Software reliability is defined as the probability of the failure-free operation of a software system for a specified period of time in a specified environment. Traditional approaches for software reliability analysis treat the software as a whole, use test data during software test phase and model only the interactions with outside world; these are known as black box models. Day-by-day the complexity of software application is growing and there is more emphasis on reuse. So, Component-Based Software (CBS) applications have come into existence. The black box models ignore the structure of the software made out of components and reliability of individual components, and thus are not appropriate to model the CBS. This has generated the research interest of architecture-based reliability. This paper introduces a reliability model and reliability analysis technique for architecture-based reliability evaluation.

Architecture-based software reliability analysis has been the current topic in research (Kubat, 1989; Krishnamurthy and Mathur, 1997; Gokhale and Trivedi, 1998; Gokhale et al., 1998; and Gokhale et al., 2004) as software applications have grown in size and complexity. The objective of this paper is to introduce a reliability model and reliability analysis technique for architecture-based reliability evaluation.

Architecture-based analysis technique can be of two kinds:

1. State-based approach

2. Path-based approach

In this model, the control flow between components has been taken into consideration. It is assumed that components failed independently. In these models, transfer between components has been considered as Markov behavior, which means that current behavior of components is independent of past behavior. From this model, architectural models using discrete Markov chain (Cheung, 1980) or semi-Markov processes have been proposed (Kubat, 1989). Using these models, the software reliability may be estimated by combining the behavioral description of the system with the component’s failure behavior.

The models proposed by Krishnamurthy and Mathur (1997) and Yacoub et al. (2004) assume that components fail independently. Reliability of systems by considering sequences of components executed along each path along with the path reliability by multiplication of their reliability can be estimated. According to these models:

Computer Sciences Journal, Minimum Spanning Tree, Software Applications, Software Reliability, Black Box Models, Component-Based Software Engineering, Interconnection Mechanisms, Kruskals Algorithm, Analytical Hierarchical Process, Software Models, Reliability Model.