Automated Reliability Analysis of Redundancy Architectures Using Statistical Model Checking

Hongbin He, Hongyu Kuang, Lin Yang, Feng Yang, Qiang Wang, Weipeng Cao

Abstract: Reliability is a fundamental property for mission and safety-critical systems, and adopting redundancy architectures is a common and prominent practice to increase system reliability. This paper proposes a novel approach for the modeling and quantitative reliability analysis of redundancy architectures based on the SBIP framework. Our approach supports modeling the nominal system behavior and the system faults in a unified formal model, which can be further integrated into the rigorous component-based system design paradigm advocated by BIP. We also propose two categories of metrics for formal reliability evaluation of redundancy architectures in terms of whether the system can operate correctly or provide reduced functionalities in the presence of faults. We take a computation unit as the running example and apply the proposed approach to analyze static redundancy and dynamic redundancy, which are Triple Module Redundancy architecture and Cold Standby architecture respectively. The experimental results show that our approach can accurately model various redundancy architectures and provide a comprehensive analysis of reliability and related properties in an automated manner. Moreover, our approach can be easily extended to a wide range of fault types and behaviors.