基于改进系统依赖图的影响域分析方法
Impact Domain Analysis Method Based on Improved System Dependence Graph
王翊泽 1金大海1
作者信息
- 1. 北京邮电大学计算机学院网络与交换技术全国重点实验室，北京 100876
- 折叠
摘要
软件系统的变更会破坏系统稳定性,因此需要通过影响域分析算法评估变更可能引发的风险,以降低潜在故障.传统基于系统依赖图（SDG,System Dependence Graph）的分析方法虽然能捕捉代码间基本依赖关系,但对复杂结构变量（如结构体,数组,指针）的依赖表示精度不足,易导致影响域过度估计.本文在传统SDG的基础上引入复合数据结构依赖弧和指针依赖弧,提出PC-SDG（Pointer-Composite System Dependence Graph）模型,通过构建复合类型内部依赖与指针访问路径,提升依赖关系的完整性和精确性.进一步基于PC-SDG设计了影响域分析算法.实验结果显示,提出的方法相较于已有的影响域分析算法结果集减少16.45\%,表明该方法能够有效提升依赖关系表示以及影响域分析算法精确性.
Abstract
Changes in software systems can compromise system stability, necessitating the use of impact domain analysis algorithms to assess the risks of potential failures triggered by modifications. Traditional analysis methods based on Program Dependence Graphs (PDG) can capture basic inter-code dependencies but lack precision in representing dependencies involving complex structural variables (such as structs, arrays, and pointers), often leading to overestimation of the impact domain. This paper introduces composite data structure dependence arcs and pointer dependence arcs into the traditional PDG, proposing the PC-SDG model to explicitly construct internal dependencies of composite types and pointer access paths, thereby enhancing the completeness and accuracy of dependence relations. Building on PC-SDG, this paper further designs an efficient impact domain analysis algorithm. Experimental results show that the result set of the proposed method is reduced by 16.45\% compared to existing impact domain analysis algorithms, demonstrating that this approach effectively improves the representation of dependence relations and the precision of impact domain analysis.关键词
计算机软件与理论/影响域分析/系统依赖图/复合数据结构Key words
Computer Software and Theory/Impact Domain Analysis/System Depedence Graph/Composite Data Structure引用本文复制引用
王翊泽,金大海.基于改进系统依赖图的影响域分析方法[EB/OL].(2026-03-12)[2026-03-15].http://www.paper.edu.cn/releasepaper/content/202603-111.学科分类
计算技术、计算机技术
评论