Assessment of High Integrity Software Components for Completeness, Consistency, Fault-Tolerance, and Reliability

Abstract

Abstract. The use of formal model based (FMB) methods to evaluate the quality of components is an important research area. Except for a growing number of exceptions, FMB methods are still not really used in practice. This chapter presents two case studies that illustrate the value of FMB approaches for developing and evaluating componentbased software. In the first study, Zed (or Z) and Statecharts are used to evaluate (a priori) the software requirement specification of a Guidance Control System for completeness, consistency and fault-tolerance. The second study evaluates (post-priori) the reliability of a complex vehicle system using Stochastic Activity Networks (SANs). The FMB approach presented here provides further evidence that such methods can indeed be useful by showing how these two different industrial strength systems were assessed and the results. Clearly, future investigations of this nature will help to convince software system developers using component based approaches that such FMB methods should be considered as a valuable tool toward improving the software product lifecycle (quality, schedule and cost).

Measuring the constraint complexity of automotive embedded software systems

“…CBSS approaches prove that software component descriptions can be used for measuring semantic properties [5,6,7]. However, unlike business systems, a direct application of such component-based development approaches on software components in real-time systems is yet not possible [10] as deriving abstract system configuration boundaries using UML design tools, in order to define end-to-end timing constraints, is a complicated task [20,25,36,37]. Hence, the development of robust measurement techniques becomes inevitable for obtaining a precise estimation of constraint complexity at the specification level and thereby quantifying the dynamic nature of automotive software components [29,30,38], and requires the employment of sophisticated methods and heuristics.…”

“…From a component-based development perspective, the issue of software complexity measurement of software components has already been addressed in terms of their interface structure, interaction and constraints associated with usage [5,6,7,16,23,30,49]. Approaches dealing with constraint complexity [5,6,7] are however very scarce, and typically only cover the structural aspects by means of static inspection of the software artifacts [36,37], rarely taking temporal constraints into account which restrict the expected behavior of real-time systems at run-time [10,20]. This paper presents a method for measuring the structural as well as the timing constraint complexity of automotive embedded software systems at the specification level.…”

A survey of component based system quality assurance and assessment

“…Kim et al [31] presents a formal model based (FMB) approach to evaluate the quality of the components by using Zed (Z), State-charts and Stochastic Activity Network (SANs). The FMB frameworks are used to assess the component properties like completeness and consistency of requirement specifications using Z and state-harts; and approaches for verifying properties like reliability using stochastic modelling formalisms.…”

A Method for Measuring the Constraint Complexity of Components in Automotive Embedded Software Systems

“…Despite the availability of the event-driven AUTomotive Open System ARchitecture (AUTOSAR) methodology [1], which provides guidelines to automotive manufacturers on TIMing MOdeling (TIMMO) and the analysis of end-to-end timing constraints in order to maintain complexity at a manageable level [8,22], the complexity measurement of timing constraints remains a challenging task [26]. The more we are able to measure the complexity of a system, the more we can understand its associated reliability [17,35]. Figure 1 shows that similar to business systems, individual component descriptions of embedded software systems can be speci¯ed using Uni¯ed Modeling Language (UML) design tools [8,11,13,22].…”