Multiscale navigation in large trace data

One solution to handle large data sets is to organize and visualize them in a hierarchical model. This paper investigates hierarchical management of large trace logs and proposes an interactive zoomable timeline view to visualize the multiple levels of information generated by analyzing the execution trace logs. The view supports both the semantic (data) zooming and physical (visual) zooming. It displays a coarser layer first and provides some operations to explore and navigate the different layers of data. The method mainly facilitates the comprehension of the execution trace logs and can also be used to improve root cause analysis. The paper also discusses the hierarchical data model designed for organizing and visualizing the information at multiple levels.