421: Are We Done Yet? Certifying Application Performance Under New Database Versions or Architectures.
David A Haas , The Nielsen Company - Consumer Panel Services    Biography

Objectives: Insight into the issues with testing application performance on a complex decision support system.
A method for capturing real-world test cases to insure the results are relevant to the business.
Examples of a proven statistics based approach that compares performance before/after a change.
Abstract: Fear of the unknown can cripple a cautious development team and prevent them from making upgrades to the database or architecture necessary to maintain technological and business relevance. Lack of knowlege on the impact of changes can empower a reckless development team with the false confidence to make upgrades that render their application unusable. In the end neither of these options is viable in today's fast changing world.

Application Development Teams require proven techniques to evaluate the impact of a change, empowering them to make the calculated risks necessary for success and continued relevance.

In this presentation, I will share the methods used to evaluate the performance impact of an Oracle 8i to 10g upgrade (combined with underlying partition structure and block size changes), as well as a radical shift in database architecture (isolating static and dynamic data content into separate instances joined into a single virtual instance using DBlinks).

Audiernce Focus: Managers
Expertise: Intermediate
Track: Development
Date: Monday, April 14, 2008
Time: 3:30 PM - 4:30 PM