Skip to main content

AWS S3 an Object Storage System

Amazon S3 is classified as object storage because of these key characteristics:

Flat Structure Instead of Hierarchical

  • Stores data as objects within buckets (no true folders or directories)
  • No nested file system like in block or file storage
  • "Folders" in S3 are actually prefix patterns in object keys

Objects as Complete Units

  • Each object includes the data, metadata, and a unique identifier (key)
  • Objects are stored and retrieved as complete, independent units
  • Cannot mount S3 as a drive or modify just parts of objects

Rich Metadata

  • Each object has system metadata (size, date, content type)
  • Supports custom metadata as key-value pairs
  • Metadata is stored with the object and retrieved together

HTTP/HTTPS Interface

  • Accessed via REST API over standard web protocols
  • Objects have unique URLs for direct access
  • Not accessed through file system commands

Scalability

  • Virtually unlimited storage capacity
  • Flat namespace allows for massive horizontal scaling
  • No volume or partition size limits to manage

Immutability

  • Objects are generally immutable (whole-object updates)
  • You replace rather than modify objects
  • Versioning preserves previous states of objects

These characteristics distinguish S3 from block storage (like EBS) which allows partial updates to storage volumes, and file storage (like EFS) which maintains a hierarchical structure with standard file system operations.