Software architectural risk assessment

Risk assessment is an essential part of the software development life cycle. Performing risk analysis early in the life cycle enhances resource allocation decisions, enables us to compare alternative software architectural designs and helps in identifying high-risk components in the system. As a result, remedial actions to control and optimize the process and improve the quality of the software product can be taken. In this thesis we investigate two types of risk---reliability-based and performance-based risk. The reliability-based risk assessment takes into account the probability of the failures and the severity of failures. For the reliability-based risk analysis we use UML models of the software system, available early in the life cycle to come up with the risk factors of the scenarios and use cases. For each scenario we construct a Markov model to assess the risk factors of the scenarios and its risk distribution among the various classes of severity. Then we investigate both independent use cases and use cases with relationships, while obtaining the system-level risk factors. (Abstract shortened by UMI.)