How to Develop High-Performance OLTP system on Oracle
Summary: In the presentation we’ll use live demo to build a critical OLTP system, we’ll use 4 different approaches to implements the same system. The performance different between different approaches is over 100X times. During this interactive session, we will work together to identify inefficiencies and improve the system performance, base on the data from ASH and performance views. We’ll show various code technique which cause huge performance difference, and how to avoid common development error, some error might be very hard to debug and might bring down the OLTP system.
How to Develop High-Performance ETL and DW system on Oracle
Summary: In the presentation we’ll use live demo to show the modern way to process large data set, by using Oracle set-based technique. We’ll build up an ETL system to process hundreds of million rows. We’ll have 4 racers which cost the hardware resource and software license, each with different technique:
1. Legacy row-by-row (aka, slow by slow)
2. Array interface
3. Home-grow parallelism
4. Native set-based technique.
The ETL process includes data loading, de-duplication, data transformation and analytical query. In this interactive session, the audience will learn there are orders of magnitude performance gain, by using set-base technique like direct path loading, parallel execution, window function and outer join.