Operations

With 50 percent or more of all critical data processing being performed in batch, the demand for batch processing is clearly growing. And there’s no doubt that batch processing is a key aspect of real-time online systems. From generating the invoices that collect revenue, to reconciling customer accounts for online access, delays in the batch cycle directly and negatively impact revenue.

IT must meet batch processing commitments to keep the business running. However, online systems require longer availability time and new business services and growth create additional data for processing. In addition, constant maintenance or application changes can significantly alter the batch resource requirements, all of which increase the pressure on the batch window.

IT typically has two choices to address changing batch needs. One option is to tune batch jobs to reduce resource requirements. This option is so time- and resource-intensive that many shops simply don’t tune batch. As a result, most organizations choose a second option—they meet batch commitments by acquiring more hardware. With analyst estimates of total cost/MIPS exceeding $9,500 for hardware, software and administration, these upgrades can lead to unplanned costs and substantial impacts on the IT budget.

Using this estimate, even adding a single z9EC engine (~580 MIPS) can result in total first-year costs of more than $5.5 million. With costs of this magnitude, companies can develop significant IT savings by focusing on how they meet their batch needs. But how can they obtain these costs savings? The following are five techniques that IT organizations have successfully used to meet batch commitments and still control costs.

1. Automatically reduce the time required to process work: Static settings that govern batch jobs usually fail to deliver optimal performance, and the high cost of rewriting old code or modifying JCL forces many jobs to continue to function—inefficiently—night after night, year after year.

Software optimization tools reduce job elapsed time and thereby reduce the time used in the batch window. This can enable IT to meet service level agreements for batch and online processing and accommodate growth and application change—without frequent unplanned upgrades.

2. Automatically eliminate costly reruns caused by spacerelated job failures: Recovering from space-related ABENDs, often the cause of reruns and batch delays, continues to be a challenge, even with additional functions available in the IBM DFSMS (SMS) product. Storage management tools can proactively and automatically prevent or recover from ABENDs with no application changes and enable jobs to complete successfully the first time.

Storage capacity expands while the storage management staffing level remains static. To address this, IT can automate the recovery of space shortages to ensure availability, reduce delays, and enable batch to meet business deadlines without costly upgrades. qq3. Easily and accurately pinpoint and tune bottlenecks constraining batch: New and growing applications plus hardware and software changes continually impact the batch environment. As a result, resource bottlenecks arise that decrease the efficiency of batch applications and create delays. qqPerformance management tools can proactively identify the source of bottlenecks and test-drive alternatives before making changes to ensure the problem is solved. These tools can pinpoint problems in processors and I/O, and detect interference from other applications, LPARs, or systems.

4. Identify the largest resource-consuming jobs and pinpoint inefficient code: Millions are wasted each year by inefficient and poorly performing applications. Detailed application performance analysis is useful for spotting and tuning key resource-consuming jobs. DB2 applications in particular can introduce inefficiencies that lead to upgrades. Maximizing SQL statement and DB2 object efficiency can improve batch application run-times and reduce the need for upgrades.

5. Automatically schedule and manage batch processing cycles: A comprehensive batch scheduling solution for the enterprise can provide a single automation methodology for all platforms while helping reduce IT operational costs. Automated scheduling can incorporate cross-application/platform scheduling capabilities, workload balancing and status-based job execution to prevent scheduling problems from developing into business problems, to enable companies to comply with regulatory requirements, and to avoid higher IT costs.

Summary

Fortunately, there are alternatives to using upgrades to meet batch commitments. These options deliver better and more consistent batch performance without the high cost of repeated upgrades or service failures. IT organizations should explore how these solutions can reduce the cost of batch processing. Z