Storage Spaces Direct data storage on disk

Q. How does Storage Spaces Direct store data on disk?

A. Storage Spaces Direct (S2D) enables local disks on nodes in a cluster to be aggregated together to form resilient storage pools that are visible as shared cluster shared volumes within the cluster. S2D utilizes SMB3 (and therefore SMB Direct where available) for the communication and replication of data between the nodes in the cluster. Virtual disks (not VHDs that are used with virtualization like Hyper-V) within the Storage Pools.

The virtual disks are made up of 1GB extents which means a 1TB disk has 1000 extents. Each extent has two replicas on different fault domains (read nodes) meaning there are three copies of all data (which is very similar to how Azure Storage stores and replicates data). Should a disk or node fail then a new copy is created of the extents that were stored on. Additional nodes and disks can be added to increase capacity and to rebalance data automatically as capacity changes. If a server is unavailable temporarily then when it comes back online data is resynchronized automatically.