Customize Virtual Device Engine › Virtual Volume Compression › How Virtual Volume Compression Works

How Virtual Volume Compression Works

When compression is active, CA Vtape always compresses the first 5 MBs of every 50 MBs of data that is written to the Virtual Volume. CA Vtape always compresses the first 5 MB of data to determine the best compression method for the remaining 45 MB of data. If the compression rate falls below the MinimumCompressionRate setting, then CA Vtape discontinues compression until the next 50 MB segment is written.

When you set the MinimumCompressionRate to a low value, CA Vtape compresses a larger percentage of the Virtual Volumes and more CPU is used for compression. When the MinimumCompressionRate is set to a high value, CA Vtape compresses a smaller percentage of the Virtual Volumes and uses less CPU for compression.

Compression reduces I/O and storage space at the cost of more CPU. The compression routines in CA Vtape rely heavily on the use of certain hardware instructions to perform compression. The CPU expense for using these instructions varies based on the CPU hardware in your environment. In some instances, the additional CPU expense might not justify the use of compression.

MaximumCompressionCPU lets you control this additional CPU expense by controlling the percentage of data compression is performed on:

• 100: CA Vtape compresses all data that is written to a Virtual Volume. 100 percent produces the maximum compression at the highest CPU cost.
• 50: CA Vtape compresses 50 percent of the data that is written to a Virtual Volume. On average this results in approximately 50 percent less compression at 50 percent less CPU cost.
• 0: CA Vtape does not compress data except for the first 5 MBs of each 50 MB of data written.

If a zIIP Engine is installed (and available) and the zIIPExploitation and PercentRunOnzIIP Parmlib attributes are set appropriately, the compression work is reduced, lowering the CPU compression cost.