Vulnerability Detection Algorithm of Lightweight Linux Internet of Things Application with Symbolic Execution Method

Abstract

The security of Internet of Things (IoT) devices has become a matter of great concern in recent years. The existence of security holes in the executable programs in the IoT devices has resulted in difficult to estimate security risks. For a long time, vulnerability detection is mainly completed by manual debugging and analysis, and the detection efficiency is low and the accuracy is difficult to guarantee. In this paper, the mainstream automated vulnerability analysis methods in recent years are studied, and a vulnerability detection algorithm based on symbol execution is presented. The detection algorithm is suitable for lightweight applications in small and medium-sized IoT devices. It realizes three functions: buffer overflow vulnerability detection, encryption reliability detection and protection state detection. The robustness of the detection algorithm was tested in the experiment, and the detection of overflow vulnerability program was completed within 2.75 seconds, and the detection of encryption reliability was completed within 1.79 seconds. Repeating the test with multiple sets of data showed a small difference of less than 6.4 milliseconds. The results show that the symbol execution detection algorithm presented in this paper has high detection efficiency and more robust accuracy and robustness.