Object storage

Object storage holds data as objects, a file plus its metadata and a unique key, inside containers called buckets, accessed over HTTP APIs rather than mounted as a drive. It’s cheap per gigabyte, scales to effectively unlimited size, and is engineered for extreme durability. AWS S3, Azure Blob Storage, and Google Cloud Storage are the canonical examples.

In plain terms

It’s an infinite coat check. You hand over a file and get back a key (a URL); present the key later and you get the file. No drives to size, no volumes to fill, no folders in the real sense, just keys and objects. That makes it the home for things you store and retrieve rather than query: user uploads, backups, exports, video, log archives, a static website’s files. It is not a database and not a disk your app can treat like C:\, reads and writes go through an API.

Why it matters when you migrate

• It’s usually the cheapest, safest first move. Backups, archives, and media can land in object storage on day one, low risk, immediate value, and it shrinks the on-premise footprint before anything hard begins.
• Apps that write to local disk need a code change. “Save to a folder” becomes “put to a bucket”, a small change, but one to find and plan for before cutover, not during it.
• Storing is cheap; leaving is not. Egress, pulling data back out of the cloud, is where the bill grows, and it’s the same trap on all three providers. Put it in your estimate and your bill-shock defences.