Managing workload with Tivoli Workload Automation

Tivoli Workload Automation is a portfolio of products provided by IBM® to automate all workload management tasks. The scheduling features of Tivoli Workload Automation help you plan every phase of your workload production. During the processing period, the production control programs manage the production environment and automate most operator activities. The schedulers prepare jobs for running, resolve interdependencies, and launch and track jobs. Because jobs start running as soon as their dependencies are satisfied, idle time is minimized, and throughput improves significantly. Jobs never run out of sequence, and, if a job fails, the schedulers can handle the recovery process with little or no operator intervention.

Workload management is based on a database that contains the definitions of the scheduling objects. There are two versions of the scheduling objects database depending on the placement of the main workload controller: it can be based on a mainframe computer (in this case, z/OS®) or on a distributed platform. Some of the scheduling objects can exist in both of the databases, some apply only to the distributed platform, and others might apply only to the mainframe platform.

The minimum set of object definitions that are required to produce a workload consists of a workstation, a job, and a job stream. Other required scheduling objects might be predefined and exist by default.

A workstation is a definition that represents a computer system or another entity that is capable of running specific tasks, and that has the ability to report the status of task execution to the scheduler. With the Tivoli Workload Automation interfaces, you can identify the physical resources associated with the workstations.

A job is the representation of a task (an executable file, program, or command) that is scheduled and launched by the scheduler. The job is run by a workstation and, after running, has a status that indicates if the run was successful or not. A job definition can specify information on what to do whenever its run was not successful. Jobs not included in a job stream do not have any attribute for running, and are only the description of a task with a definition on how to perform it in a form that is known to the specified workstation.

A job stream represents a container for related jobs and organizes them, in terms of run time, sequencing, concurrency limitations, repetitions, assigning priority or resources, and so on. Job streams are the macro elements of the workload that you manage.

The scheduling plan is the to-do list that tells Tivoli Workload Scheduler or Tivoli Workload Scheduler for z/OS what jobs to run, and what dependencies must be satisfied before each job is launched. Tivoli Workload Scheduler or Tivoli Workload Scheduler for z/OS builds the plan using the elements that are stored in the scheduling database.

The running of a plan requires tracking to identify possible problems that can impact the effective delivery of the work products. It is possible to perform the tracking from a Web-based Java™ interface, the Dynamic Workload Console, on either of the platforms (z/OS and distributed). As an alternate interface to the Dynamic Workload Console on the z/OS platform you can also use the ISPF panel interface, and on the distributed platforms you can use the command-line interface.

See A business scenario which describes a possible implementation of workload management based on Tivoli Workload Automation.