Figure 9-12 was created using the same data as the graph in Figure 9-11. The only difference is that I included the numbers and used standard words (for example, "response time") and metrics (for example, "exec" for executions and "ms" for milliseconds). If asked why the SQL execution metric is relevant, I may respond that there is a CPU bottleneck, and in this system, the number of SQL statement executions directly impacts CPU consumption, which affects the response time. As I'll detail in later sections, you can also state that your proposed solutions are aimed at reducing the execution rate and the impact of each execution.

After you have shown and described an abstracted response-time curve like the one in Figure 9-11 or Figure 9-12, if your audience members are technical and will benefit from seeing real data, and you have multiple samples, then show them a graph containing real data, like the ones in Figure 9-9 and Figure 9-10. If you do show real data, be very well prepared to keep control of the presentation, because you will be peppered with questions, many of which will be irrelevant.

Figure 9-12. Shown is the same data as in Figure 9-11, but with slightly less abstraction. Notice I use more traditional words, such as "response time" and "arrival rate," and include numeric values. If you include technical words and numbers, be prepared to explain what they mean and how they relate to the performance situation.