Impact of statechart implementation techniques on the effectiveness of fault detection mechanisms

This work presents the analysis of an experiment series aiming at the discovery of the impact of two inherently different statechart implementation methods on the behavior of the resulting executables in the presence of faults. The discussion identifies the key features of implementation techniques influencing the effectiveness of standard fault detection mechanisms (memory protection, assertions etc.) and an advanced statechart-level watchdog scheme used for detecting the deviations from the abstract implementation-independent behavioral specification.

1. Introduction

When selecting components for dependable systems the dominant requirements are not only the low price and high performance but the dependability of the resulting application as well. For example in case of software for railway control designed according to EN-50128, if COTS software is to be used in higher software integrity levels, then the following precautions shall be taken: (1) an analysis of possible failures shall be carried out and (2) a strategy shall be defined to detect failures of the COTS software and to protect the system from these failures. Fault detection mechanisms are the substantial facilities for developing fault tolerant systems. As automatic code generation plays a more and more central role in modern software development methodologies, the impact of code generation patterns on the effectiveness of fault detection mechanisms is becoming an important issue. This paper presents an experiment series aiming at investigation of two different statechart implementation methods and identifies the key differences between behaviors of applications in the presence of faults.