A three-dimensional visualization tool for software fault analysis of a distributed system

We propose a software visualization tool named Software Visualization Supporting Space (SVSS) utilizing three-dimensional (3D) graphics in order to detect and analyze software faults in a large-scale distributed system. When a failure occurs in a software testing process, it is necessary for fault detection to inspect a large number of software processes from various viewpoints. 3D graphical representations allow a greater quantity of data as well as complex structures and relationships between components to be displayed on the screen effectively. SVSS can analyze the source code regularly to generate structural data of the target software. Moreover, it can also analyze the behavior of the processes using trace data obtained automatically by embedding a trace data acquisition function in the target machine. The results of these analysis functions can be visualized collectively on a common display so that the developer can trace execution flows and static connections simultaneously. The work reduction ratio using SVSS was estimated by sampling faults that had occurred in practical development of a distributed power control system.