Object Storage

Object storage is the storage model behind services like Amazon S3, Google Cloud Storage, Azure Blob Storage, and MinIO. It is built for large amounts of unstructured data: images, videos, backups, logs, exports, ML datasets, and data lake files.

The basic idea is simple. Instead of exposing a disk or a mounted file system, object storage exposes an API. You store an object with a key, and later retrieve it by that key.

That design sounds modest, but it changes the trade-offs completely. Object storage gives you massive scale, strong durability, HTTP access, lifecycle policies, and low storage cost. In exchange, you give up normal file-system behavior such as in-place writes, file locks, and POSIX-style operations.

This chapter explains how object storage works, where it fits in system design, and where it should not be used.