n/a

n development processes with high code production rates testing typically triggers fault diagnosis to localize the detected failures. However, current test prioritization algorithms are tuned for failure detection rate rather than diagnostic information. Consequently, unnecessary diagnostic effort might be spent to localize the faults. We present a dynamic test prioritization algorithm that trades fault detection rate for diagnostic performance, minimizing overall testing and diagnosis cost. The algorithm exploits pass/fail information from each test to select the next test, optimizing the diagnostic information produced per test. Experimental results from synthetic test suites, and suites taken from the Software-artifact Infrastructure Repository show possible diagnostic cost reductions up to 10 and 19 percent, respectively, compared to the best of random selection, FEP, and ART. The cost reduction is sensitive to the quality of the test coverage matrices and component health, but tends to grow with the number of faults.

Although epidemic or gossip-based multicast is a robust and scalable approach to reliable data dissemination in distributed systems, its inherent redundancy may result in high resource consumption both on links and nodes. This problem is aggravated in settings that have costlier or resource constrained links, as happens in Cloud Computing infrastructures composed by several interconnected large data centers across the globe. The goal of this work is therefore to improve the efficiency of gossip-based reliable multicast used in infrastructure management systems by reducing the load imposed on those constrained links. In detail, the proposed CLON protocol combines an overlay that gives preference to local links and a dissemination strategy that takes into account locality. Extensive experimental evaluation using a very large number of simulated nodes shows that this results in a reduction of traffic in constrained links by an order of magnitude, while at the same time preserving the resilience properties that make gossip-based protocols so attractive.

This technical report is meant to report our findings and ideas with respect to spectrum-based fault localization and model-based diagnosis. In the following we want to introduce and compare model-based diagnosis (MBD), spectrum-based fault localization (SFL) and our contributions using 3-inverters as a running example (which is simple, yet sufficiently interesting).The remainder of this paper is organized as follows. The concepts and definitions used in this paper are given in the next section. The combination of model-based diagnosis and Bayesian reasoning, as it is normally applied to, e.g., digital circuits, is discussed in Section 3. Spectrum-based fault localization, including the system transformation for instrumentation to collect data to reason about failures is discussed in Section 4.1. In Section 4.2 we investigate several novel approaches for applying model-based diagnosis, and notably Bayesian reasoning to systems that have been prepared for spectrum-based fault localization.

This tutorial paper gives an overview of automated diagnosis applied to software faults. The emphasis is on a particular technique called spectrum-based fault localization, which is well-suited for diagnosing software systems, and which can easily be integrated with existing testing schemes. We discuss applications of the technique, including the specific application domain of embedded software, and provide pointers to recent research on factors that influence its diagnostic accuracy. In addition, we give instructions for quickly getting started with applying spectrum-based fault localization to existing projects.