KGUP processes keys on a disk copy of a CKDS which is a VSAM data set. KGUP uses key-direct update processing to process the keys. To access keys, VSAM uses the key's label as the VSAM key. This means that keys are added to the data set in collating sequence. That is, if two keys named A and B are in the data set, A appears earlier in the data set than B. As a result, adding keys to the data set can cause multiple VSAM control interval splits and control area splits. For example, a split might occur if the data set contains keys A, B, E and you add C (C must be placed between B and E). These splits can leave considerable free space in the data set.
The amount of control area splits and control interval splits in the CKDS affects performance. You may want to periodically use the TSO LISTCAT command to list information about the number of control area splits and control interval splits in a CKDS.
You can help reduce the frequency of control interval and control area splits by ensuring that key generator utility control statements are always in the correct collating sequence, A-Z, 0-9, if possible. When adding keys to a new CKDS, add the key entries in sequential order. Also, when adding new entries to the CKDS, you can reorganize the data set to reduce control area splits and control interval splits. To do this, copy the disk copy of the CKDS into another disk copy using the AMS REPRO command or AMS EXPORT/IMPORT commands. You may want to reorganize the data set after every KGUP run.
For a detailed explanation of keyed-direct update processing and a description of what happens when control area and control interval splits occur, refer to z/OS DFSMS Access Method Services Commands.