A file system is the set of data structures, rules and software mechanisms an operating system uses to organise, name, store, retrieve and manage data on a storage device or other storage medium. It provides the abstraction of files and directories over underlying storage such as hard-disk drives, solid-state drives, USB flash devices, memory cards and sometimes network or virtual storage.
Without a file system, storage would generally appear as an undifferentiated collection of blocks or another lower-level representation rather than a structured hierarchy of named objects. A file system determines how filenames, directory relationships, metadata, allocation information and file contents are represented. Different file systems make different trade-offs involving compatibility, performance, reliability, scalability, security and feature support.
A file is a logical unit of stored data identified through a name and location within a filesystem namespace. A directory, often called a folder in graphical interfaces, contains references to files and potentially other directories, creating a hierarchical namespace. File systems normally associate metadata with objects, including attributes such as size, timestamps, ownership or permissions, depending on the implementation. The exact metadata model differs considerably.
Unix-like systems commonly represent permissions and ownership as inode-associated metadata, while NTFS provides a richer attribute and security model. The visible filename is therefore only one component of how a filesystem identifies and manages data. Paths combine directory information and filenames into a representation that applications and users can use to locate objects.
At the storage level, many file systems divide storage into units such as blocks or clusters. When a file is created, the file system allocates appropriate storage units and records where the file's contents are located. The file system must also track which storage units are free and which are allocated. Fragmentation can occur when a file's contents are spread across non-contiguous areas. File systems use allocation structures to manage these mappings, but the specific mechanism differs. FAT-family systems use allocation tables, while NTFS uses the Master File Table and associated metadata structures. Modern copy-on-write file systems use substantially different approaches. File allocation is therefore a fundamental implementation problem, and the choice of data structures affects performance, scalability and recovery behaviour.
File systems also define how changes are made safely when operations are interrupted by crashes or power loss. Some file systems use journalling, in which intended metadata or other filesystem changes are recorded in a journal so that the system can recover to a consistent state after an interruption. Journalling does not automatically guarantee that every user-data write survives a power failure, because storage caches and filesystem configuration influence durability. Other file systems use copy-on-write techniques, checksums or transactional mechanisms. Reliability features vary substantially between file systems. A filesystem's apparent ability to store files is therefore only part of its design; consistency, recovery and corruption detection can be equally important in real-world use.
File systems can support additional functions beyond basic file storage. These may include access-control permissions, symbolic links, hard links, extended attributes, sparse files, compression, encryption, quotas, snapshots, deduplication, checksums and alternate data streams. Not every filesystem provides every feature. Network filesystems can make remote resources appear within a filesystem-like namespace, while virtual filesystems can present kernel or device information as if it were a hierarchy of files.
Operating systems may support many filesystem types simultaneously through a common filesystem abstraction layer. This allows applications to use standard file APIs without needing to understand the internal layout of each filesystem. The abstraction is particularly important because the same application may access files stored on different physical and logical media.
A filesystem is distinct from a partition, disk and volume. A physical disk is the hardware storage device, while a partition is a defined region of that device. A volume is a logical storage unit that an operating system can generally mount or otherwise access, and a filesystem can be created within that volume. The exact relationship varies among operating systems and storage technologies.
A single physical disk can contain multiple partitions and filesystems, while logical storage systems can aggregate physical devices into a volume before a filesystem is placed on top. Understanding these layers is important when diagnosing storage problems because formatting a filesystem is not the same operation as partitioning a disk, and deleting a partition can affect access to the filesystem it contains.
File systems also impose practical limits and compatibility constraints. Maximum file size, maximum volume size, filename rules, path-length behaviour, case sensitivity, supported characters, permission semantics and available metadata features can differ significantly. A file created on one filesystem may therefore not behave identically when copied to another. For example, a filesystem that supports Unix-style ownership and permissions may lose those semantics when files are transferred to a simpler filesystem designed primarily for removable-media compatibility. Applications may also depend on filesystem-specific features. Choosing a filesystem consequently involves considering the operating system, storage medium, intended workload, required reliability, security model and interoperability requirements rather than simply selecting whichever filesystem is newest or most familiar.
Common filesystem families include FAT variants, NTFS, exFAT, ext4, XFS, Btrfs, APFS, HFS Plus, ZFS and numerous network and specialised filesystems. FAT32 and exFAT are especially common on removable media because of their broad compatibility, while NTFS is strongly associated with modern Windows system and data volumes. Unix-like systems commonly use filesystems such as ext4 or XFS, while macOS commonly uses APFS. Modern filesystems increasingly incorporate mechanisms for integrity, snapshots, encryption and advanced storage management, but simpler filesystems remain useful where interoperability is more important than sophisticated features. The filesystem is therefore a foundational layer connecting applications and logical files to the underlying storage technology.
