Hotsos Symposium Speaker – Felix Castillo

Felix Castillo is a new Hotsos Symposium presenter.

Biography

Felix CastilloFelix Castillo has been working with Oracle since 1991. At that time, he was responsible for OS, database, network, users and development. His focus drifted more and more to Oracle administration and performance tuning.

Since 2003, he also works as a trainer for different companies teaching Oracle topics.

He presented the results of his research at the last German Oracle Users Group conference (DOAG) in 2011, DOAG Regio Conference 2012, and will speak about his topic at the DOAG 2012 as well as the newly created database conference "Frankfurter Datenbanktage" in 2013.

Presentation Title

The Oracle Performance Signature

Abstract

Every database system has a specific way of being used by applications. After collecting and analyzing many systems over the last few years, and diving into the evaluation of performance data with graphical analysis, Felix has discovered that these patterns are application-specific. They can be identified and analyzed, and they allow us to make deductions on system performance, which can help us to understand why statements, jobs, etc., sometimes take longer.

There are many different ways to process performance data and to prepare them graphically. Every graph shows other aspects, which helps us to understand the system better, and therefore, helps us to better interpret AWR/ASH Reports.

The collection of graphs and mathematical results to describe identified patterns form the Oracle Performance Signature. With this signature we can compare systems, compare time periods, monitor and forecast performance.

Using Oracle-independent collecting and storing scripts to maintain a performance data warehouse not only gives us the power to analyze performance with significantly better accuracy and minimal impact, it is also version independent and license-free. Furthermore, it can be combined with performance data like SAR, iostat, etc. It can also be used for performance forecasting and with machine learning algorithms, the process of pattern recognition can be automated and integrated into an automatic performance monitoring.

The emphasis of this presentation is on showing examples of how performance data can be collected and presented graphically, and to help us understand the way a system is used. In addition, it will show us how the Performance Signature can be used to monitor and forecast performance.

Presentation Materials

Presentation materials are available online to attendees only.

Schedule

The speaker schedule is as follows: