Managing workload groups

Workload groups provide the capability to manage work on a system.

The workload groups function can be used to limit the processing capacity of a workload to a subset of processor cores in a partition. A workload group can be created with a limit on the number of processor cores. Jobs can then be assigned to the workload group. The system enforces this processing core assignment by ensuring that any job and its associated threads cannot run on more processor cores then have been designated by the workload group.

Workload groups can be used to get better control of a workload and to ensure products only use a designated number of processor cores. Additionally, software vendors can employ workload groups to support sub-LPAR licensing. To take advantage of the enhanced licensing controls for products, IBM® i License Management must be used to register and manage the enforcement of workload groups.

Collection Services, Performance Explorer, and Job Watcher have performance metrics that can help you manage and understand the performance of jobs running in a workload group.

Example of workload groups use

A user has a multithreaded batch job that is CPU intensive. The user must run this job during the day but cannot afford to affect the performance of the production system. Assigning this batch job to a workload group puts the job into a ‘processing container'. A workload group ensures that this job is kept to a limited amount of system processing capacity. If the workload group has a processor core limit of one, the batch job and any threads running under that job can only run on a single processor core. If this job is running on a multiple threaded core, multiple threads can be running for that designated batch job, but only a single core can be used at a time. This same concept applies to jobs running under a subsystem that has been assigned to a workload group. In this case, all jobs and their associated threads are limited to the number of processor cores specified in the group.