Identifying Data Inconsistencies Using After-State Database Testing (ASDT) Framework

Database applications retrieve data from database, process the data and then store them for future processing. If the data from database is incorrect or inconsistent with the application's business rules, the output of the application will be misleading. When such data is retrieved by the application for further processing, it may lead to incorrect results even though the program is correct. In this paper, we investigate methods in revealing inconsistent data stored in a database. We propose a framework to test the consistency of data with respect to the system specifications and design choices of the database applications. It is a black-box testing approach focusing on finding data inconsistency in database state after the execution of applications. We also build a prototype to demonstrate that our method is capable of handling different types of system rules, regardless of how they are implemented in the system.