3rd Increment – MIDAS Version 0.3
The objectives of the 3rd increment were the following.
- All scenarios for the three major MIDAS use cases running in principle, although not 100%.
- Usage of partially implemented pilots for verification.
- Partial adoption of the MIDAS DSL.
- Further iterations of integration steps 4 and 5 are completed.
The developed prototype met all objectives. The constraint validation for the MIDAS DSL was moved into a separate service on the MIDAS platform. The DSL itself was fully specified and the constraints as well as the UML profile partially implemented. Furthermore, the generator for TTCN-3 code was implemented for functional, usage-based, and behavioral fuzz test cases. The components for test execution were completed, but required additional effort than initially planned, due to several bugs and anomalies that occurred when working with the TTworkbench platform. This required several iterations with Testing Technologies in order to get stable wrapper components on the MIDAS platform.
The test method developers extended their functionalities from the mock-ups that were available in the 2nd increment. The components for usage-based testing were completed without support for the DSL. Usage monitoring facilities were deployed with both pilots and the results were used to train first usage profiles, generate tests and calculate usage scores. However, the generated tests were still only with available in internal formats and not compatible with the MIDAS DSL. Similarly, it was only possibly to calculate the usage scores, the definition of a MIDAS DSL compliant scheduling was not yet finished. The security testing was partially implemented. The TTCN-3 adapter for the data fuzzing library was partially implemented with some details missing. Similarly, the behavioral fuzzing was also partially implemented. The components for functional test generation were under heavy development for this increment, and considerable effort was required for their development, i.e., the underlying research and the definition of a machine readable model based on an XML specification. These efforts, together with unexpected problem during the development of the execution component, caused the delay of the increment.
Additionally, the dynamic orchestration of test methods through a generic Gen&Run component was made available to all partners in August, 2014 and was thus, also already achieved in the 3rd increment and ahead of schedule.
This increment was released on September 9th.