CSR: Medium: Systems Support for Scalable, Easy-to-Implement, and Multilingual Static Analyses of Modern Software

CSR：中：对现代软件的可扩展、易于实施和多语言静态分析的系统支持

• 批准号: 1763172
• 负责人: Ardalan Amiri Sani
• 金额: $ 119.97万
• 依托单位: University of California-Irvine
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2018
• 资助国家: 美国
• 起止时间: 2018-08-15 至 2023-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1763172&HistoricalAwards=false
• 关键词: CSR; Medium; Systems; Support; Scalable

Static program analysis has been widely used in academia and industry to find bugs, security vulnerabilities, and performance optimization opportunities. Supporting sophisticated analysis algorithms on large codebases has been a key challenge in the program analysis research for decades. This inability is the major factor that prevents analysis-based techniques from being widely adopted in industry. This project revisits this problem from a data-driven perspective and develops novel system solutions that can make static program analyses easy to implement, support large codebases, and support programs written in multiple languages.The project will develop a transformative approach with four elements: (1) develop disk-based out-of-core systems to parallelize and scale constraint-based path-sensitive analysis; (2) develop system support for flow-sensitive analysis by treating it as evolving graph processing; (3) develop a distributed system solution to SAT solving --- a 50-year old problem --- to enable SAT-based applications to solve larger problems with the resources available in modern computing; and (4) analyze the Android operating system and apps together to find complicated (a) bugs, (b) security vulnerabilities, and (c) performance problems, which involve interactions between the system and an app as well as multiple apps.The systems developed by the project will make precise static analysis algorithms more efficient and scalable, enabling them to process modern software programs that existing techniques could not analyze. Since these programs are used every day by many users and businesses, making them more robust and secure extends the benefit to a broad community. The project will develop big data systems for sophisticated code analysis, opening a new direction to scale program analysis. It will involve several PhD students and infuse research into the undergraduate and graduate curricula to train developers of the future.All the system implementations, experimental data, and documents from the project will be publicly available and maintained at https://www.ics.uci.edu/~guoqingx/research/projects/analysis.html .This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

静态程序分析已广泛应用于学术界和工业界，以发现错误，安全漏洞和性能优化机会。在大型代码库上支持复杂的分析算法一直是程序分析研究的一个关键挑战。这种能力是阻碍基于分析的技术在工业中广泛采用的主要因素。本项目从数据驱动的角度重新审视这一问题，并开发新的系统解决方案，使静态程序分析易于实现，支持大型代码库，并支持用多种语言编写的程序。本项目将开发一种变革性的方法，包括四个要素：（1）开发基于磁盘的核外系统，以并行化和扩展基于约束的路径敏感分析;（2）通过将流敏感性分析视为进化的图形处理来开发对流敏感性分析的系统支持;（3）开发一种分布式系统解决方案来解决SAT问题--一个50年前的问题--使基于SAT的应用程序能够利用现代计算中可用的资源来解决更大的问题;以及（4）一起分析Android操作系统和应用以发现复杂的（a）错误、（B）安全漏洞和（c）性能问题，这涉及到系统与应用程序以及多个应用程序之间的交互。该项目开发的系统将进行精确的静态分析算法更高效和可扩展，使他们能够处理现有技术无法分析的现代软件程序。由于许多用户和企业每天都在使用这些程序，因此使它们更加强大和安全可以为广泛的社区带来好处。该项目将开发用于复杂代码分析的大数据系统，为规模化程序分析开辟新的方向。它将涉及几个博士生和注入研究到本科生和研究生课程，以培养未来的开发人员。所有的系统实现，实验数据和文件，从项目将公开提供和维护在www.example.com。这个奖项反映了NSF的法定使命，并已被认为是值得的支持，通过评估使用基金会的智力价值和更广泛的影响审查标准。

项目成果

Dorylus: Affordable, Scalable, and Accurate GNN Training with Distributed CPU Servers and Serverless Threads

• 发表时间: 2021-05
• 作者: John Thorpe;Yifan Qiao;Jon Eyolfson;Shen Teng;Guanzhou Hu;Zhihao Jia;Jinliang Wei;Keval Vora-Keval-Vor

Accelerating Graph Mining Systems with Subgraph Morphing

• DOI: 10.1145/3552326.3567489
• 发表时间: 2023-05
• 期刊: Proceedings of the Eighteenth European Conference on Computer Systems
• 作者: Kasra Jamshidi;Harry Xu;Keval Vora

Systemizing Interprocedural Static Analysis of Large-scale Systems Code with Graspan

使用 Graspan 系统化大型系统代码的过程间静态分析

• DOI: 10.1145/3466820
• 发表时间: 2021-07
• 期刊: ACM Transactions on Computer Systems, (TOCS'21)
• 作者: Zuo Zhiqiang;Wang Kai;Hussain Aftab;Sani Ardalan Amiri;Zhang Yiyu;Lu Shenming;Dou Wensheng;Wang Linzhang;Li Xu;ong;Wang Chenxi;Xu Guoqing Harry
• 通讯作者: Xu Guoqing Harry

Gerenuk: thin computation over big native data using speculative program transformation

• DOI: 10.1145/3341301.3359643
• 发表时间: 2019-10
• 期刊: Proceedings of the 27th ACM Symposium on Operating Systems Principles
• 作者: Christian Navasca;Cheng Cai;Khanh Nguyen;Brian Demsky;Shan Lu;Miryung Kim;G. Xu

Understanding and automatically detecting conflicting interactions between smart home IoT applications

• DOI: 10.1145/3368089.3409682
• 发表时间: 2020-11
• 期刊: Proceedings of the 28th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering
• 作者: R. Trimananda;S. A. H. Aqajari;Jason Chuang;Brian Demsky;G. Xu;Shan Lu