As a result of performing IV&V (Independent Verification and Validation) on Space Station software, a number of interface faults were found during integrated testing or actual software deployment. Faults found at this late phase of the software development lifecycle are very expensive to fix. Other research indicates that significant cost savings can be realized if these types of faults can be discovered at earlier software development lifecycle phases, such as specification or coding. A need was determined for processes, procedures, and tools that will reliably identify interface faults during these earlier software development lifecycle phases. The ability to perform interface validation during earlier phases will reduce costly fixes due to interface faults discovered during later software development phases.
The approach is to leverage insight gained through analysis of real interface faults to develop analytical approaches for identifying interface faults in various types of artifacts, including specifications and source code. The unique contribution of this technology is the model-based representation of interface information that permits semi-automated analysis. Currently, interface analysis is largely manual and therefore time-consuming and expensive. Functional operation is to capture interface information from text-based artifacts into a model-based representation. Once in a model-based representation, consistency checks and flow analysis are performed to identify potential interface faults.
The technology has widespread applicability in any kind of software development that uses distributed applications. It is particularly applicable to critical real-time systems because interface faults in these systems can be catastrophic.
This work was done by Jack Smith of Goddard Space Flight Center and Pavan Rajagopal of Geocontrol Systems, Inc. GSC-16082-1