Cache partitioning function

When multiple hosts share one storage system, if a specific host reads and writes large amounts of data, the read and write data on the host might occupy a lot of space in cache memory. In such a condition, the other hosts must wait for the data to be written to the cache, which might slow down the data writing process.

The cache partitioning function can divide cache memory built into the storage system into multiple virtual cache partitions and allocates the available cache capacity to the host in advance to avoid a condition in which the specific host occupies a lot of cache memory.

The following figure shows an example of a company's network on which cache memory is divided into three virtual caches that are allocated each to three branch offices. The host at branch office A is accessing the storage system a lot, however, the available cache capacity is limited to 40 GB so that the read and write data on the host at branch office A does not occupy the entire cache. No matter how much data the host at branch office A accesses, the hosts at the other branch offices can always have 40 GB of cache space available. Therefore, the data transfer speeds at the other branch offices are not slowed down due to the data access volume of branch office A.

The virtual cache memory divided by using the cache partitioning function is referred to as CLPR and up to four CLPRs can be created on one storage system, including the default CLPR0.

You can assign one of the following to a CLPR:

• An external volume group
• A virtual volume

Use the cache partitioning function only for the following purposes:

• Divide the cache capacity between external volumes and resources that are not external volumes.
• Divide the cache capacity for each external storage system if there are external volumes from multiple external storage systems.