Managing any work in the system requires a set of performance objectives
in terms of resources used, response goals, and service requirements.
To achieve these objectives, an installation monitors how work is
running in the system and then tunes the system to try to meet the
performance objectives.
Incorporating new work, such as APPC, into a tuned system can affect
system performance. The purpose of this chapter is to provide guidance
to help you get the best performance possible for APPC work without
upsetting the balance of existing work in the system.
The steps for measuring the effect of APPC work in your system
and for tuning the system are listed below. Each step is described
in more detail in the sections that follow.
- Consider the storage requirements of APPC/MVS
- Monitor APPC performance
- If resources used by TPs are charged to accounts, use SMF to audit
APPC work.
- Use RMF reports to measure the performance of APPC as a whole,
by class, by account number, or by single transaction program.
- Use the DISPLAY APPC and DISPLAY ASCH operator commands to take
a "snapshot" of APPC work in the system.
- Improve performance by program design and administration
- Make efficient use of callable services.
- Be aware of keywords that cause performance degradation in the
TP profile's JCL.
- Consider using multi-trans scheduling to enhance performance.
- Define classes for APPC work with appropriate response time goals.
- Put each multi-trans TP in its own class.
- Associate TPs to LUs with the appropriate level of access to eliminate
unnecessary searching.
- Eliminate unnecessary use of the TP message log.
- Improve performance through system changes
- Consider the amount of buffer storage needed for received data.
- Improve verification performance of RACF® by
using VLF. For complete information, see z/OS Security Server RACF System Programmer's Guide.
- Eliminate unnecessary SMF recording.
- Improve network performance.
- Minimize use of APPC component tracing.