What is object storage?

Object storage: definition and structure

With object storage, files are managed as objects. An object consists of the payload data, descriptive metadata and a unique identifier. It sits in a bucket, a kind of container, and is addressed not by a path in a file system but by its identifier. Folder structures are merely part of the object name.

Access runs over HTTP interfaces. The S3 interface has established itself as the de facto standard, and many European providers and open-source solutions support it as well. This makes it comparatively easy to switch or combine applications, backup programs and tools between providers.

How object storage differs from file and block storage

Block storage behaves like a hard drive attached to a server; it suits operating systems and databases. File storage provides network drives with a folder structure. Object storage dispenses with that structure and gains scalability as a result: very large amounts of data can be stored comparatively cheaply and robustly.

There are also features that are useful in operations: versioning preserves earlier versions, lifecycle rules move old data to cheaper storage classes or delete it, and an object lock prevents data from being changed for a defined period—an additional protection against ransomware.

Real-world example: construction site docu­menta­tion and backups

A construction company documents its projects with thousands of photos, plans and acceptance reports that take up more and more space on file servers over the years. The data must be retained but is only occasionally retrieved. Expanding its own servers would be expensive and laborious.

The company stores the documents in object storage with a provider in Germany. An app accesses it through an interface, older projects move automatically to a cheaper storage class, and nightly backups are also stored there, protected by versioning and object lock. The storage grows along with the company, and billing is based on the volume used.

Benefits, limits and what to watch for

Object storage is scalable, inexpensive, robust thanks to multiple copies and easy to integrate into applications. It is not, however, a replacement for databases or drives that need frequent small changes and low latency. Also note fees for retrievals and outbound traffic, as well as accidentally public buckets—a frequent cause of data leaks.

Nano Concepts integrates object storage into websites, platforms and backup concepts, sets up access rights, encryption and lifecycle rules, and keeps location, cost and data protection in mind, so that your files are stored securely, remain quickly accessible when needed and can later be moved to another provider without problems.

• Keep buckets private by default and grant access selectively
• Issue access keys per application and rotate them regularly
• Enable versioning and object lock for backups
• Define lifecycle rules for old data
• Budget for retrieval and outbound traffic costs