Alessandro Bacioccola’s research while affiliated with University of Florence and other places

Software plays an increasingly important role in complex systems, especially for high-tech applications involved in important fields, such as transportation, financial management, communication, biomedical applications and so on. For these systems, performances such as efficient operation, fault tolerance, safety and security have to be guaranteed by the software structure, whose quality in use is assuming a growing importance from the industrial point of view. The basic problem is that the complexity of the task which software has to perform has often grown more quickly than hardware. In addition, unlike hardware, software cannot break or wear out, but can fail during its life cycle (dynamic defects) [1]. Software problems, essentially, have to be solved with quality assurance tools such as configuration management, testing procedures, quality data reporting systems and so on [2]. In this context, the paper proposes a new approach concerning the automated software testing as an aid to maximize the test plan coverage within the time available and also to increase software reliability and quality in use [3]. In this paper a method which combines accelerated automated tests for the study of software regression and memory overflow will be shown, in order to guarantee software with both a high quality level and a decrease of the testing time. The software will be tested by using test sequences reproducing the actual operating conditions and accelerated stress level. Moreover the research wishes to define some parameters of the software life and to show the generality of the proposed technique.

Software plays an increasingly important role in equipment and systems, both in terms of technical relevance and of development cost (often higher than 50% even for small systems). Unlike HW, SW does not go through a production phase. Also, SW cannot break or wear out. However, it can fail to satisfy its required function because of defects which manifest themselves while the system is operating (dynamic defects). A fault in the SW is thus caused by a defect, even if appears randomly in time, and SW problems are basically quality problems which have to be solved with quality assurance tools (configuration management, testing, and quality data reporting systems). In order to grant a software high quality level against a reasonable cost, the testing planning phase has to be study in detail. In order to do so, an adequate coverage of the product functionality has to be supplied to reduce the test time. The aim of the paper is suggesting an automated software testing as a solution to the problem of having to maximize the test plan coverage within the available time and to increase software reliability and quality in use.

In order to guarantee a software with high quality level vs a reasonable cost, the testing planning phase has to be study in detail, also to reduce the test time [1]. This paper introduces the result of the application of software automatic test to one of the most common software failure causes: the overflow due to memory leaks. The analysis carried out on a single software failure cause wants to show the total generality of the technique and to define some parameters of life of the software. The general value of the method is shown applying the technique to different software and for everyone of them defining the mean time to failure, where with failure we mean only the failure due to overflow; even if the overflow it isn't the only analyzable failure cause through the application of such software automatic test.