The Flash Pool technology enables you to add one or more RAID groups composed of SSDs to an aggregate that consists of RAID groups of HDDs.
The SSDs provide a high-performance cache for the active data set of the data volumes provisioned on the Flash Pool, which offloads I/O operations from the HDDs to the SSDs. For random workloads, this can increase the performance of the volumes associated with the aggregate by improving the response time and overall throughput for I/O-bound data access operations. (The performance increase is not seen for predominantly sequential workloads.)
The SSD cache does not contribute to the size of the aggregate as calculated against the maximum aggregate size. For example, even if an aggregate is at the maximum aggregate size, you can add an SSD RAID group to it. The SSDs do count toward the overall spindle limit.
The HDD RAID groups in a Flash Pool behave the same as HDD RAID groups in a standard aggregate, following the same rules for mixing disk types, sizes, speeds, and checksums.
The checksum type, RAID type, and RAID group size values can be configured for the SSD cache RAID groups and HDD RAID groups independently.
There is a platform-dependent maximum size for the SSD cache. For information about this limit for your platform, see the Hardware Universe.