Federator® Scale-out storage allows horizontal scaling of storage capacity and storage processing by adding more storage nodes with their individual CPU processors and storage into a cluster.
Ceph, a clustered platform that simultaneously offers object, file, and block services, is rapidly becoming the de facto scale-out storage solution in the open source community. However, today’s high-density storage servers typically house a large number of disks in a single node, deploying and managing these individual disks as Ceph Object Storage Devices (OSD’s) becomes a challenge. In addition, the physical limitations of the disks (size and performance) become constraints when configuring Ceph patameters such as Placement Groups (PG’s).
At the cost a very small amount of overhead, Federator® can deploy a Ceph cluster using the storage hypervisor to create software-defined OSD’s for Ceph. Not only can this architecture greatly simplify disk management for large-scale deployments, it allows Ceph to access OSD’s with advanced enterprise features such as compression and deduplication without capacity or performance limitations typically associated with physical disk drives. Individual disk failures are no longer visible to the Ceph cluster and the recovery process is strictly limited to the involved storage node instead of triggering high volumes of data copy activities between the OSD’s across the network.