🗄️

Ceph Distributed Storage Networking

Complex

Free preview · 2 of 5 steps

Ceph: no single point of failure

Ceph is a distributed storage system where any node can fail and the cluster continues. Unlike traditional SAN/NAS, there's no central metadata server for object storage — clients compute data location directly using the CRUSH algorithm.

Ceph separates traffic onto two networks: a public network for client I/O and a cluster network for internal replication and recovery — preventing recovery storms from impacting client latency.

Packets in motion — click to inspect:

1 / 2

Space to play/pause · ← → to step