Abstract:

At present, most IoT devices are remotely managed through Web service interfaces, exploiting Web service vulnerabilities to launch attacks is an important threat facing current IoT devices. This paper proposed a static mining method for Web service vulnerabilities in IoT devices. Firstly, based on the presence of certain correlation characteristics between the front-end script file and the boundary binary program (a binary program specifically designed to process Web service data), the boundary binary program in the firmware was identified by the association analysis of the front-end and back-end files, the Web input data corresponded to the processing position of the boundary binary program. Then the taint analysis technique was used to determine whether the input data will be processed by the vulnerability trigger function. Locate the area of the hazard function that existed in the program. Finally, the prototype system FBIR of the method was realized, and 10 firmwares were tested, 45 known vulnerabilities were verified by locating dangerous functions, and 12 zero-day vulnerabilities were excavated. On the basis of the false negative rate of only 14.9%, the analysis range of program was reduced by 86%.

Key words: firmware vulnerability analysis, correlation analysis, border binary program