Research On Bugs Of TinyOS Programs And Implementation Of A Benchmark Suite

TinyOS is a widely used node operating system in Wireless Sensor Networks, which is implemented by nes C, a component-based programming language. TinyOS applications have the following characteristics which make them error-prone due to severe resource constraint of nodes: Access protection mechanism of the system is unavailable while accessing hardware,such as memory; Unsafe-type nesC programming language makes the applications fail to apply dynamic type detection mechanism; Interleaving execution of event-driven and tasks complicates the concurrent execution model of TinyOS applications.Research on testing, debugging technologies and tools of TinyOS programs are based on the characteristics analyses of the common bugs of TinyOS applications. Such analyses also provide important guidance for implementing high-quality WSNs applications. As a significant research work of software testing and debugging, researchers need to choose appropriate benchmark suites to evaluate the software testing and debugging technologies during empirical studies.Each of the exiting research on TinyOS testing and debugging is aimed at no more than three kinds of bugs, and in its empirical studies, usually no more than three subjects are used and the subjects are self-chosen without a benchmark. As a result, in order to promote the research of testing and debugging technologies, it is necessary to provide a comprehensive analysis of common TinyOS bugs and to build corresponding benchmark suites.In the thesis, the characteristics of 17 common TinyOS applications bugs are systematically analyzed and corresponding benchmark suites are implemented, in which 18 benchmark subjects are included. Firstly, common TinyOS program bug types and the corresponding features are systematically described, including causes, patterns and repairing strategies. For each type of bug, related benchmark subject in this paper is implemented, including the error edition, right edition and the corresponding test cases. The details on how to use the benchmark suite under the simulation environment of Avrora are also described, including how to trigger the bug and how to distinguish the error symptom of the subjects.