| With the widespread use of relational database, the functionality and performance of that has direct effects to the quality of information systems. Enhancing the database testing is of great and practical significance. Traditional manual testing methods cannot meet the practical needs, so it is necessary to develop and implement an automated testing framework against the database in order to realize the high efficiency of database testing through reducing the cost and increasing the quality of database products. For this purpose, this thesis, on the background of CUBRID database, studies to implement an automated testing framework which could be applied to the CUBRID database. The outcomes of the author’s work are as follow:(1) Adopted "ActiveMQ", a technology of message oriented middleware in open source frameworks, as the controller to initiate in a whole automation. This can largely increase the flexibility of test control and realize the design philosophy of High Cohesion & Low Coupling in any testing automation.(2) Selected "Shell" and "SQL", two readable, reusable and easy to maintain test scripts in test execution, designed and developed two functional testing frameworks for these test scripts; Designed and implemented a functional testing tool based on YCSB, to support automated performance test of CUBRID database; Designed and implemented an automated testing tool for memory leak detection on the basis of Valgrind, to support the CUBRID memory leak test and analysis of the problem.(3) Designed and implemented a framework which is applicable to show the CUBRID database test results so that the test results are easy to be viewed and counted, moreover, it also benefit to maintenance and upgrading of the testing framework.The automated testing framework implemented in this thesis can be applied to CUBRID database product, support the database functional testing tasks, as well as implement and control the automation of performance testing tasks. The automated testing framework has been proved by actual operation and meets expectations and has important reference meanings for the study of database testing automation. |
