Windows Server 2012 R2 Storage Spaces boosts performance
Three new features in Windows Storage Spaces -- storage tiering, write-back caching and dual parity -- boost storage performance.
What are Storage Spaces' capabilities in Windows Server 2012 R2?
In addition to basic storage virtualization, provisioning and management, Windows Server 2012 Storage Spaces protects data, using snapshots and failover capability to support storage availability. Storage Spaces uses data striping (RAID 0) to break up data between disk groups to run multiple spindles and boost disk performance. When disk problems occur, rebuilds take place using spare pool capacity to preserve storage availability. Periodic background scans that preemptively check and correct disk problems enhance data integrity.
Windows Server 2012 R2 features several Storage Spaces improvements, the most noteworthy of which enables data deduplication to boost performance while minimizing storage capacity requirements. Three new features include storage tiering, write-back caching and dual parity.
Storage tiering allows Storage Spaces to create virtual disks using either magnetic or solid-state media, then places data onto the media that best fits the data usage patterns. For example, data that is frequently accessed can be dynamically located onto a hot tier using SSDs, while less frequently accessed data can be located on cold-tier disks using standard SAS HDDs. Data can also be relocated dynamically as data access patterns change over time. IT or storage administrators can also direct data to desired tiers.
The use of write-back cache caches new data to the storage array, rather than waiting for physical disk writes to complete. This improves storage write performance and allows applications to resume working faster, while reducing I/O traffic in the storage network and subsystem, which is particularly useful in shared storage deployments and tiering.
Finally, Windows Server 2012 R2 Storage Spaces supports dual parity, which can detect and recover from two simultaneous disk failures (rather than a single disk failure for standard parity). When faults occur, the disks can be rebuilt quickly using spare storage capacity to minimize the impact on storage performance during a lengthy disk-rebuild process.