Tiantian Zhu , Wenya He , Tieming Chen , Jiabo Zhang , Mingqi Lv , Hongmei Li , Aohan Zheng , Jie Zheng , Mingjun Ma , Xiangyang Zheng , Zhengqiu Weng , Shuying Wu
{"title":"LinTracer: An efficient tracking system for cyberattack chains fusing entity and event semantics","authors":"Tiantian Zhu , Wenya He , Tieming Chen , Jiabo Zhang , Mingqi Lv , Hongmei Li , Aohan Zheng , Jie Zheng , Mingjun Ma , Xiangyang Zheng , Zhengqiu Weng , Shuying Wu","doi":"10.1016/j.cose.2025.104413","DOIUrl":null,"url":null,"abstract":"<div><div>With the rapid development of information technology, advanced persistent threat (APT) attacks are becoming increasingly prevalent. This form of attack is known for its persistence, diversity, and stealth, and it results in serious security threats and economic losses for various organizations and institutions. In the face of this threat, tracing the attack chain (i.e., attack investigation) is critical to understanding the attacker’s behavior, identifying attack methods and patterns, and taking appropriate defensive measures. However, the current APT attack investigation techniques suffer from insufficient audit log refinement, attack entrance location difficulties, and attack path tracking accuracy challenges. In this paper, we propose LinTracer, which is an efficient attack investigation system based on the ATT&CK attack model for Linux systems that fuses entity and event semantics for cyber-attack chains. First, an auditing mechanism is used to stably collect the kernel data of the target operating system, and data compression techniques are used to refine the log data and reduce the overhead imposed by the attack investigation system. Second, a backward causal analysis is performed from the alarm point to construct a suspicious provenance graph. LinTracer extracts the features used to distinguish between attack events and benign events, calculates the feature scores of the events, and then uses the backward propagation algorithm to propagate the dependency scores backward from the alarm point to identify the attack entry points. Finally, entity semantic labels are designed based on the ATT&CK framework to perform forward label propagation on the attack entry points, ultimately enabling an effective attack investigation. The experimental results derived from laboratory tests and DARPA Engagement (approximately 64 million auditing events obtained from real systems) show that LinTracer has good real-time performance and can accurately identify attack chains.</div></div>","PeriodicalId":51004,"journal":{"name":"Computers & Security","volume":"154 ","pages":"Article 104413"},"PeriodicalIF":4.8000,"publicationDate":"2025-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Computers & Security","FirstCategoryId":"94","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0167404825001026","RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"COMPUTER SCIENCE, INFORMATION SYSTEMS","Score":null,"Total":0}
引用次数: 0
Abstract
With the rapid development of information technology, advanced persistent threat (APT) attacks are becoming increasingly prevalent. This form of attack is known for its persistence, diversity, and stealth, and it results in serious security threats and economic losses for various organizations and institutions. In the face of this threat, tracing the attack chain (i.e., attack investigation) is critical to understanding the attacker’s behavior, identifying attack methods and patterns, and taking appropriate defensive measures. However, the current APT attack investigation techniques suffer from insufficient audit log refinement, attack entrance location difficulties, and attack path tracking accuracy challenges. In this paper, we propose LinTracer, which is an efficient attack investigation system based on the ATT&CK attack model for Linux systems that fuses entity and event semantics for cyber-attack chains. First, an auditing mechanism is used to stably collect the kernel data of the target operating system, and data compression techniques are used to refine the log data and reduce the overhead imposed by the attack investigation system. Second, a backward causal analysis is performed from the alarm point to construct a suspicious provenance graph. LinTracer extracts the features used to distinguish between attack events and benign events, calculates the feature scores of the events, and then uses the backward propagation algorithm to propagate the dependency scores backward from the alarm point to identify the attack entry points. Finally, entity semantic labels are designed based on the ATT&CK framework to perform forward label propagation on the attack entry points, ultimately enabling an effective attack investigation. The experimental results derived from laboratory tests and DARPA Engagement (approximately 64 million auditing events obtained from real systems) show that LinTracer has good real-time performance and can accurately identify attack chains.
期刊介绍:
Computers & Security is the most respected technical journal in the IT security field. With its high-profile editorial board and informative regular features and columns, the journal is essential reading for IT security professionals around the world.
Computers & Security provides you with a unique blend of leading edge research and sound practical management advice. It is aimed at the professional involved with computer security, audit, control and data integrity in all sectors - industry, commerce and academia. Recognized worldwide as THE primary source of reference for applied research and technical expertise it is your first step to fully secure systems.