Isilon multi-tiered storage is called SmartPools. Two critical terms for SmarPools are Disk Pools and File Pools.
A Disk Pool consists of similar Isilon storage nodes. Multiple Isilon nodes of similar series can work together in a single cluster. For example, a Disk Pool of S Series nodes is applicable to I/O-intensive applications, a Disk Pool of X Series nodes is applicable to high-concurrency and sequential workloads, and a Disk Pool of NL Series nodes is applicable to nearline access. This means different performance and capacity of different node types can be mixed to meet various workflow requirements on a unified management interface.
A File Pool is a user-configurable policy which determines how data is placed, protected, accessed, and moved among Disk Pools. Such a policy enables data to be automatically moved from one type of storage to another within a single cluster to meet performance, space, cost or other requirements while retaining data protection settings. For example, a File Pool may specify that data written to /ifs/foo goes to Disk Pool 1 consisting of S series nodes and then moves to Disk Pool 3 consisting of NL Series nodes 30 days later.
Isilon provides three types of nodes: S Series nodes, X Series nodes, and NL Series nodes, and each type of nodes constitute a Disk Pool (tier).
As shown in the figure above, small files are written to Disk Pool 1 consisting of SAS disks and SSDs for fast-lightning access, files larger than 2 GB are stored on Disk Pool 2 consisting of SATA disks, and files that are not accessed in 30 days will be moved to Disk Pool 3 consisting of high-density SATA disks for lowered storage costs.
