Effectiveness of Linux rootkit detection tools
1University of Oulu, Faculty of Information Technology and Electrical Engineering, Department of Information Processing Science, Information Processing Science
|Online Access:||PDF Full Text (PDF, 1.1 MB)|
|Persistent link:|| http://urn.fi/URN:NBN:fi:oulu-202004201485
Oulu : J. Junnila,
|Publish Date:|| 2020-04-20
|Thesis type:||Master's thesis
Rootkits — a type of software that specializes in hiding entities in computer systems while enabling continuous control or access to it — are particularly difficult to detect compared to other kinds of software. Various tools exist for detecting rootkits, utilizing a wide variety of detection techniques and mechanisms. However, the effectiveness of such tools is not well established, especially in contemporary academic research and in the context of the Linux operating system.
This study carried out an empirical evaluation of the effectiveness of five tools with capabilities to detect Linux rootkits: OSSEC, AIDE, Rootkit Hunter, Chkrootkit and LKRG. The effectiveness of each tool was tested by injecting 15 publicly available rootkits in individual detection tests in virtual machines running Ubuntu 16.04, executing the detection tool and capturing its results for analysis. A total of 75 detection tests were performed.
The results showed that only 37.3% of the detection tests provided any indication of a rootkit infection or suspicious system behaviour, with the rest failing to provide any signs of anomalous behaviour. However, combining the findings of multiple detection tools increased the overall detection rate to 93.3%, as all but a single rootkit were discovered by at least one tool. Variation was observed in the effectiveness of the detection tools, with detection rates ranging from 13.3% to 53.3%. Variation in detection effectiveness was also found between categories of rootkits, as the overall detection rate was 46.7% for user mode rootkits and 31.1% for kernel mode rootkits. Overall, the findings showed that while an individual detection tool‘s effectiveness can be lacking, using a combination of tools considerably increased the likelihood of a successful detection.
© Juho Junnila, 2020. This publication is copyrighted. You may download, display and print it for your own personal use. Commercial use is prohibited.