- Overview: What's a Mix?
- Real-World Low-Level Technology Stack Test Input and Mixes
- Testing and Tuning for Daily System Loads
- Testing and Tuning for Business Peaks
- Identifying Key Transactions and Business Processes
- Real-World Access Method Limitations
- Best Practices for Assembling Test Packages
- SAP Component and Other Cross-Application Test Mix Challenges
- Tools and Approaches
9.4 Testing and Tuning for Business Peaks
Though the tuning made possible by daily expected workload testing will benefit you 99% of the time, you must still work through an ROI exercise designed to determine whether it's financially advisable to stress test the remaining 1% of scenarios. To an outsider, 1% might seem trivial, but if the 1% represents any of the following scenarios, the financial impact associated with planning for, setting up, configuring, executing, monitoring, analyzing, and tuning might easily be justified:
If you bring in a significant portion of your revenue via seasonal peaks
If you must process a time-sensitive workload in a particular execution window (e.g., 2 hours start to finish), such as payroll checks or tax-return data, the failure of which would produce financial or other significant losses
If specific SLAs are in place that penalize you significantly for "missing" key response-time or throughput metrics
If your business peaks represent an opportunity to compress financial cycles or make it possible to gain an edge on the competition in terms of insight into product sales, order fulfillment, inventory turns, and so forth.
If customer satisfaction is impacted beyond what is otherwise grudgingly acceptable
Business peaks vary nearly as much as the companies that endure them. As I mentioned before, almost every organization must suffer through some kind of seasonal peak workloadtesting and tuning for this peak can mitigate much of the suffering. Similarly, working through issues surrounding weekly, month-end, and quarter-end financial closings, warehouse inventories, payroll runs, HR benefit update windows, and so on will not only help you avoid self-imposed or externally driven SLA penalties but also provide a competitive long-term advantage. And like the value of knowing your daily load, understanding your business peaks positions you that much better when it comes time to perform a major functional upgrade or technology refresh.
Data input required for your business peaks should not be that difficult to identify, because it's almost always associated with a particular business activity or functional area, as we have already seen. But it's all about quantities! The key is to ensure you have enough data to do the job of stress testing, along with access to the proper number and mix of virtual or physical users that truly represent the concurrent online or batch load you seek.