Study the Adaptive Server environment.
See “Environment analysis and planning” for details.
Analyze the behavior of all execution objects and categorize them as well as possible.
Understand interdependencies and interactions between execution objects.
Perform benchmark tests to use as a baseline for comparison after establishing precedence.
Think about how to distribute processing in a multiprocessor environment.
Identify the critical execution objects for which you will enhance performance.
Identify the noncritical execution objects that can afford decreased performance.
Establish a set of quantifiable performance goals for the execution objects identified in the last two items.
Understand the effects of using execution classes.
See “Execution class attributes” for details.
Understand the basic concepts associated with execution class assignments.
Decide whether you need to create one or more user defined-execution classes.
Understand the implications of different class level assignments—how do assignments affect the environment in terms of performance gains, losses, and interdependencies?
Assign execution classes and any independent engine affinity attributes.
After making execution precedence assignments. analyze the running Adaptive Server environment.
See “Results analysis and tuning” for details.
Run the benchmark tests you used in step 1 and compare the results.
If the results are not what you expect, take a closer look at the interactions between execution objects, as outlined in step 1.
Investigate dependencies that you might have missed.
Fine tune the results by repeating steps 3 and 4 as many times as necessary.
Monitor the environment over time.