Reliability prediction through system modeling

Abstract

Quantifying software reliability, such as performance and dependability, through stochastic behavior models (or labeled transition systems) is already a common practice in the software analysis community. However, those models are usually too fine grained to represent an accurate view of the software system by its stakeholders. Scenarios, on the other hand, are capable not only to describe the system traces as behavior models do but also depict very clearly the system components designed to provide the intended system behavior as well as to outline a high level architecture view of the system being described. In this paper, we introduce a case study of a safety critical computer based system that is running in an Indian Nuclear Power Plant. We define clear component interfaces, from which we analyze its software reliability.