Compression Analysis and Implementation › Analysis › Implementing Compression

Implementing Compression

For VSAM data sets, CA Compress supports Scheduled or Immediate compression. Scheduled compression specifies that compression is to take place as a future event, either when the data set is opened for output or loaded. Because nothing has to be done to the data set itself at implementation, the IUI is able to perform scheduled implementation online, by simply updating the Control File. Immediate compression means that the entire data set is compressed at the time it is implemented. Key position and maximum record length can change, and the data set must be redefined and reloaded. Because the data set, which can be large, must be backed up and reloaded, the IUI performs this multistep operation by creating a batch job for you to submit.

For Physical Sequential data sets, unlike VSAM, there are no keys and therefore no change in key position. Records cannot be updated in place, and fixed and variable records are not compatible in the same data set so that CA Compress cannot consistently support compressed and uncompressed records together. For these reasons, scheduled compression in the VSAM sense is not appropriate for Physical Sequential data sets, but deferred and immediate compression are still supported through the effective date (EFFDATE parameter or ANYDATE).

If you specify or default a nonzero EFFDATE, the data set is compressed the next time it is loaded, at or after the effective date, like VSAM SCHEDULED=LOAD. If you specify ANYDATE, the IUI builds JCL to back up and reload the data set, much as it does for VSAM Immediate compression.

A third possibility with Physical Sequential data sets is that the data set is already compressed, having been implemented under an older version with the SUBSYS JCL parameter. Such a data set is marked COMPRESD in the work list. If you select to implement it, the IUI implements it online with ANYDATE, using the FDT or algorithm it was compressed with, and it is immediately treated as compressed under the Physical Sequential Transparency (PST).

The IUI performs implementation in one of two modes, CHOICE and SCHEDULED, as determined by your system administrator. As distributed, the system default is CHOICE.

CHOICE lets you control how each data set is implemented. For both VSAM and Physical Sequential data sets, in addition to selecting deferred or immediate compression, you can specify various optional parameters. SCHEDULED defaults to SCHEDULED=LOAD for VSAM and EFFDATE=today for Physical Sequential. The result in each case is that the data set is compressed the next time it is loaded, and the IUI takes defaults for parameters you can control under CHOICE.

The following table shows the implementation process that takes place based on the type of routine being implemented and the setting of the System Default field. The table assumes normal operation and that the user enters all required information correctly. The simplest process is when Scheduled is the System Default and the user wants to implement compression with a standard routine, Standard Table, or Super Express. In this case CA Compress just implements compression for the selected data sets and no further action by the user is required. On the other hand, the most involvement is required when Choice is the System Default and you want to implement compression with a custom routine; Huffman, SRHVL, or Tailored. In this case, you must provide an FDT name and a mode type before CA Compress implements compression.