Main Challenges In Automated Testing Of Digital App

By Author: Tomwade
Total Articles: 34
Comment this article

A financial or a banking application facilitates the management of business processes, automates, assists and stores financial information. It can be executed as a standalone software or as a part of a financial information system.

The main characteristics of a typical banking application are as follows:

- A multi-tier functionality to support thousands of concurrent sessions
Large scale integration

- Integration with other applications such as Bill pay, trading accounts, and other real time and batch processing with high rate of transactions.

Testing banking application requires an end to end testing solution involving multiple testing techniques, mainly to ensure the efficiency following characteristics-

- Full coverage of banking and business workflows

- Functional aspects of the application

- Application Security

- Data Security

- Concurrency management

- User Experience

- Application performance

Automated testing tools that are most effective in testing the system are most often limited ...
... while testing the complete system. The limitations are a result of the following situations:

- Establishing shared services that are used by many business groups and systems.

- Use of workflow technologies where the business logic is spread over a set of systems.

- Increased implementation of services and process management techniques which focus on reusing of business components.

- Increased integration of packaged solutions.

Challenges of Testing Digital Banking Apps

Testing the above interdependencies are very critical for core applications. The banking industry interfaces with a huge volume of customers and transactions. Additionally, there are multiple implementations across geographic regions. Changes in such systems need to be rigorously tested in orders to avoid downstream inconsistencies that could eventually lead to serious consequences.

Regarding regression testing, it is required to focus that the original functionality is not disturbed due to the introduction of new functionality or changes made in the system. Regression test needs to be performed in isolation assuming the other interconnected systems are not undergoing changes. As the consequences of the errors are high, hence the end to end testing of such systems are very costly. As such the tests cases are limited to basic conditions only.

Banking applications automate a number of processes and handle multiple product lines, multiple flow paths and with a vast amount of real time information. As a result, these systems are not suitable for the traditional test automation. Here, the core application logic needs to be tested for which record/playback feature is not enough. The record/playback feature offers only a stop gap solution and creating all the scripts for the application will only make it difficult to maintain in the face of changes to the application.

Using the concepts of functional decomposition and test asset reuse can significantly solve these problems. It involves reducing all the functions and process to their lowest unit of work, each defined by its set of parameters and attributes. Test cases can then be created using these atomic processes. Errors made in this situation result in inaccurate execution of the test cases. Also, the test data needs to be separated from the test case in order to ensure test reuse.

For a rigorous automated testing, the QA team needs to have exhaustive and comprehensive test cases. This requires considering all external inputs as independent parameters to generate test cases. In practice, exploring all combinations becomes impractical. This leads the QA team to define a set of cases which do not account for all input variations. A pairwise coverage is then implemented which is considered to be a more thorough approach to cover all critical cases, while keeping the number of test cases under control. For areas which are not covered by the pairwise approach, test cases need to be created manually.

Conclusion

Test automation has always been considered as a means to achieve the desired amount of consistency and coverage in the testing process. Clearly defining the boundaries of the system that needs to be tested and optimizing the test cases to a manageable number can provide significant benefits in carrying out the testing process more effectively.