Synchronization Complexity Metric

As multitasking environments and applications become more and more common, the quality and efficiency of these applications becomes more and more important. While for single-threaded programs there are many complexity metrics in use since the 1970’s, so far there was no way to efficiently measure and compare the complexity of multitasking programs, and the affect of the concurrency on the overall program complexity. In this thesis we propose a solution to this problem. Namely, we introduce a new metric that characterizes program complexity based on the kind and amount of the various synchronization means used for coordination between its concurrent components. Similar to McCabe’s metric for single-threaded programs, the new Synchronization Complexity metric allows for assessment of the amount of tests needed to achieve a proper coverage in testing of a concurrent program. It also enables comparison between different system’s implementations based on the synchronization complexity analysis. SYNCHRONIZATION COMPLEXITY METRIC Peter Yastrebenetsky Page 3 Table of