6. DATA SOURCES › 6.5 Batch Turnaround/Response Time

6.5 Batch Turnaround/Response Time

Batch can be considered as a special case of application
response time.  An input is received in the form of a job to
process.  The job can wait on an input queue.  Time is taken
to process the job.  Output can be produced that can also
wait in a queue.  There can be network delays if the job is
received from or sent to a remote workstation or system.
For batch jobs, response time is normally referred to as
"turnaround time." However, like most response time
definitions, this one can have different meanings.  The three
most widely used are as follows:

  o  The classic definition of turnaround time refers to the
     time between submission of the job and the receipt of
     its output.

  o  Some installations that have implemented differential
     charging based upon turnaround time define it as the
     time between job submission and the start of execution.
     The charging rate varies with the time waiting for
     execution, not the time executing.

  o  A more popular definition has been to define turnaround
     as the time between job submission and the end of
     execution, with the idea that the output is then
     immediately available for viewing from a TSO terminal.
     However, the use of tools such as the QUEUE command,
     SDSF, and IOF permit the output to be viewed before the
     job has completed execution.

This section deals with the measurement of application
response time in a batch environment.  First, a diagram
showing the significant events in the life a batch job is
provided.  Next, the measured response intervals, which are
the elapsed time between the events are covered.  Finally,
considerations in the calculation of batch application
response time are discussed.

    1 - Batch Job Events and Measured Intervals
    2 - Batch Response Considerations