Documenting the workload

Before setting objectives, you must understand and document the current workload on the system.

It is very important to fully document your workload. Some of the most significant performance gains to be achieved are accomplished by workload management. The more details you know about the workload, the better you can manage it.

You should document your workload by the amount and categories of work:

• Priority of the work
• Current activity rates and response time levels
• Different periods during which objectives and priorities vary for the same work
• Resource consumption of the work
• Types of users requiring different objectives
• Ability to track and report on work according to your organization’s needs, such as by departments, groups, projects, and so on

Start with the workload categories that are currently defined for your system (for example, TSO and CMS transaction types, batch windows, and so on). Usually these categories are not fully defined. Table 1 lists some factors to consider when defining each workload category, including resources consumed by each workload. When initially defining the categories, the resource consumption will probably reflect expected resource usage; the measurement of actual resource usage on z/OS is described in more detail in Measuring resource consumption.

After you define the categories, review them for definition problems. The different categories distinguish work by resource needs, objectives that must be met, priorities, and so on. For example, all jobs submitted from similar development groups in different locations should receive the same turnaround time. However, because of distribution of the completed work to different locations and possible time differences in actually returning output to the submitters, you might want to further separate this work—to give priority, for example, to jobs that must be distributed to locations in different time zones, where delays in turnaround time can have a significant effect on the users.

You should determine the factors listed in the above table for each level of each workload category. For example, determine the factors for batch, TSO, CICS, and IMS or other subsystems. Within each subsystem, determine the factors for further breakdowns of that workload type:

• TSO, CICS, and IMS divided into transaction types
• Batch work divided into batch windows
• VM users divided into accounting groups
• All categories divided by peak hours and off shifts

z/OS lets you associate transactions with a set of performance characteristics through performance groups. You can group your transactions into categories using performance group numbers. You assign performance groups through the installation control specification. For complete information, refer to the z/OS Initialization and Tuning Guide.

VM lets you group transactions and users through user classes. For z/VM data, user classes might be CMS users, service machines, and guests.

By assigning each distinct workload type to a separate performance group, you can use IBM Z Decision Support reports to obtain data on the processor, processor storage, and I/O activity for each category. Performance groups are the basic entity for z/OS workload measurements.