crush-placement · shipping · also Controlled Replication Under Scalable Hashing, CRUSH placement

CRUSH

CRUSH is a deterministic, pseudorandom placement algorithm that maps an identifier to storage targets using a weighted cluster hierarchy and placement rules.

Related Observatory article: CRUSH on hesela.com

Clients can calculate target locations without a per-object location directory. [1] In Ceph, rules can separate replicas or erasure-code shards across modeled hosts or racks. [2] CRUSH is a placement mechanism, not an erasure code, a consensus protocol or a guarantee of availability.

Relations

related

failure-domain, placement-group, erasure-coding

Citations

1. Weil, Brandt, Miller and Maltzahn, CRUSH: Controlled, Scalable, Decentralized Placement of Replicated Data, SC 2006, sections 3-3.2 (accessed 2026-10-13)
2. Ceph Squid documentation: CRUSH Maps, hierarchy and placement rules (accessed 2026-10-13)
3. Storage availability and correlated failures: placement boundaries (accessed 2026-10-13)

JSON: /data/terms/crush-placement.json