virtual volume that spans multiple Availability Zones (AZs). DocumentDB automatically grows the storage size as your data increases, up to 128 TiB.
• Instances (Compute Layer): These are the compute resources that provide the processing power to handle database read and write operations. You can have a primary instance and up to 15 replica instances.

Endpoints

DocumentDB provides different endpoints to connect to the appropriate instances for different use cases.

• Cluster Endpoint:

  • Always connects to the cluster's current primary instance.
  • This is the main endpoint used for both read and write operations (e.g., find, insert, update, delete).

• Reader Endpoint:

  • Load-balances read-only connections across all available replica instances in the cluster.
  • Use this endpoint for read-intensive workloads to distribute the load.

• Instance Endpoint:

  • Connects to a specific DB instance within the cluster.
  • Useful for specialized workloads or administrative tasks that need to target a single, specific instance (either primary or a particular replica).

Key Features

MongoDB Compatibility

• Amazon DocumentDB emulates the MongoDB 3.6, 4.0, or 5.0 APIs, allowing you to use most of the same drivers, applications, and tools you use with MongoDB.

Performance

• Designed to deliver high throughput and millisecond latency. According to AWS, it can achieve millions of requests per second and offers up to twice the throughput of self-managed MongoDB.

Scalability

• Storage Scaling: Storage starts at 10 GB and automatically scales up to 128 TiB in 10 GB increments with no impact on cluster performance.
• Compute Scaling: You can scale compute resources vertically by changing the instance class for the DB instances in your cluster.
• Read Scaling: You can horizontally scale read throughput by adding up to 15 low-latency read replicas. Replication lag is typically under 100 milliseconds.

High Availability & Reliability

• Multi-AZ Durability: The cluster volume maintains six copies of your data across three Availability Zones, providing high durability.
• Automatic Failover: DocumentDB supports automatic failover. If the primary instance fails, a replica is automatically promoted to primary. You can set a promotion priority tier for replicas to influence which one is chosen.
• Fast Recovery: In the event of a crash, database restart time is typically under a minute.
• Fault-Tolerant Replicas: Replicas can serve as failover targets with no data loss. For maximum availability, create replicas in multiple AZs.

Backup & Restore

• Automated Backups: Continuous, automated backups are always enabled by default with no performance impact.
• Point-In-Time Restoration (PITR): You can restore your cluster to any point in time within your backup retention period, down to a granularity of five minutes.
• Snapshots: You can create manual cluster snapshots for long-term archival. Encrypted manual snapshots can be shared with other AWS accounts, and you can copy snapshots across AWS Regions for disaster recovery.

Security

• Authentication:
  • Connections are authenticated using the standard MongoDB SCRAM (Salted Challenge Response Authentication Mechanism).
  • Management APIs are authenticated and authorized using AWS IAM users, roles, and policies.
• Encryption:
  • Encryption at Rest: Data is encrypted at rest by default using keys you manage through the AWS Key Management Service (KMS).
  • Encryption in Transit: Data is encrypted in transit using Transport Layer Security (TLS).
• Network Isolation:
  • DocumentDB clusters are deployed within an Amazon VPC, allowing you to isolate them in your own private network.
• Role-Based Access Control (RBAC):
  • DocumentDB supports RBAC using MongoDB-compatible, built-in roles to enforce the principle of least privilege for database users.

Pricing

You are billed based on four main components:

1. On-Demand DB Instances: Priced per second of usage (with a 10-minute minimum).
2. Database Storage: Billed per GB-month.
3. I/O Operations: Billed per million requests.
4. Backup Storage: Billed per GB-month for storage exceeding your provisioned database storage size.

Limitations

• While highly compatible, DocumentDB does not support every MongoDB API and aggregation pipeline stage. Always verify that your specific application's required operations are supported.
• Server-side Javascript execution (e.g., db.eval(), $where operator) is not supported for security reasons.
  Sources