Architectural-Level Risk Analysis FOR UML Dynamic Specifications

Risk assessment is an essential process in managing software development. Performing risk assessment in the early development phases enhances the resource allocation decisions. Several methodologies for risk assessment were developed, mostly based on subjective judgment. In this paper we use the Unified Modeling language (UML), and a dynamic metrics based methodology developed in earlier work. We use commercial tools such as Rose RealTime modeling and simulation environment to obtain simulation statistics for which an automated architectural-risk assessment environment can be developed. We use Microsoft Excel sheets and Macros in the development of this environment. The dynamic metrics based methodology is a heuristic risk assessment methodology that is based on dynamic complexity factors and severity analysis. Model execution is used for obtaining dynamic complexity and dynamic coupling measures for all architecture elements. Severity analysis is performed using Failure Mode and Effect Analysis. Heuristic risk factor for each architectural component is obtained. A component dependency graph is constructed and traversed to obtain the overall system/subsystem risk factor.