This topic provides consideration for determining how many storage
pools you need in your environment.
Typically, you would assign volumes to storage pools based on common characteristics,
such as device capabilities (or quality of service) or data security.
When you assign volumes based on device capabilities (for example, performance,
reliability, and availability), you first need to classify your data and applications
according to its value to the business. These are some examples of assigning
volumes based on device capabilities:
- A storage pool that requires volumes that have the same RAID level, pathing
redundancy, and controller caching behavior.
- A critical storage pool that requires volumes that are multi-pathed and
RAID 5 with a large controller cache, such as with IBM® Enterprise Storage Server® (ESS).
- A storage pool that uses cheap storage.
- A storage pool that requires fast spindles for random I/O, such as in
OLTP.
- A storage pool that requires volumes on a storage device that supports
high transfer rates for serving music files to the web.
Storage pools defined in terms of data security would be used to limit
access. For example, a medical-records database would be kept in a storage
pool whose volumes are zoned to limit access.
Define your storage pool using factors from device capabilities and data-security
policies. Creating too few storage pools indicates that there is not enough
differentiation in quality of service or security. Because a reserve is typically
kept in each storage pool, creating too many storage pools might cause wasted
space.
Tip: To ease administration of the storage pools, make volumes
in a pool homogenous so that they can be administered uniformly.