Model-based analysis is a well-established approach to assess the influence of several factors on system availability within the context of system structure. Prevalent availability models in the literature consider all failures to be equivalent in terms of their consequences on system services. In other words, all the failures are assumed to be of the same level of severity. In practice, failures are typically classified into multiple severity levels, where failures belonging to the highest severity level cause a complete loss of service, while failures belonging to levels below the highest level enable the system to operate in a degraded mode. This makes it necessary to consider the influence of failure severities on system availability. In this paper we present a Markov model which considers failure severities of the components of the system in conjunction with its structure. The model also incorporates the repair of the components. Based on the model, we derive a closed form expression which relates system availability to the failure and repair parameters of the components. The failure parameters in the model are estimated based on the data collected during acceptance testing of a satellite system. However, since adequate data are not available to estimate the repair parameters, the closed form expressions are used to assess the sensitivity of the system availability to the repair parameters

This paper describes research performed to augment the capabilities of the statistical modeling and estimation of reliability functions for systems (SMERFS) software suite. Enhancements to the Schneidewind software reliability model, which features optimal selection of failure data for predicting failure counts, were made to enable clearer interpretation of that model's results. These enhancements included the addition of risk assessment measures and modifications to the model's optimal point selection procedure. An analysis of an actual data set from a NASA Goddard Space Flight Center project that was used to validate these additions to SMERFS is included.

Exhaustive testing of computer software is intractable, but empirical studies of software failures suggest that testing can in some cases be effectively exhaustive. We show that software failures in a variety of domains were caused by combinations of relatively few conditions. These results have important implications for testing. If all faults in a system can be triggered by a combination of n or fewer parameters, then testing all n-tuples of parameters is effectively equivalent to exhaustive testing, if software behavior is not dependent on complex event sequences and variables have a small set of discrete values.

Software plays an increasing role in the safety critical systems. Increasing the quality and reliability of the software has become the major objective of software development industry. Researchers and industry practitioners, look for innovative techniques and methodologies that could be used to increase their confidence in the software reliability. Fault tree analysis (FTA) is one method under study at the Software Assurance Technology Center (SATC) of NASA's Goddard Space Flight Center to determine its relevance to increasing the quality and the reliability of software. This paper briefly reviews some of the previous research in the area of software fault tree analysis (SFTA). Next we discuss a roadmap for application of the SFTA to software, with special emphasis on object-oriented design. This is followed by a brief discussion of the paradigm for transforming a software design artifact (i.e., sequence diagram) to its corresponding software fault tree. Finally, we discuss challenges, advantages and disadvantages of SFTA.