Ceph: Designing and Implementing Scalable Storage Systems

Erasure Coding for Better Storage Efficiency

Ceph's default replication level provides excellent protection against data loss by storing three copies of your data on different OSDs. The chance of losing all three disks that contain the same objects, within the period that it takes Ceph to rebuild from a failed disk, is verging on the extreme edge of probability. However, storing three copies of data vastly increases both the purchase cost of the hardware and also associated operational costs such as power and cooling. Furthermore, storing copies also means that for every client write, the backend storage must write three times the amount of data. In some scenarios, either of these drawbacks may mean that Ceph is not a viable option.

Erasure ...