Assurance Technology for High Confidence Software and Systems

This research will develop a feasible and applicable assessment environment, called MOCHA (MOdel-based Certification for High Assurance), that assists in control, composition, management, and documentation of different assurance activities conducted bydifferent groups and evidences collected from multiple sources during the entire life cycle of a real-time, embedded, high-confidence software system. MOCHA uses UMLas a base model of a software system and creates such an open and integrated environment byextending a popular COTS CASE tool, Rational Rose RealTime, with additional evidence organizing and composing features. The MOCHA data repository uses the emerging XML documentation technology as the common repository format for managing and organizingmodel and evidence data of a software system. MOCHA also provides a unified interface mechanism to support transformation of the base model to other models so that it can be assessed with external tools and that corresponding results can be fed back to theMOCHA data repository. As such, different models and assessment results can be easily shared, exchanged, reused, and manipulated among different tools and during different stages of a software life cycle. AS a proof-of-concept, the technology will beapplied into flight control software.The use of software-intensive systems in safety critical applications increasing. The combined complexity poses a significant challenge on building and verifying high confidence software systems. As recognized byPresident's Information Technology Advisory Committee,