Storage
Three kinds of storage: object (S3), block (EBS and instance store) and file (EFS and FSx). Knowing which to use in each scenario is the core of this topic.
S3
Simple Storage Service
Object storage designed for 99.999999999% (11 nines) durability. Objects live in buckets and are identified by an object key.
Key points
- Bucket names are globally unique, 3 to 63 characters long, using only lowercase letters, numbers, dots and hyphens.
- Versioning keeps previous versions of objects (
version ID). - Lifecycle policies move objects between classes or delete them automatically.
- Uses: backups, data lakes, static websites, media files.
On the exam
Rarely accessed data that still needs immediate retrieval goes to Standard-IA or Glacier Instant Retrieval. Long-term archive, Glacier Deep Archive.
Storage classes
| Class | When to use |
|---|---|
| Standard | Frequently accessed data. No minimum duration. |
| Intelligent-Tiering | Unknown or changing access patterns: moves objects between tiers automatically. |
| Express One Zone | Hot, latency-sensitive data in a single AZ. |
| Standard-IA | Infrequent access with fast retrieval. 30-day minimum. |
| One Zone-IA | Infrequent access, re-creatable data. One AZ, 30-day minimum. |
| Glacier Instant Retrieval | Archive accessed a few times a year, in milliseconds. 90-day minimum. |
| Glacier Flexible Retrieval | Backups and archives retrieved in minutes or hours. 90-day minimum. |
| Glacier Deep Archive | Long-term retention and compliance, retrieval within 12 hours. 180-day minimum. |
EBS
Elastic Block Store
Persistent block storage volumes attached to EC2 instances within one Availability Zone.
Key points
- The volume exists independently of the instance and lives in the same AZ.
- Snapshots are incremental backups stored in S3.
- Encryption at rest with AWS KMS keys.
- Instance store is temporary block storage: its data is lost when the instance stops or terminates.
On the exam
A persistent disk for an EC2 instance: EBS. A temporary, very fast disk: instance store.
EFS
Elastic File System
A managed, elastic NFS file system that grows and shrinks as you add and remove files.
Key points
- Can be mounted by many instances at once.
- Data replicated across multiple AZs in the Region.
- For Linux workloads, with no capacity to provision.
On the exam
Many Linux instances reading and writing the same files: EFS.
FSx
Amazon FSx
Fully managed third-party file systems.
Key points
- FSx for Windows File Server: native Windows SMB file shares.
- FSx for Lustre: high performance for HPC and machine learning.
- Options also exist for NetApp ONTAP and OpenZFS.
On the exam
Windows file shares (SMB, Active Directory): FSx for Windows File Server.
Storage Gateway
AWS Storage Gateway
A hybrid service that gives on-premises applications access to cloud storage, with a local cache.
Key points
- Types: S3 File Gateway, Volume Gateway and Tape Gateway (virtual tapes).
- Keeps the most used data in a local cache.
- Used for backup, archiving and migration.
On the exam
On-premises applications using AWS storage as if it were local: Storage Gateway.
Backup
AWS Backup
Centralizes and automates backups across AWS services and hybrid environments.
Key points
- Backup plans with schedules and retention.
- Covers EBS, EC2, RDS, DynamoDB, EFS, FSx, S3 and more.
- Cross-Region and cross-account copies.
On the exam
"Manage backups of many services in one place": AWS Backup.
Elastic Disaster Recovery
AWS DRS
Continuously replicates on-premises or cloud servers to AWS and recovers them within minutes after a disaster.
Key points
- Continuous block-level replication, with an RPO of seconds.
- Recovery (RTO) in minutes, without keeping idle servers.
- Lets you run recovery drills without affecting production.
On the exam
"Recover servers quickly on AWS with minimal data loss": Elastic Disaster Recovery.