2012Advanced materials researchOpen access

Taint Graph of System Call Arguments for Intrusion Detection in Mobile Intelatrac

Dan Nie, Yu Hui Wang

Open full text 0 citations

Abstract

The intended data-flow in a vulnerable program is subject to be subverted by attacks which exploit buffer overflows or format string vulnerabilities to write data to unintended location. In Mobile Telecommunication it is especially important on data safety. These attacks can be classified into two types: control-flow-attacks exploit buffer overflows or other vulnerabilities to overwrite a return address, a function pointer, or some other piece of control-data; non-control-data attacks exploit similar vulnerabilities to overwrite security critical data without subverting the intended control-flow in the program. The control-flow attacks are well studied and widely used, so there are several typical approaches to prevent them, which monitor the sequence of system calls emitted by the application being monitored and utilize control-flow information of the system calls for intrusion detection. However, the non-control-data attacks are rare for the reason that they rely on specific semantics of the target applications, and there are only few works that defend them to some extent. In order to prevent non-control-data attacks, we leverage dynamic taint technique to track the instruction level relationship between different system call arguments and construct taint graph which can represent behavior profile of a benign program in this paper..

About this research paper

What this paper is about

The intended data-flow in a vulnerable program is subject to be subverted by attacks which exploit buffer overflows or format string vulnerabilities to write data to unintended location. In Mobile Telecommunication it is especially important on data safety. These attacks can be classified into two types: control-flow-attacks exploit buffer overflows or other vulnerabilities to overwrite a return address, a function pointer, or some other piece of control-data; non-control-data attacks exploit similar vulnerabilities to overwrite security critical data without subverting the intended control-flow in the program. The control-flow attacks are well studied and widely used, so there are several typical approaches to prevent them, which monitor the sequence of system calls emitted by the application being monitored and utilize control-flow information of the system calls for intrusion detection. However, the non-control-data attacks are rare for the reason that they rely on specific semantics of the target applications, and there are only few works that defend them to some extent. In order to prevent non-control-data attacks, we leverage dynamic taint technique to track the instruction level relationship between different system call arguments and construct taint graph which can represent behavior profile of a benign program in this paper..

Why it matters

A significance statement is not available in the OpenAlex record.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

The intended data-flow in a vulnerable program is subject to be subverted by attacks which exploit buffer overflows or format string vulnerabilities to write data to unintended location. In Mobile Telecommunication it is especially important on data safety. These attacks can be classified into two types: control-flow-attacks exploit buffer overflows or other vulnerabilities to overwrite a return address, a function pointer, or some other piece of control-data; non-control-data attacks exploit similar vulnerabilities to overwrite security critical data without subverting the intended control-flow in the program. The control-flow attacks are well studied and widely used, so there are several typical approaches to prevent them, which monitor the sequence of system calls emitted by the application being monitored and utilize control-flow information of the system calls for intrusion detection. However, the non-control-data attacks are rare for the reason that they rely on specific semantics of the target applications, and there are only few works that defend them to some extent. In order to prevent non-control-data attacks, we leverage dynamic taint technique to track the instruction level relationship between different system call arguments and construct taint graph which can represent behavior profile of a benign program in this paper..

Key concepts: Exploit, Buffer overflow, Computer science, Computer security, Control flow, Pointer (user interface), Leverage (statistics), Taint checking

Related papers

Back to paper searchBrowse research topicsOriginal source
Taint Graph of System Call Arguments for Intrusion Detection in Mobile Intelatrac — Research Paper | ScholarLens