BDLS
The standard BDLS process as implemented in the SCP template consists of the following steps:
- Call a process definition to prepare the database. If no such process definition exists, then this step is skipped
- Remove the target logical system name from BD54
- Reset the BDLS status
- For all configured conversions, call BDLS in test mode. The test runs are scheduled to run in parallel
- For all configured conversions, call BDLS in conversion mode. The conversion runs are scheduled to run sequentially
- Amend variants for APO (if needed)
- For all configured conversions, change the description of the logical systems in BD54
- Call a process definition to revert changes to the database. If no such process definition exists, then this step is skipped
- Collect statistics of all BDLS runs
This process is known to work for all scenarios, although it might not always provide the best performance. As such you can supply a process definition to prepare the database such that the BDLS conversion runs faster.
The supplied process definitions from step 1 can be used to improve the speed of BDLS conversion by:
- Adding indexes to the database tables
- Increasing the parallelization mode on database tables
- Switching off the archive logging mode
- Changing the database parameters for example buffer sizes
- Changing the OS parameters
These changes can be reverted in step 8.
The actual performance of BDLS depends on many factors, with the performance of the database being the most important one.
The following conditions have an impact on the BDLS performance:
- Number of CPUs of the database server
- Capabilities of the database to handle many parallel processes
- Location of the database server (if database and application servers are running on the same machine, they will compete for resources, for example CPU power).
- I/O performance
- Number of entries to convert
- Size of the BDLS commit block
See Also
sscpTopic