Skip to main content
Cloud & AI Hub
Browse
Glossary AI Directory Playgrounds Models Prompts Explainers Strategy Matrix Benchmark Decoder
Storage

Aliases: S3, Blob storage, Bucket

Object Storage

Cloud storage that keeps files as objects in flat buckets accessed over HTTP — cheap, effectively unlimited, and the backbone of cloud data.

What is object storage?

Object storage stores data as objects — the file bytes plus metadata plus a key — in flat containers called buckets, accessed via HTTP APIs rather than mounted filesystems. Amazon S3 defined the model; Google Cloud Storage, Azure Blob Storage, and S3-compatible systems (Cloudflare R2, MinIO) followed. It scales to exabytes with 11-nines durability and pennies-per-GB pricing.

Object vs block vs file

TypeLooks likeBest for
ObjectHTTP API + keysBackups, media, data lakes, static sites, ML datasets
BlockA raw disk attached to one VM (EBS)Databases, boot volumes — anything needing a filesystem
FileA shared network drive (EFS, NFS)Legacy apps, shared home directories

Objects are immutable — you replace, not edit — and there are no real directories, just key prefixes that look like paths.

Storage classes and lifecycle

Providers tier by access frequency: hot (standard) → infrequent access → archive (Glacier-class, hours to retrieve). Lifecycle rules move objects between tiers automatically — the difference between a data lake that costs 2k/monthandthesameoneat2k/month and the same one at 400.

Cost reality

The sticker price (~0.02/GBmonth)isthesmallpart.Thebillsthatsurpriseteams:egress(gettingdataoutcosts 0.02/GB-month) is the small part. The bills that surprise teams: **egress** (getting data *out* costs ~0.09/GB on major clouds — moving 100 TB out is ~$9k, which is R2’s whole zero-egress pitch), request charges on millions of small objects, and cross-region transfer. Design rule: process data in the region where it lives.

What people get wrong

  • Public buckets. The classic cloud breach; default-deny, use presigned URLs for sharing, audit with provider tools.
  • Using it as a filesystem. Listing “directories” with millions of keys and rewriting objects for small edits fights the model; that workload wants block or file storage.
  • Ignoring lifecycle from day one. Data accumulates forever by default — the bill only goes up until you add rules.

Historical figures and technical concepts for informational purposes only. Not technical, professional, legal, or financial advice. Sources: Official Documentation.