Track Talk Th2

Shaping Testing into Models for Automated Test Case Generation and Execution

Minh Nguyen

09:00-09:45 CEST Thursday 9th June

In this track talk I will share our practical experiences of implementing model-based API testing of a complex and highly intensive rule-based Master Data Management system, at one of the biggest banks in Norway.

We started our journey to improve our test approach 3 years ago. The ultimate goal is to make API testing smarter and more effective in order to align to our agile development. Currently, we have managed to implement a transition from manual testing to automated, continuous (event-based), and lastly model based testing. During this talk we will share our experiences, benefits gained, and lessons learnt. We will focus on the latest achievement implementing model based testing by using ecFeed and SoapUI. The main benefits are:

  • Automatic generation of test cases
  • Flexible mechanism for adjusting test case coverage according to a given test purpose
  • Automatic test case execution and evaluation followed by export of test result
  • Cost effective maintenance of API tests e.g. requiring changes to the model and not to the API tests at business requirement changes

 

We will also share our gained experiences with respect to the model based test technique, implementation process and tool integration between ecFeed and SoapUI. If there’s time, we can mention our work in establishing a framework for automated API testing, which provides the following benefits:

  • API tests are executed independent of test environments and test data (on-the-fly test data retrieval)
  • Writing a single API test and executing it to verify multiple scenarios
  • Developing API tests “for reuse and with reuse”

 

Finally, we will briefly tell you about our next step in the journey, i.e. adaptation of AI/ML for analyzing generated test cases, and continuously gathering test results to identify potential risks or error-prone components, for smarter on-the-fly generation of new test cases.