Business process models created in the blueprinting phase of an SOA or ERP project can be leveraged for managing and improving the end-to-end integration, user acceptance and non-regression testing. The blueprinting exercise can deliver you much more value than just documenting and scoping of a new IT solution.
IT solutions should deliver comprehensive and qualified support for business processes of an organization. Three main challenging aspects for risk, quality and costs can be identified concerning the quality assurance of complex IT solutions:
- Not all relevant test cases are captured and no reliable quality metric exists
- Incomplete tests may have serious impacts on production environment.
- Untested scenarios and technical interfaces lead to failures in the productive ERP system.
- Untested changes of risk-related processes (e.g. SOX) could have critical consequences.
- Bad test coverage potentially leads to low a software quality and so may have a high business impact.
- An incomplete matching of the test data will increase the negative impact.
- Insufficient reviews of work products and deliverables add to the negative impact.
- The manual creation of test plans and case results in high project efforts.
- The traditional design of test cases blocks qualified and costly team members.
- Insufficient integration testing means low production efficiency and high maintenance costs
A typical process model contains different potential execution paths when going from the start to the end of a business process due to given logical rules inside a process model. The test plan and cases for the test of the business process can be reliable generated depending on quality metrics called “coverage” for the testing of process actions, the transitions between process actions and the process paths. In addition known execution probabilities of rules decisions may be used to compute the execution probability of a specific end-to-end path. This can be used to lower test execution efforts and to concentrate on most probable business scenarios e.g. by implementing test automation.
Our team developed a tool that can be feed with business process models from a modeling tool to generate test plans and cases to overcome the mentioned challenges. The number of generated test plans can be optimized depending on the selected testing metrics. Deliverables can have a format of Excel® or data that can be loaded into HP ALM/QC, IBM RQM, SAP® Solution Manager’s Test Workbench or other testing tools. Based upon these deliverables the integration, user acceptance or non-regression test may then be executed in the test tool.