While file storage uses complete files when storing data and block storage divides files into blocks of equal size, object storage stores data as objects. An object can be imagined as a package. It contains the data as well as numerous metadata and a unique ID. This ID enables a flat structure. Instead of a complex folder structure, data in object-based storage is stored side by side. The actual location (local or remote) is irrelevant.
In contrast to conventional file storage, the metadata for object-based storage is freely selectable. Instead of only storing information about the name, creation date, file type, etc. individual metadata can be defined for each file. For example, information from the corresponding application can be integrated or user data can be inserted.
Access to the data, i.e. communication with the storage, occurs via Hypertext Transfer Protocol (HTTP) and via a REST API. This simplifies the connection between client and object storage, because any computer can handle HTTP, which is common practice on the Internet. The storage can be addressed with a few commands:
- PUT: creates an object
- GET: reads an object
- DELETE: deletes an object
- LIST: lists all objects