CRUSH stands for Controlled Replication Under Scalable Hashing
CRUSH is the algorithm used by CEPH to determine data placement and retrieval across OSDs based on defined rules and a map of the cluster.
Analogy: CRUSH is like the library’s card catalog, mapping each book’s location (data object) based on its title (object ID) and the library’s layout (cluster map).
Why It Matters: CRUSH enables intelligent data placement, rebalancing, and recovery without relying on a centralized lookup table, ensuring scalability and efficiency.