PCS description of your workload

The mission of ThruPut Manager is to automate and optimize the running of your z/OS JES2 batch workload. PCS is the part of ThruPut Manager that understands that some workloads are scheduled (by CA 7) and can make decisions about individual jobs as part of a scheduled workload. PCS understands where the critical path lies and which jobs are part of it, when the critical path shifts to another path, when a job is expected to complete and when a job is projected to be late.

PCS does more than "know" and "understand". PCS is proactive. For instance, deploying the ThruPut Manager engine, it favors higher importance jobs over lower importance jobs, and, for late jobs, changes Service Classes to one with higher WLM service goals. PCS provides comprehensive automation to manage your scheduled batch production workload, in an optimal manner.

Because of its unique perspective, PCS can present an integrated display that includes CA 7, JES2 and ThruPut Manager information, giving the operations staff the information they need to know to ensure ThruPut Manager is handling the scheduled workload in the optimal manner.

PCS keeps a detailed view of your scheduled workload. It interrogates the CA 7 database and builds up a picture of each application, the job trees within each application, the job paths within each job tree, and the jobs within each job path. Over time it builds a history or "experience" data for each job that contributes to the expectations for how long a path, tree and application is expected to run. It is very good at predicting when an application "will be late" early in the game.

Here's an application with 2 job trees, one of which has a subtree. The job trees are different shades of blue; the subtree within the darker blue tree also has a drop shadow. The solid and dashed red lines each trace a different path. There are 6 paths and 16 jobs in this application.

PCS has four sources of information:

• The CA 7 database: PCS captures the application-tree-path-job dependency structure; due out times, if entered; and, the actual jobs running on a specific day.
• The Application Management Database (AMD): Populated by the installation, you can add values to indicate the importance of the work, when it should be considered late, whether it should have early analysis, when to raise an alert, and other setting to control how PCS views and processes your workload.
• The SLM Policy: Populated by the installation, you provide information pertaining to the Production Services (PS) queue and Services Classes to be used for this workload. The PS queue and service classes are also used by the Production Service Groups, if deployed in your installation.
• The experience data: Captured automatically, PCS aggregates and uses this information to predict elapsed times, completion times, and if a job will be or is late.

Of course the existing sources that ThruPut Manager uses are still relevant when the job is processed by ThruPut Manager: job profile determined in the analysis process, JAL statements, JECL statements, binding and limiting agents and so on.

This section provides information about the following topics: