AWS storage can be grouped by access model: block storage, object storage, and file storage.
Block Storage
Block storage behaves like a disk and suits frequent low-latency updates.
EC2 Instance Store is temporary local storage. It is fast, but data may be lost when the instance stops or terminates.
EBS is persistent block storage, usually attached to one EC2 instance.
EBS snapshots are point-in-time backups:
- Incremental data block backup
- Stored in S3
- Useful for recovery, migration, resizing, and cloning
Object Storage: S3
S3 stores objects. Each object includes data, metadata, and a unique key. S3 does not have real folders; a/b/c.png is just a key that looks like a path.
S3 fits:
- Images and static assets
- Logs
- Backups
- Data lakes
- Large-file archives
S3 Storage Classes
S3 Standard: default, multi-AZ, high availability, frequent access.
S3 Express One Zone: single AZ, low latency, good for reconstructable high-speed data.
S3 Standard-IA: infrequent access with seconds-level retrieval, multi-AZ.
S3 One Zone-IA: cheaper, single AZ, for reconstructable data.
S3 Intelligent-Tiering: automatically moves objects based on access patterns.
S3 Glacier Instant Retrieval: rare access but millisecond retrieval.
S3 Glacier Flexible Retrieval: minutes-to-hours retrieval, good for backup and disaster recovery.
S3 Glacier Deep Archive: lowest cost, usually 12+ hours retrieval, good for long-term compliance retention.
Lifecycle Policies
Lifecycle policies automatically transition storage classes or expire objects.
Example:
- Move to Standard-IA after 30 days
- Move to Glacier after 180 days
- Delete after 7 years
File Storage
EFS is a managed NFS file system that multiple EC2 instances can mount. It scales automatically.
FSx provides managed file systems:
- FSx for Windows
- FSx for Lustre
- FSx for ONTAP
- FSx for OpenZFS
Storage Gateway
Storage Gateway is a hybrid cloud storage service that lets on-premises systems access cloud storage. It is common for backup, archive, disaster recovery, and local cache use cases.