Deriving Safety-Related Scenarios to Support Architecture Evaluation

This chapter introduces an analysis process that combines the different perspectives of system decomposition with hazard analysis methods to identify the safety-related use cases and scenarios. It argues that the derived safety-related use cases and scenarios, which are the detailed instantiations of system safety requirements, serve as input to future software architectural evaluation. Furthermore, by modeling the derived safetyrelated use cases and scenarios into UML (Unified Modeling Language) diagrams, the authors hope that visualization of system safety requirements will not only help to enrich the knowledge of system behaviors but also provide a reusable asset to support system development and evolution. Deriving Safety-Related Scenarios to Support Architecture Evaluation 33 Copyright © 2005, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of Idea Group Inc. is prohibited. Introduction This chapter defines a technique to identify and refine the safety-related requirements of a system that may constrain the software architecture. The purpose of the approach presented here is to provide a relatively complete set of scenarios that can be used as a reusable asset in software architectural evaluation and software evolution for safetycritical systems. For this purpose, we will identify, refine, and prioritize the safety-related requirements in terms of scenarios by using safety analysis methods. The resulting scenarios can serve as input to software architectural evaluation. By evaluating various kinds of architectural decisions against the input safety-related requirements, the evaluation approach will assist in the selection of an architecture that supports the system safety. The resulting scenarios are reusable during software evolution. By reusing those common scenarios, and hence the common architectural decisions, the cost of development and the time to market can be reduced. The objective of this chapter is to introduce a technique that: (1) Identifies and refines the safety-related requirements that must be satisfied in every design and development step of the system, (2) instantiates the nonfunctional requirement – safety – into misuse cases and misuse scenarios that are further modeled by UML, and (3) provides a reusable asset – utility tree – that may either support the engineering decision making during software development or become input to future software architectural evaluation.