Google Cloud Storage

Google Cloud Storage is a managed object storage service. You store data as objects inside buckets. An object combines bytes, a name, and metadata. A bucket supplies the shared location, access, protection, and lifecycle settings around those objects.

This model differs from a disk or shared file system. An application reaches Cloud Storage through an API, a client library, the Google Cloud CLI, or an HTTP endpoint. It does not receive raw blocks to format. It also does not receive a normal operating-system file system with arbitrary in-place writes.

That distinction guides every design decision. Cloud Storage fits data that your application can read or replace as whole objects. Common examples include media, documents, static web assets, backups, logs, data exchange files, and analytics input.

The resource model

Every bucket belongs to a Google Cloud project. The bucket has a globally unique name and a location. Objects live inside that bucket and have names such as reports/2026/july.csv.

The slash characters in that name can support a folder-like view. The object still has one name unless you use hierarchical namespace features. Do not assume that every object name maps to a directory on a mounted disk.

Cloud Storage assigns a generation number when it creates an object. If an upload replaces an object with the same name, the replacement receives a new generation. Generation and metageneration values also support request preconditions. A precondition lets you reject a write when the object changed after your application last observed it.