A cloud file system keeps a company's files in cloud object storage, but shows them to staff as an ordinary shared drive in every office. Staff open, save and share files exactly as they did on the office file server. Only the place where the files are kept has changed. It is also called a hybrid cloud file service, because one part sits in each office and the other part sits in the cloud. The page on hybrid cloud explains that wider idea.
Before cloud file systems, each office had its own file server, usually a box from an enterprise storage system maker. Files had to be copied between offices, so people often worked on an old version.
How does it work? The master copy of every file is kept in object storage in a cloud. Object storage is cheap and almost unlimited, but it is slow to open files one at a time, and it does not behave like a shared drive. So each office runs a small cache. The cache is a server, or a virtual machine, that keeps a copy of the files that the office uses most. When an architect opens a drawing, the cache serves it from its own disk, at office speed. When she saves it, the cache sends only the changed parts to the cloud. The other offices see the new version within minutes. When two people in different offices open the same file, the system locks it. So one person cannot overwrite the other's work.
Take, for example, an architecture firm with offices in Chennai, Kolkata and Bhubaneswar. Its 400 staff work on drawing files of up to 500 MB each, and the firm holds 60 TB of files in all. In the past, a project shared between Chennai and Kolkata meant copying the files every night. Now all three offices see one set of files. The Kolkata cache holds the 3 TB that the office is using this month. The rest stays in the cloud, and it is fetched when someone opens it.
Protection from ransomware is built in. The system takes a snapshot every few minutes. A snapshot is a read-only record of every file as it was at that moment. Say ransomware encrypts the files on the Chennai office's drive at 11:40 a.m. The IT team restores every affected file to the snapshot from 11:35 a.m. Because the snapshots cannot be changed, many firms no longer run separate nightly backups of these files. They still use backup and recovery for their servers and databases.
A cloud file system differs from online backup for one person's computer. It is a shared drive for a whole company, with permissions, file locks and snapshots.