Automated tracing and visualization of software security structure and properties.

VizSec '12: Proceedings of the Ninth International Symposium on Visualization for Cyber Security(2012)

引用 20|浏览0
暂无评分
摘要
Visualizing a program's structure and security characteristics is the intrinsic part of in-depth software security assessment. Such an assessment is typically an analyst-driven task. The visualization for security analysis is usually labor-intensive, since analysts need to read documents and source code, synthesize trace data from multiple sources (e.g., system utilities like lsof or strace). To help address this problem, we propose SecSTAR, a tool that dynamically collects the key information from a system and automatically produces the necessary diagrams to support the first steps of widely-used security analysis methodologies, such as Microsoft Threat Modeling and UW/UAB First Principles Vulnerability Assessment (FPVA). SecSTAR uses an efficient dynamic binary instrumentation technique, self-propelled instrumentation, to collect trace data from production systems during runtime then automatically produces diagrams. Furthermore, SecSTAR allows analysts to interactively view and explore diagrams in a web browser. For example, analysts can navigate the diagrams through time and at different levels of detail. We demonstrated the usefulness of using SecSTAR to produce FPVA-style diagrams for a widely used and complex distributed middleware system, the Condor high-throughput scheduling system. Compared with the original manual approach in FPVA, SecSTAR shortened the initial diagram construction time from months to hours and constructed a more accurate diagram visualizing the complete runtime structure of Condor.
更多
查看译文
关键词
software security structure,fpva-style diagram,trace data,production system,middleware system,in-depth software security assessment,system utility,widely-used security analysis methodology,security analysis,security characteristic,condor high-throughput scheduling system
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要