Cross-System Work extract

The Cross-System Work Extract accepts performance class data from a single or multiple CICSĀ® systems and correlates the data by network unit-of-work. A single performance class record is then written to the Extract data set. That one record represents all the work done on behalf of the network unit-of-work.

The default is to extract only the CMF performance class records that are contained in a unique network unit-of-work that includes multiple performance records.

Note: The Cross-System Work Extract will also include multiple performance class records from a single system.

You can request an extract that processes all available input records, or you can specify criteria for record selection to extract only the data that meets specific requirements.

The extract records have the same format as CMF performance class records written by the latest CICS release supported by CICS PA (VRM 700), regardless of the CICS releases of the input records.

After a Cross-System Work Extract data set has been created, it can be used as input to CICS PA for further processing. For example, the Performance List, Performance List Extended, Performance Summary, and Performance Totals Reports can be run against this data set.

Note: If you are using conversational transactions, and you have specified MNCONV=YES in your system initialization parameters to get separate CMF records for each pair of terminal I/O requests, or you have specified MNSYNC=YES in your system initialization parameters to get separate CMF records for each unit-of-work, or you have applications that are using user event monitoring points (EMPs) with the DELIVER option, all records will still be part of the same network unit-of-work. Since they are part of the same network unit-of-work, they will all be merged into one record in the Cross-System Work Extract Data Set. If you, for example, run the Performance Summary Report against this data set, the response time does not represent the response time of an individual screen display, but the complete lifetime of this conversational transaction. The AVE, DEV, MAX, MIN, and TOT statistics might also be skewed in the same way.