Shared Record Positioning

Shared Record Positioning (SRP) is used when a data set that resides on DASD has to be shared between a writer and reader pair of steps that are running in parallel. The writer and reader steps can run on the same or on different systems in the Parallel Sysplex.

SRP allows data blocks that are being written to DASD (writing process not completed) to be read by subsequent job steps that are running in parallel with the writer step. Job Optimizer periodically passes the location of the last written block on DASD to the reader step. This allows the reader step to know how much data is available.

The following figuer shows a traditional job stream. Step 1 writes data blocks to DASD. When all of the data blocks have been written, the data set is closed and Step 1 terminates. Step 2 then reads the data from DASD. When Step 2 finishes processing all of the records, the data set is no longer required and can be deleted. In effect, the transfer of data from one step to the other is at a data set level.

Two steps in a traditional job stream

The following figuer shows a job stream using SRP. In Step 1, data blocks are written to DASD. Periodically Job Optimizer sends the location of the last written record on DASD to Step 2. In Step 2 the Reader uses the SRP information to know the position of the last data block written in the data set. The Reader can read all information up to the last data block written.

Shared Record Positioning

The steps are executed in different systems. SRP allows Step 2 to read blocks as they are written by Step 1.

Was this page helpful? Yes No Submitting... Thank you

Comments