Explain like I'm five · Cloud

What are Block, File, and Object Storage?

Block, file, and object storage, told as the difference between raw bricks, a folder tree, and a warehouse of tagged boxes.

raw brickpaper folderstagged boxesevery tenant wants to store things in a different shape

Every tenant wants to store things in a different shape.Bricks, folders, boxes — all different shapes.

only paper foldersbricks don't fit in a foldercan't even tag a box herewith only one storage shape, somebody's stuff never fits

With only one storage shape, somebody's stuff never fits.In a folder-only room, neither bricks nor box tags work.

block: raw bricksfile: folder treeobject: tagged boxesblock=bricks, file=folder tree, object=tagged boxesblock, file, and object differ in HOW data gets stored and retrieved

Block=bricks, file=folder tree, object=tagged boxes.Block, file, and object differ in how data gets stored and retrieved.

Block = raw bricksstructureless pieces, and usually the fastest of the three
File = a folder treemost familiar to people
Object = tagged boxesfetched by its unique ID through an API
Object storage piles up almost without limitthe box warehouse can keep growing sideways
block vs file vs objectstructure: raw bricks vs folder tree vs tagged boxesaccess: usually fastest vs human-familiar vs via API lookupscalability: moderate vs moderate vs near-unlimitedtypical use: databases vs shared docs vs photos, backups, logs

The three storage types differ on structure, access, scalability, and use.Use cases split clearly — block for databases, file for shared docs, object for photos and backups.

object: even one tweak means resending the whole boxblock: swap just the one brick, so edits are quick

Even a small edit to an object means resending the whole box, so it's slower than block.Block storage rewrites only the changed piece, so edits are quick.

In one breath

Block is raw bricks (usually fastest, with no structure of its own). File is a folder tree (familiar to people). Object is tagged boxes (near-unlimited scale, fetched by unique ID).

Block storage is accessed in fixed-size chunks, suited to fast I/O like databases. File storage uses a directory/file hierarchy that's easy for people to navigate. Object storage keeps data as objects tagged with metadata and a unique ID, queried via an API, with near-unlimited horizontal scale — ideal for unstructured bulk data like photos, backups, and logs. But objects are inefficient to partially modify.

When grown-ups say it

Block storage
Storage carved into brick-like chunks. Usually the fastest of the three, but with no file structure of its own.
File storage
Storage organized like a folder tree. Most familiar for a person to navigate.
Object storage
A warehouse of boxes tagged with metadata. Fetched by its unique ID through an API.
Metadata
The tag attached to the box. Details that describe an object, like its creation date or type.
DAS / NAS / SAN
A difference in connection method and sharing scope. A different yardstick — how storage connects, not what shape data takes. → my drawer, the shared file room, the dedicated network
Horizontal scaling
Adding more warehouse space sideways, endlessly. This is how object storage scales almost without limit.
Backup
Keeping one more copy of the data elsewhere. A common use case for object storage.
Cloud service model
Bare shell / interior done / move-in ready. With self-built models like IaaS, you pick among these three. → pick one to rent
Next story
The Best-Known Name Behind the Auto-Assign Badge →
Want another term as a picture book? Request →