Ceph

Ceph is an open-source software-defined storage platform that provides object, block, and file storage in a single distributed system. Designed to run on commodity hardware, it aims to scale from small clusters to large deployments and is widely used to back private clouds, HPC environments, and archival storage. The project originated at UC Berkeley under Sage Weil and has evolved into a core component of many cloud and data-center infrastructures.

The core component is the Reliable Autonomic Distributed Object Store (RADOS), which stores data as objects across a cluster of storage daemons (OSDs). Clients access data through librados or native interfaces for Ceph Block Device (RBD) and CephFS. The cluster is coordinated by monitor daemons (ceph-mon) that maintain maps and quorum. Data placement is governed by the CRUSH algorithm, which avoids centralized metadata. The Ceph Object Gateway (RADOS Gateway, RGW) provides S3- and Swift-compatible object access. CephFS offers a POSIX-like file system backed by metadata servers (MDS). BlueStore is the default storage backend on OSDs, improving performance and efficiency.

Features include data durability through replication and erasure coding, self-healing and automated recovery, live rebalancing, and capacity expansion without downtime. Ceph supports snapshots and clones for block devices and file systems, and offers flexible pool configuration for data placement, compression, and encryption. It integrates with cloud and container platforms via drivers and APIs, including Kubernetes through RBD and CephFS CSI, and various OpenStack components.

Ceph’s design emphasizes scalability, reliability, and interoperability, making it a popular choice for enterprises seeking a unified storage solution on commodity hardware.