WLP can study the effects of changing the processing performed on a particular day or even a particular shift. For example, a data center manager can use the History Management report to locate a log history file containing run information for the time frame in question. With this file used as input, WLP simulates a previous time frame run taking abends, reruns and on-request work into account, and creates workload planning reports. The reports are a key tool in studying processing alternatives. They provide job and resource performance data for the particular time frame.
The manager can now begin to study, for example, the effects of omitting one job from the run or adding a tape drive to the available resource pool.
A simulated workload model can be produced, using CA WA CA 7 Edition database job and resource data reflecting the same time frame. The online FWLP command extracts this data from the database and creates a file that can be edited before a simulation run. Once the file has been edited, WLP can be used to simulate and report on the run. The resulting reports, when compared to those reports produced from log history data, project what effect changing the workload or the processing objectives has on production work flow and resource use.
WLP is also helpful in establishing a proper balance between production processing, testing, and stand-alone time such as Preventive Maintenance (PM) or dedicated test time. CA WA CA 7 Edition does not capture or report on testing or PM. Workload planning reports on production requirements can, however, identify what time and resources are not available for testing and stand-alone time. Decisions on when and how much time and resources to provide programmers, customer engineers, and so forth, can be easily and more accurately made based on the slack time identified by these reports.
Data center management is continuously involved in the process of finding an optimum balance between job work flow and resource use. Ultimately, the goal of a data center is to run every job or application on schedule while maximizing use of the available resources. WLP is designed to provide a means of reaching and maintaining this goal. Automated Performance Analysis (APA) graphs show how many jobs are late or early. WLP lets managers simulate specific alternatives in job schedules or processing environments to achieve better throughput. WLP shows how you can reasonably expect work flow to occur when, for example, a group of jobs that have historically run late are submitted earlier.
If the jobs and available resources remained constant, the task of balancing work flow requirements and resource use is relatively simple. In most cases, however, the demands on data centers are continually growing. Growth can be significant although no new applications are being implemented. Manufacturers are constantly adding new parts to inventory. Banks are constantly adding new customers to their customer files. Payroll files grow as companies gain new employees. In each case, the elapsed time of jobs using these files can increase significantly.
Eventually, such increases in production workload can require hardware upgrades like faster CPUs or more tape drives and initiators. Occasional use of WLP helps prevent these subtle increases in demand from suddenly exhausting the available resources.
CA WA CA 7 Edition workload balancing provides real-time balancing of production work to yield optimum use of resources while monitoring work flow through completion and delivery deadlines. Processing objectives, as defined to workload balancing, are subject to change as new requirements or environments occur. Changes to these definitions can be simulated with WLP before implementing the changes in the production environment. Assume, for example, new peripherals or a new CPU can be expected to cause all jobs to run 10 percent faster. WLP can simulate this using a user-supplied elapsed time factor. An Hourly Usage Projection report can show that, under normal circumstances, the use of one type of tape drive approaches maximum capacity at all times. If requirements for this type of drive are expected to increase, WLP can simulate the use of additional drives.
Typically resources available for production batch processing change on a shift-to-shift basis. In the daytime, for example, a majority of CPU resources are allocated for processing online transactions. At night, a reduction in the number of these transactions frees up available resources. The freed resources enable an entire online system such as IMS to be shut down, and thus providing more resources for batch processing. Workload balancing can handle these environmental changes in real time whether they are scheduled or not scheduled. Simulating their impact on overnight batch production processing, however, assists a data center in developing and planning procedures for handling the changes. This simulation is true for unexpected, last-minute requests to keep up the online system beyond the normal scheduled time.
|
Copyright © 2015 CA Technologies.
All rights reserved.
|
|