SeaweedFS

SeaweedFS is a distributed storage system that handles billions of files with O(1) disk access, usually a single read per file. A central master tracks only volumes, while volume servers manage files and their metadata, which keeps the master uncontended and lets the cluster scale horizontally across many machines.

A built-in blob store handles small files efficiently, and an optional Filer layer adds directories and POSIX attributes on top, with pluggable metadata stores such as MySQL, Postgres, Redis, and Cassandra. Files are reachable through an S3-compatible API, a Hadoop-compatible interface, a FUSE mount, and WebDAV.

Replication is rack and data-center aware, erasure coding reduces cost for warm data, and tiered cloud storage transparently offloads cold data. A built-in Iceberg REST catalog turns the same cluster into a self-contained lakehouse. SeaweedFS is Apache-licensed open source.