To make the most of workload management, work needs to be properly distributed so that MVS™ can manage the resources. It is essential that the subsystems distributing work are configured properly for workload distribution in a sysplex. You do this with the controls provided by each subsystem. For example, in a JES2 and JES3 environment, you spread initiator address spaces across each system.
Initial cooperation between MVS and the transaction managers (CICS®, IMS™, DB2®) allows you to define performance goals for all types of MVS-managed work. Workload management dynamically matches resources (access to the processor and storage) to work to meet the goals.
CICS, however, goes further with the CICSplex Systems Manager (CICSPlex® SM) to dynamically route CICS transactions to meet the performance goals. CPSM monitors the performance of CICS resources and systems and presents the resources as though they were part of a single system. This type of cooperation greatly improves CICS transaction workload balancing.
Other subsystems also have automatic and dynamic work balancing in a sysplex. For example, DB2 can spread distributed data facility (DDF) work across a sysplex automatically. DB2 can also distribute work in a sysplex through its sysplex query parallelism function. CICS, TSO, and APPC cooperate with VTAM® and workload management in a sysplex to balance the placement of sessions. SOMobjects® can automatically spread its servers across a sysplex to meet performance goals and to balance the work.
For more detail on workload management in different subsystem environments, see Workload management participants.