- Modify/create JCL
- Update job schedule
- Set up conversion to run in multiple batches over
a period of time if the volume of data is high or
your batch window constraints are tight.
- For example, if ten billing cycles are run each
month, it may make better sense to do the data
conversion in a job preceding each billing cycle
(convert as touched) rather than one massive job.
- Or, do data conversion incrementally as records are
accessed by a bridge program, rather than having separate
conversion programs and jobs.
You can avoid bridging for data conversion if you add
compliant dates to records and retain the legacy
dates; your programs just use the format that they need.
- Modify/create JCL to run jobs at appropriate times and
locations.
- Update job schedule
Batch jobs are obviously required for batch systems but
may also be required to support on-line systems that
access data after it has been run through a filter-style
bridge.
The setup of bridge jobs is affected by the error-handling
model:
- In an error-end design, errors are not
passed through the bridge and must therefore be handled
by the sending partition
- In an error-thru design, errors are
passed through the bridge and therefore must be handled
by the receiving partition.
Either design may employ error log and transaction-record
suspense files.
Changes to dataset names often require changes to job
scheduling and report-distribution systems. The job scheduling or
report system can be triggered by the creation of the dataset.
©1997 IBM Corporation