This section covers Amazon S3, the object store at the centre of most AWS data architectures, and S3 Transfer Acceleration, which speeds up long-distance uploads and downloads. The overview below places S3 among the other AWS storage services so it is clear when object storage is the right choice.
Object storage and the transfer behaviour that determines how fast data gets in and out.
S3 stores objects (bytes plus metadata) under keys in buckets that live in one region and are replicated across availability zones. Analytics services read the data in place, which is what makes S3 the storage layer of an AWS data lake. Transfer Acceleration routes traffic from clients far from the bucket's region into the nearest CloudFront edge location and then over the AWS backbone, which can help large transfers over long distances; for clients in the same region it adds cost without benefit.
| Service | Type | Use when |
|---|---|---|
| Amazon S3 | Object storage, HTTP API | Data lakes, backups, static assets, any scale |
| Amazon EBS | Block volume for one EC2 instance | Databases and boot volumes needing low latency |
| Amazon EFS | Shared NFS file system | Many instances need a POSIX file system |
| S3 Glacier classes | Archive tiers of S3 | Rarely accessed data, retrieval delay acceptable |
Within S3, storage classes (Standard, Intelligent-Tiering, Standard-IA, Glacier tiers) trade storage price against retrieval cost and latency. Lifecycle rules can move objects between classes automatically as they age.