Cloud Storage and Databases Explained

Four shapes, four different jobs

Cloud storage is not one product — it is a family of products, each optimised for a different pattern of use. Picking the wrong one usually still works at first, but becomes expensive or slow as an application grows.

Object storage: for files at massive scale

Object storage holds files — images, videos, backups, documents — each accessed as a whole unit through a unique address, similar to a URL. It is built to scale to enormous amounts of data cheaply and reliably, which is why it is the default home for things like user-uploaded photos, video files, and backups across almost every cloud application.

Block storage: for a server's own disk

Block storage behaves like a traditional hard drive attached directly to a virtual server — the operating system reads and writes to it exactly like local disk space. It is what a virtual machine actually boots from and runs its software on, and it is built for fast, low-latency access rather than for holding enormous, rarely touched files.

File storage: for shared folders

File storage provides a shared folder structure that multiple servers can read and write to at the same time over a network, much like a shared drive in an office. It fits situations where several machines genuinely need to work from the same set of files simultaneously, which object and block storage are not designed to do as naturally.

Managed databases: for structured, queryable data

A managed database stores structured information — user accounts, orders, inventory — in a way that can be searched, filtered, and updated precisely, with the cloud provider handling backups, patching, and reliability behind the scenes. This is the natural home for the kind of data an application constantly asks specific questions about, rather than simply storing and retrieving whole files.