Data-centric Testing

Enterprise Information Management SIG: Data-centric Testing - Why is it different? 16/10/2012 Presenters: Richard Ng This event topic was about Data-centric Testing. Data-centric Testing is based on data warehouse or business intelligence system. Unlike transaction system, its test scenario is totally different. To begin with, Richard stated the testing myths and realities. It can not be ensured that all the faulty of the system are corrected by the testing. Testing is divided into functional test, usability test, performance test, stress test, recovery test, security test, regress test. Different test scenarios should be used in specific purpose. Take functional system for example, it is clear that testing is to ensure the transaction goes fluently in a designed user case. Then he moved to the key challenges in the data centric testing. Compared with transaction, data centric testing involved a large amount of entities and relationships. Besides, it was not sequential and was difficult to define a use case for it. This is partly due to the fact that data centric testing involves numerous data in present layer. Also, it is subject oriented. This means the tester should have enough about the enterprise data and understand about the content in the report or decision support system. Finally, he gave us a solution in regard to this problem. Specific skills are necessary in carry out success testing, which includes data model aware, DW-BI architecture knowledge and enterprise data or at least data subject area understanding. In addition, a data scenario should be a good approach in testing. In terms of testing the front-end, functional testing must necessarily involve the end users since they can more easily detect abnormality in data. Overall, I think this presentation is very inspiring in data centric testing. I will use this knowledge in my recent business intelligence project.