UNIXFs

UnixFS is a data format used by the InterPlanetary File System (IPFS) to represent files and directories within its content-addressable Merkle DAG. It provides Unix-like semantics inside a distributed, immutable storage system; UnixFS defines how file contents and directory listings are encoded into DAG nodes so that files can be chunked, cached, and retrieved efficiently.

Structurally, a UnixFS object is stored inside a DAG node (typically a Dag-PB node) and contains a data field that specifies the kind of object (file, directory, symlink, or metadata) and carries the associated payload. For files, the payload is split into blocks; each block is addressed by its own content identifier (CID) and the file’s total size is recorded in the metadata. For directories, the payload encodes a list of named links to child nodes, effectively representing the directory tree. UnixFS also stores metadata such as file size, Unix-style permissions (mode), and modification time.

UnixFS is designed to be extensible and efficient for large datasets. It supports streaming and random access through chunking, and its content-addressed nature enables deduplication across files and versions. The format is implemented in multiple IPFS language libraries and tools, and is used by higher-level features such as IPFS's Mutable File System (MFS), which presents a familiar file-system view atop the DAG.