Section 4.2. How to Find the Economically Optimal Performance Improvement Activity

4.2 How to Find the Economically Optimal Performance Improvement Activity

By now, you've probably noticed that a central theme of Method R is careful targeting . How you respond to your diagnostic data fits the theme. Your response consists of three targeting steps:

First, the analyst targets the user actions for which performance improvement provides the best potential economic value to the business (Chapter 2).
Next , the analyst targets the correct time scope and action scope for diagnostic data collection (Chapter 3).
Finally, the analyst targets for implementation the performance improvement activity with the best expected net payoff (Chapter 4).

As I described in Chapter 1, the data format that best facilitates this third act of targeting is the resource profile . However, a resource profile is only part of the information that you'll need. After you learn from your resource profile where a user action's time went, your next step is to mine the diagnostic data to determine why a targeted component of response time took so long. Fortunately, if you have collected well-scoped extended SQL trace data, then everything you'll need is already in your possession. If you have not collected well-scoped extended SQL trace data, then you probably shouldn't have escaped the bounds of Chapter 3 just yet.

Once you have collected properly time- and action-scoped diagnostic data for each of your few (one to five) targeted user actions, the following algorithm combines the technical and financial analysis elements required to reveal the optimal action for your business:

For each targeted user action:
1. Assemble your diagnostic data into a format that helps you attribute root causes to elements of the action's response time consumption.
2. Estimate the net payoff of the best few options for performance improvement. Add the option and its expected net payoff to the list of candidate performance improvement activities.
From the list of candidate performance improvement activities, determine which activities will provide the best net payoff for the business.

When you've performed these steps, you've identified the performance activities that will most benefit your business. The next thing you'll do is set this book down and go convert your performance improvement plan into reality.

Method R is a significant departure from conventional "tuning" approaches. Method R is not about chasing down a list of ratios or even wait events that look suspicious. It's about aligning the priorities of performance improvement with business need . Your business drives prioritization decisions in an optimized performance improvement project, not your technology.

Top