Section 1.2. Requirements of a Good Method

   

1.2 Requirements of a Good Method

What distinguishes a good method from a bad one? When we started hotsos.com in 1999, I began spending a lot of time identifying the inefficiencies of existing Oracle performance improvement methods . It was a fun exercise. After much study, my colleagues and I were able to construct a list of objectively measurable criteria that would assist in distinguishing good from bad in a method. We hoped that such a list would serve as a yardstick that would allow us to measure the effectiveness of any method refinements we would create. Here is the list of attributes that I believe distinguish good methods from bad ones:

Impact

If it is possible to improve performance, a method must deliver that improvement. It is unacceptable for a performance remedy to require significant investment input but produce imperceptible or negative end- user impact.

Efficiency

A method must always deliver performance improvement results with the least possible economic sacrifice. A performance improvement method is not optimal if another method could have achieved a suitable result less expensively in equal or less time.

Measurability

A method must produce performance improvement results that can be measured in units that make sense to the business . Performance improvement measurements are inadequate if they can be expressed only in technical units that do not correspond directly to improvement in cash flow, net profit, and return on investment.

Predictive capacity

A method must enable the analyst to predict the impact of a proposed remedy action. The unit of measure for the prediction must be the same as that which the business will use to measure performance improvement.

Reliability

A method must identify the correct root cause of the problem, no matter what that root cause may be.

Determinism

A method must guide the analyst through an unambiguous sequence of steps that always rely upon documented axioms, not experience or intuition. It is unacceptable for two analysts using the same method to draw different conclusions about the root cause of a performance problem.

Finiteness

A method must have a well-defined terminating condition, such as a proof of optimality.

Practicality

A method must be usable in any reasonable operating condition. For example, it is unacceptable for a performance improvement method to rely upon tools that exist in some operating environments but not others.

Method C suffers brutally on every single dimension of this eight-point definition of "goodness." I won't belabor the point here, but I do encourage you to consider, right now, how your existing performance improvement methods score on each of the attributes listed here. You might find the analysis quite motivating. When you've finished reading Part I of this book, I hope you will revisit this list and see whether you think your scores have improved as a result of what you have read.


   
Top


Optimizing Oracle Performance
Optimizing Oracle Performance
ISBN: 059600527X
EAN: 2147483647
Year: 2002
Pages: 102

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net