Performance and tuning externals are scattered between MVS™ and various subsystems, as well as throughout monitoring and reporting products. This is true for the OPT parameters, and, prior to z/OS® V1R3, for the MVS installation performance specifications (IPS) and the installation control specifications (ICS). MVS and the subsystems each have their own terminology for similar concepts, each have their own controls, and the controls are not coordinated.
Many of the MVS externals are geared towards implementation. You tell MVS how to process work, not your expectations of how well it should run work. There is no single way to make sure that important work is getting the necessary system resources.
The multiple monitoring and reporting products show many different views of how well MVS is doing, or how well individual subsystems are managing work. Since there is no single way to specify performance goals for your installation, it is also difficult to get feedback from monitors and reporters on how well your installation actually achieved what you expected it to. There is little sense of which reports and fields relate to the externals you specified.
Some installations configure their systems to handle a peak load. This could result in inefficient use of expensive resources. Today's externals do not allow you to reflect your performance expectations for work. Today, a system programmer must completely understand the implications of changing a parameter before he or she can configure an installation to achieve performance objectives. Since there is no direct path between specification and expectation, it is difficult to predict the effects of changing any one control or parameter.
Given this mix of problems, workload management has some high level objectives to provide a solution.