Visualization analysis framework for large-scale software based on software network

Abstract

Large-scale software systems, which are the most sophisticated human-designed objects, play more and more important role in our daily life. Consequently effective analysis for large-scale software has become an urgent problem to be solved with the increasing issues of software security and the continuous expansion of software applications scope. For the characteristics of large scale and complex structure in large-scale software, the traditional software analysis techniques are difficult to be used. With the problem of difficulty in presentation, storage and low efficiency in the process of large-scale software analysis, the visualization analysis framework for large-scale software based on software network, named SoNet, is proposed with the combination of complex network theory and program slicing technique. Constraint logic attributes of the programs will be obtained through source code parsing. Then we will construct a global view by the theory of complex network after extracting software structure and behavior, improving user’s perception of software architecture in a macro perspective. Use case slicing will be realized combined with Redis cluster, and accessibility analysis when given a keyword to be analyzed. We evaluate our prototype implementation on an open source software project named SoundSea in Github, and the results suggest that our approach can realize the analysis for large-scale software.